fix: keep embargo tied to accepted folds and enforce PR-comment decision logs

Reviewed-on: #274

.env.example

@@ -1,23 +1,82 @@
+# ============================================================
+# The Ouroboros — Environment Configuration
+# ============================================================
+# Copy this file to .env and fill in your values.
+# Lines starting with # are comments.
+
+# ============================================================
 # Korea Investment Securities API
+# ============================================================
 KIS_APP_KEY=your_app_key_here
 KIS_APP_SECRET=your_app_secret_here
 KIS_ACCOUNT_NO=12345678-01
-KIS_BASE_URL=https://openapivts.koreainvestment.com:9443
 
+# Paper trading (VTS): https://openapivts.koreainvestment.com:29443
+# Live trading:        https://openapi.koreainvestment.com:9443
+KIS_BASE_URL=https://openapivts.koreainvestment.com:29443
+
+# ============================================================
+# Trading Mode
+# ============================================================
+# paper = 모의투자 (safe for testing), live = 실전투자 (real money)
+MODE=paper
+
+# daily = batch per session, realtime = per-stock continuous scan
+TRADE_MODE=daily
+
+# Comma-separated market codes: KR, US, JP, HK, CN, VN
+ENABLED_MARKETS=KR,US
+
+# Simulated USD cash for paper (VTS) overseas trading.
+# VTS overseas balance API often returns 0; this value is used as fallback.
+# Set to 0 to disable fallback (not used in live mode).
+PAPER_OVERSEAS_CASH=50000.0
+
+# ============================================================
 # Google Gemini
+# ============================================================
 GEMINI_API_KEY=your_gemini_api_key_here
-GEMINI_MODEL=gemini-pro
+# Recommended: gemini-2.0-flash-exp or gemini-1.5-pro
+GEMINI_MODEL=gemini-2.0-flash-exp
 
+# ============================================================
 # Risk Management
+# ============================================================
 CIRCUIT_BREAKER_PCT=-3.0
 FAT_FINGER_PCT=30.0
 CONFIDENCE_THRESHOLD=80
 
+# ============================================================
 # Database
+# ============================================================
 DB_PATH=data/trade_logs.db
 
+# ============================================================
 # Rate Limiting
-RATE_LIMIT_RPS=10.0
+# ============================================================
+# KIS API real limit is ~2 RPS. Keep at 2.0 for maximum safety.
+# Increasing this risks EGW00201 "초당 거래건수 초과" errors.
+RATE_LIMIT_RPS=2.0
 
-# Trading Mode (paper / live)
-MODE=paper
+# ============================================================
+# External Data APIs (optional)
+# ============================================================
+# NEWS_API_KEY=your_news_api_key_here
+# NEWS_API_PROVIDER=alphavantage
+# MARKET_DATA_API_KEY=your_market_data_key_here
+
+# ============================================================
+# Telegram Notifications (optional)
+# ============================================================
+# Get bot token from @BotFather on Telegram
+# Get chat ID from @userinfobot or your chat
+# TELEGRAM_BOT_TOKEN=1234567890:ABCdefGHIjklMNOpqrsTUVwxyz
+# TELEGRAM_CHAT_ID=123456789
+# TELEGRAM_ENABLED=true
+
+# ============================================================
+# Dashboard (optional)
+# ============================================================
+# DASHBOARD_ENABLED=false
+# DASHBOARD_HOST=127.0.0.1
+# DASHBOARD_PORT=8080

.gitignore

@@ -174,3 +174,7 @@ cython_debug/
 # PyPI configuration file
 .pypirc
 
+# Data files (trade logs, databases)
+# But NOT src/data/ which contains source code
+data/
+!src/data/

CLAUDE.md

@@ -1,255 +1,186 @@
-# CLAUDE.md
+# The Ouroboros
 
-This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
+AI-powered trading agent for global stock markets with self-evolution capabilities.
 
-## Git Workflow Policy
+## Quick Start
 
-**CRITICAL: All code changes MUST follow this workflow. Direct pushes to `main` are ABSOLUTELY PROHIBITED.**
-
-1. **Create Gitea Issue First** — All features, bug fixes, and policy changes require a Gitea issue before any code is written
-2. **Create Feature Branch** — Branch from `main` using format `feature/issue-{N}-{short-description}`
-3. **Implement Changes** — Write code, tests, and documentation on the feature branch
-4. **Create Pull Request** — Submit PR to `main` branch referencing the issue number
-5. **Review & Merge** — After approval, merge via PR (squash or merge commit)
-
-**Never commit directly to `main`.** This policy applies to all changes, no exceptions.
-
-## Agent Workflow
-
-**Modern AI development leverages specialized agents for concurrent, efficient task execution.**
-
-### Parallel Execution Strategy
-
-Use **git worktree** or **subagents** (via the Task tool) to handle multiple requirements simultaneously:
-
-- Each task runs in independent context
-- Parallel branches for concurrent features
-- Isolated test environments prevent interference
-- Faster iteration with distributed workload
-
-### Specialized Agent Roles
-
-Deploy task-specific agents as needed instead of handling everything in the main conversation:
-
-- **Conversational Agent** (main) — Interface with user, coordinate other agents
-- **Ticket Management Agent** — Create/update Gitea issues, track task status
-- **Design Agent** — Architectural planning, RFC documents, API design
-- **Code Writing Agent** — Implementation following specs
-- **Testing Agent** — Write tests, verify coverage, run test suites
-- **Documentation Agent** — Update docs, docstrings, CLAUDE.md, README
-- **Review Agent** — Code review, lint checks, security audits
-- **Custom Agents** — Created dynamically for specialized tasks (performance analysis, migration scripts, etc.)
-
-### When to Use Agents
-
-**Prefer spawning specialized agents for:**
-
-1. Complex multi-file changes requiring exploration
-2. Tasks with clear, isolated scope (e.g., "write tests for module X")
-3. Parallel work streams (feature A + bugfix B simultaneously)
-4. Long-running analysis (codebase search, dependency audit)
-5. Tasks requiring different contexts (multiple git worktrees)
-
-**Use the main conversation for:**
-
-1. User interaction and clarification
-2. Quick single-file edits
-3. Coordinating agent work
-4. High-level decision making
-
-### Implementation
-
-```python
-# Example: Spawn parallel test and documentation agents
-task_tool(
-    subagent_type="general-purpose",
-    prompt="Write comprehensive tests for src/markets/schedule.py",
-    description="Write schedule tests"
-)
-
-task_tool(
-    subagent_type="general-purpose",
-    prompt="Update README.md with global market feature documentation",
-    description="Update README"
-)
-```
-
-Use `run_in_background=True` for independent tasks that don't block subsequent work.
-
-## Common Command Failures
-
-**Critical: Learn from failures. Never repeat the same failed command without modification.**
-
-### tea CLI (Gitea Command Line Tool)
-
-#### ❌ TTY Error - Interactive Confirmation Fails
-```bash
-~/bin/tea issues create --repo X --title "Y" --description "Z"
-# Error: huh: could not open a new TTY: open /dev/tty: no such device or address
-```
-**💡 Reason:** tea tries to open `/dev/tty` for interactive confirmation prompts, which is unavailable in non-interactive environments.
-
-**✅ Solution:** Use `YES=""` environment variable to bypass confirmation
-```bash
-YES="" ~/bin/tea issues create --repo jihoson/The-Ouroboros --title "Title" --description "Body"
-YES="" ~/bin/tea issues edit <number> --repo jihoson/The-Ouroboros --description "Updated body"
-YES="" ~/bin/tea pulls create --repo jihoson/The-Ouroboros --head feature-branch --base main --title "Title" --description "Body"
-```
-
-**📝 Notes:**
-- Always set default login: `~/bin/tea login default local`
-- Use `--repo jihoson/The-Ouroboros` when outside repo directory
-- tea is preferred over direct Gitea API calls for consistency
-
-#### ❌ Wrong Parameter Name
-```bash
-tea issues create --body "text"
-# Error: flag provided but not defined: -body
-```
-**💡 Reason:** Parameter is `--description`, not `--body`.
-
-**✅ Solution:** Use correct parameter name
-```bash
-YES="" ~/bin/tea issues create --description "text"
-```
-
-### Gitea API (Direct HTTP Calls)
-
-#### ❌ Wrong Hostname
-```bash
-curl http://gitea.local:3000/api/v1/...
-# Error: Could not resolve host: gitea.local
-```
-**💡 Reason:** Gitea instance runs on `localhost:3000`, not `gitea.local`.
-
-**✅ Solution:** Use correct hostname (but prefer tea CLI)
-```bash
-curl http://localhost:3000/api/v1/repos/jihoson/The-Ouroboros/issues \
-  -H "Authorization: token $GITEA_TOKEN" \
-  -H "Content-Type: application/json" \
-  -d '{"title":"...", "body":"..."}'
-```
-
-**📝 Notes:**
-- Prefer `tea` CLI over direct API calls
-- Only use curl for operations tea doesn't support
-
-### Git Commands
-
-#### ❌ User Not Configured
-```bash
-git commit -m "message"
-# Error: Author identity unknown
-```
-**💡 Reason:** Git user.name and user.email not set.
-
-**✅ Solution:** Configure git user
-```bash
-git config user.name "agentson"
-git config user.email "agentson@localhost"
-```
-
-#### ❌ Permission Denied on Push
-```bash
-git push origin branch
-# Error: User permission denied for writing
-```
-**💡 Reason:** Repository access token lacks write permissions or user lacks repo write access.
-
-**✅ Solution:**
-1. Verify user has write access to repository (admin grants this)
-2. Ensure git credential has correct token with `write:repository` scope
-3. Check remote URL uses correct authentication
-
-### Python/Pytest
-
-#### ❌ Module Import Error
-```bash
-pytest tests/test_foo.py
-# ModuleNotFoundError: No module named 'src'
-```
-**💡 Reason:** Package not installed in development mode.
-
-**✅ Solution:** Install package with dev dependencies
 ```bash
+# Setup
 pip install -e ".[dev]"
-```
+cp .env.example .env
+# Edit .env with your KIS and Gemini API credentials
 
-#### ❌ Async Test Hangs
-```python
-async def test_something():  # Hangs forever
-    result = await async_function()
-```
-**💡 Reason:** Missing pytest-asyncio or wrong configuration.
+# Test
+pytest -v --cov=src
 
-**✅ Solution:** Already configured in pyproject.toml
-```toml
-[tool.pytest.ini_options]
-asyncio_mode = "auto"
-```
-No decorator needed for async tests.
-
-## Build & Test Commands
-
-```bash
-# Install all dependencies (production + dev)
-pip install ".[dev]"
-
-# Run full test suite with coverage
-pytest -v --cov=src --cov-report=term-missing
-
-# Run a single test file
-pytest tests/test_risk.py -v
-
-# Run a single test by name
-pytest tests/test_brain.py -k "test_parse_valid_json" -v
-
-# Lint
-ruff check src/ tests/
-
-# Type check (strict mode, non-blocking in CI)
-mypy src/ --strict
-
-# Run the trading agent
+# Run (paper trading)
 python -m src.main --mode=paper
 
-# Docker
-docker compose up -d ouroboros          # Run agent
-docker compose --profile test up test   # Run tests in container
+# Run with dashboard
+python -m src.main --mode=paper --dashboard
 ```
 
-## Architecture
+## Telegram Notifications (Optional)
 
-Self-evolving AI trading agent for Korean stock markets (KIS API). The main loop in `src/main.py` orchestrates four components in a 60-second cycle per stock:
+Get real-time alerts for trades, circuit breakers, and system events via Telegram.
 
-1. **Broker** (`src/broker/kis_api.py`) — Async KIS API client with automatic OAuth token refresh, leaky-bucket rate limiter (10 RPS), and POST body hash-key signing. Uses a custom SSL context with disabled hostname verification for the VTS (virtual trading) endpoint due to a known certificate mismatch.
+### Quick Setup
 
-2. **Brain** (`src/brain/gemini_client.py`) — Sends structured prompts to Google Gemini, parses JSON responses into `TradeDecision` objects. Forces HOLD when confidence < threshold (default 80). Falls back to safe HOLD on any parse/API error.
+1. **Create bot**: Message [@BotFather](https://t.me/BotFather) on Telegram → `/newbot`
+2. **Get chat ID**: Message [@userinfobot](https://t.me/userinfobot) → `/start`
+3. **Configure**: Add to `.env`:
+   ```bash
+   TELEGRAM_BOT_TOKEN=1234567890:ABCdefGHIjklMNOpqrsTUVwxyz
+   TELEGRAM_CHAT_ID=123456789
+   TELEGRAM_ENABLED=true
+   ```
+4. **Test**: Start bot conversation (`/start`), then run the agent
 
-3. **Risk Manager** (`src/core/risk_manager.py`) — **READ-ONLY by policy** (see `docs/agents.md`). Circuit breaker halts all trading via `SystemExit` when daily P&L drops below -3.0%. Fat-finger check rejects orders exceeding 30% of available cash.
+**Full documentation**: [src/notifications/README.md](src/notifications/README.md)
 
-4. **Evolution** (`src/evolution/optimizer.py`) — Analyzes high-confidence losing trades from SQLite, asks Gemini to generate new `BaseStrategy` subclasses, validates them by running the full pytest suite, and simulates PR creation.
+### What You'll Get
 
-**Data flow per cycle:** Fetch orderbook + balance → calculate P&L → get Gemini decision → validate with risk manager → execute order → log to SQLite (`src/db.py`).
+- 🟢 Trade execution alerts (BUY/SELL with confidence)
+- 🚨 Circuit breaker trips (automatic trading halt)
+- ⚠️ Fat-finger rejections (oversized orders blocked)
+- ℹ️ Market open/close notifications
+- 📝 System startup/shutdown status
 
-## Key Constraints (from `docs/agents.md`)
+### Interactive Commands
 
-- `core/risk_manager.py` is **READ-ONLY**. Changes require human approval.
-- Circuit breaker threshold (-3.0%) may only be made stricter, never relaxed.
-- Fat-finger protection (30% max order size) must always be enforced.
-- Confidence < 80 **must** force HOLD — this rule cannot be weakened.
-- All code changes require corresponding tests. Coverage must stay >= 80%.
-- Generated strategies must pass the full test suite before activation.
+With `TELEGRAM_COMMANDS_ENABLED=true` (default), the bot supports 9 bidirectional commands: `/help`, `/status`, `/positions`, `/report`, `/scenarios`, `/review`, `/dashboard`, `/stop`, `/resume`.
 
-## Configuration
+**Fail-safe**: Notifications never crash the trading system. Missing credentials or API errors are logged but trading continues normally.
 
-Pydantic Settings loaded from `.env` (see `.env.example`). Required vars: `KIS_APP_KEY`, `KIS_APP_SECRET`, `KIS_ACCOUNT_NO` (format `XXXXXXXX-XX`), `GEMINI_API_KEY`. Tests use in-memory SQLite (`DB_PATH=":memory:"`) and dummy credentials via `tests/conftest.py`.
+## Smart Volatility Scanner (Optional)
 
-## Test Structure
+Python-first filtering pipeline that reduces Gemini API calls by pre-filtering stocks using technical indicators.
 
-54 tests across four files. `asyncio_mode = "auto"` in pyproject.toml — async tests need no special decorator. The `settings` fixture in `conftest.py` provides safe defaults with test credentials and in-memory DB.
+### How It Works
 
-- `test_risk.py` (11) — Circuit breaker boundaries, fat-finger edge cases
-- `test_broker.py` (6) — Token lifecycle, rate limiting, hash keys, network errors
-- `test_brain.py` (18) — JSON parsing, confidence threshold, malformed responses, prompt construction
-- `test_market_schedule.py` (19) — Market open/close logic, timezone handling, DST, lunch breaks
+1. **Fetch Rankings** — KIS API volume surge rankings (top 30 stocks)
+2. **Python Filter** — RSI + volume ratio calculations (no AI)
+   - Volume > 200% of previous day
+   - RSI(14) < 30 (oversold) OR RSI(14) > 70 (momentum)
+3. **AI Judgment** — Only qualified candidates (1-3 stocks) sent to Gemini
+
+### Configuration
+
+Add to `.env` (optional, has sensible defaults):
+```bash
+RSI_OVERSOLD_THRESHOLD=30    # 0-50, default 30
+RSI_MOMENTUM_THRESHOLD=70    # 50-100, default 70
+VOL_MULTIPLIER=2.0           # Volume threshold (2.0 = 200%)
+SCANNER_TOP_N=3              # Max candidates per scan
+```
+
+### Benefits
+
+- **Reduces API costs** — Process 1-3 stocks instead of 20-30
+- **Python-based filtering** — Fast technical analysis before AI
+- **Evolution-ready** — Selection context logged for strategy optimization
+- **Fault-tolerant** — Falls back to static watchlist on API failure
+
+### Realtime Mode Only
+
+Smart Scanner runs in `TRADE_MODE=realtime` only. Daily mode uses static watchlists for batch efficiency.
+
+## Documentation
+
+- **[Workflow Guide](docs/workflow.md)** — Git workflow policy and agent-based development
+- **[Command Reference](docs/commands.md)** — Common failures, build commands, troubleshooting
+- **[Architecture](docs/architecture.md)** — System design, components, data flow
+- **[Context Tree](docs/context-tree.md)** — L1-L7 hierarchical memory system
+- **[Testing](docs/testing.md)** — Test structure, coverage requirements, writing tests
+- **[Agent Policies](docs/agents.md)** — Prime directives, constraints, prohibited actions
+- **[Requirements Log](docs/requirements-log.md)** — User requirements and feedback tracking
+- **[Live Trading Checklist](docs/live-trading-checklist.md)** — 모의→실전 전환 체크리스트
+
+## Core Principles
+
+1. **Safety First** — Risk manager is READ-ONLY and enforces circuit breakers
+2. **Test Everything** — 80% coverage minimum, all changes require tests
+3. **Issue-Driven Development** — All work goes through Gitea issues → feature branches → PRs
+4. **Agent Specialization** — Use dedicated agents for design, coding, testing, docs, review
+
+## Requirements Management
+
+User requirements and feedback are tracked in [docs/requirements-log.md](docs/requirements-log.md):
+
+- New requirements are added chronologically with dates
+- Code changes should reference related requirements
+- Helps maintain project evolution aligned with user needs
+- Preserves context across conversations and development cycles
+
+## Project Structure
+
+```
+src/
+├── analysis/        # Technical analysis (RSI, volatility, smart scanner)
+├── backup/          # Disaster recovery (scheduler, cloud storage, health)
+├── brain/           # Gemini AI decision engine (prompt optimizer, context selector)
+├── broker/          # KIS API client (domestic + overseas)
+├── context/         # L1-L7 hierarchical memory system
+├── core/            # Risk manager (READ-ONLY)
+├── dashboard/       # FastAPI read-only monitoring (8 API endpoints)
+├── data/            # External data integration (news, market data, calendar)
+├── evolution/       # Self-improvement (optimizer, daily review, scorecard)
+├── logging/         # Decision logger (audit trail)
+├── markets/         # Market schedules and timezone handling
+├── notifications/   # Telegram alerts + bidirectional commands (9 commands)
+├── strategy/        # Pre-market planner, scenario engine, playbook store
+├── db.py            # SQLite trade logging
+├── main.py          # Trading loop orchestrator
+└── config.py        # Settings (from .env)
+
+tests/               # 551 tests across 25 files
+docs/                # Extended documentation
+```
+
+## Key Commands
+
+```bash
+pytest -v --cov=src              # Run tests with coverage
+ruff check src/ tests/           # Lint
+mypy src/ --strict               # Type check
+
+python -m src.main --mode=paper  # Paper trading
+python -m src.main --mode=paper --dashboard  # With dashboard
+python -m src.main --mode=live   # Live trading (⚠️ real money)
+
+# Gitea workflow (requires tea CLI)
+YES="" ~/bin/tea issues create --repo jihoson/The-Ouroboros --title "..." --description "..."
+YES="" ~/bin/tea pulls create --head feature-branch --base main --title "..." --description "..."
+```
+
+## Markets Supported
+
+- 🇰🇷 Korea (KRX)
+- 🇺🇸 United States (NASDAQ, NYSE, AMEX)
+- 🇯🇵 Japan (TSE)
+- 🇭🇰 Hong Kong (SEHK)
+- 🇨🇳 China (Shanghai, Shenzhen)
+- 🇻🇳 Vietnam (Hanoi, HCM)
+
+Markets auto-detected based on timezone and enabled in `ENABLED_MARKETS` env variable.
+
+## Critical Constraints
+
+⚠️ **Non-Negotiable Rules** (see [docs/agents.md](docs/agents.md)):
+
+- `src/core/risk_manager.py` is **READ-ONLY** — changes require human approval
+- Circuit breaker at -3.0% P&L — may only be made **stricter**
+- Fat-finger protection: max 30% of cash per order — always enforced
+- Confidence 임계값 (market_outlook별, 낮출 수 없음): BEARISH ≥ 90, NEUTRAL/기본 ≥ 80, BULLISH ≥ 75
+- All code changes → corresponding tests → coverage ≥ 80%
+
+## Contributing
+
+See [docs/workflow.md](docs/workflow.md) for the complete development process.
+
+**TL;DR:**
+1. Create issue in Gitea
+2. Create feature branch: `feature/issue-N-description`
+3. Implement with tests
+4. Open PR
+5. Merge after review

README.md

@@ -10,27 +10,41 @@ KIS(한국투자증권) API로 매매하고, Google Gemini로 판단하며, 자
 │  (매매 실행) │     │ (거래 루프)  │     │  (의사결정)  │
 └─────────────┘     └──────┬──────┘     └─────────────┘
                            │
-                    ┌──────┴──────┐
-                    │Risk Manager │
-                    │  (안전장치)  │
-                    └──────┬──────┘
-                           │
-                    ┌──────┴──────┐
-                    │  Evolution  │
-                    │ (전략 진화)  │
-                    └─────────────┘
+              ┌────────────┼────────────┐
+              │            │            │
+       ┌──────┴──────┐ ┌──┴───┐ ┌──────┴──────┐
+       │Risk Manager │ │  DB  │ │  Telegram   │
+       │  (안전장치)  │ │      │ │ (알림+명령)  │
+       └──────┬──────┘ └──────┘ └─────────────┘
+              │
+     ┌────────┼────────┐
+     │        │        │
+┌────┴────┐┌──┴──┐┌────┴─────┐
+│Strategy ││Ctx  ││Evolution │
+│(플레이북)││(메모리)││ (진화)   │
+└─────────┘└─────┘└──────────┘
 ```
 
+**v2 핵심**: "Plan Once, Execute Locally" — 장 시작 전 AI가 시나리오 플레이북을 1회 생성하고, 거래 시간에는 로컬 시나리오 매칭만 수행하여 API 비용과 지연 시간을 대폭 절감.
+
 ## 핵심 모듈
 
-| 모듈 | 파일 | 설명 |
+| 모듈 | 위치 | 설명 |
 |------|------|------|
-| 설정 | `src/config.py` | Pydantic 기반 환경변수 로딩 및 타입 검증 |
-| 브로커 | `src/broker/kis_api.py` | KIS API 비동기 래퍼 (토큰 갱신, 레이트 리미터, 해시키) |
-| 두뇌 | `src/brain/gemini_client.py` | Gemini 프롬프트 구성 및 JSON 응답 파싱 |
-| 방패 | `src/core/risk_manager.py` | 서킷 브레이커 + 팻 핑거 체크 |
-| 진화 | `src/evolution/optimizer.py` | 실패 패턴 분석 → 새 전략 생성 → 테스트 → PR |
-| DB | `src/db.py` | SQLite 거래 로그 기록 |
+| 설정 | `src/config.py` | Pydantic 기반 환경변수 로딩 및 타입 검증 (35+ 변수) |
+| 브로커 | `src/broker/` | KIS API 비동기 래퍼 (국내 + 해외 9개 시장) |
+| 두뇌 | `src/brain/` | Gemini 프롬프트 구성, JSON 파싱, 토큰 최적화 |
+| 방패 | `src/core/risk_manager.py` | 서킷 브레이커 + 팻 핑거 체크 (READ-ONLY) |
+| 전략 | `src/strategy/` | Pre-Market Planner, Scenario Engine, Playbook Store |
+| 컨텍스트 | `src/context/` | L1-L7 계층형 메모리 시스템 |
+| 분석 | `src/analysis/` | RSI, ATR, Smart Volatility Scanner |
+| 알림 | `src/notifications/` | 텔레그램 양방향 (알림 + 9개 명령어) |
+| 대시보드 | `src/dashboard/` | FastAPI 읽기 전용 모니터링 (8개 API) |
+| 진화 | `src/evolution/` | 전략 진화 + Daily Review + Scorecard |
+| 의사결정 로그 | `src/logging/` | 전체 거래 결정 감사 추적 |
+| 데이터 | `src/data/` | 뉴스, 시장 데이터, 경제 캘린더 연동 |
+| 백업 | `src/backup/` | 자동 백업, S3 클라우드, 무결성 검증 |
+| DB | `src/db.py` | SQLite 거래 로그 (5개 테이블) |
 
 ## 안전장치
 
@@ -41,6 +55,7 @@ KIS(한국투자증권) API로 매매하고, Google Gemini로 판단하며, 자
 | 신뢰도 임계값 | Gemini 신뢰도 80 미만이면 강제 HOLD |
 | 레이트 리미터 | Leaky Bucket 알고리즘으로 API 호출 제한 |
 | 토큰 자동 갱신 | 만료 1분 전 자동으로 Access Token 재발급 |
+| 손절 모니터링 | 플레이북 시나리오 기반 실시간 포지션 보호 |
 
 ## 빠른 시작
 
@@ -66,7 +81,11 @@ pytest -v --cov=src --cov-report=term-missing
 ### 4. 실행 (모의투자)
 
 ```bash
+# 기본 실행
 python -m src.main --mode=paper
+
+# 대시보드 활성화
+python -m src.main --mode=paper --dashboard
 ```
 
 ### 5. Docker 실행
@@ -75,23 +94,90 @@ python -m src.main --mode=paper
 docker compose up -d ouroboros
 ```
 
+## 지원 시장
+
+| 국가 | 거래소 | 코드 |
+|------|--------|------|
+| 🇰🇷 한국 | KRX | KR |
+| 🇺🇸 미국 | NASDAQ, NYSE, AMEX | US_NASDAQ, US_NYSE, US_AMEX |
+| 🇯🇵 일본 | TSE | JP |
+| 🇭🇰 홍콩 | SEHK | HK |
+| 🇨🇳 중국 | 상하이, 선전 | CN_SHA, CN_SZA |
+| 🇻🇳 베트남 | 하노이, 호치민 | VN_HNX, VN_HSX |
+
+`ENABLED_MARKETS` 환경변수로 활성 시장 선택 (기본: `KR,US`).
+
+## 텔레그램 (선택사항)
+
+거래 실행, 서킷 브레이커 발동, 시스템 상태 등을 텔레그램으로 실시간 알림 받을 수 있습니다.
+
+### 빠른 설정
+
+1. **봇 생성**: 텔레그램에서 [@BotFather](https://t.me/BotFather) 메시지 → `/newbot` 명령
+2. **채팅 ID 확인**: [@userinfobot](https://t.me/userinfobot) 메시지 → `/start` 명령
+3. **환경변수 설정**: `.env` 파일에 추가
+   ```bash
+   TELEGRAM_BOT_TOKEN=1234567890:ABCdefGHIjklMNOpqrsTUVwxyz
+   TELEGRAM_CHAT_ID=123456789
+   TELEGRAM_ENABLED=true
+   ```
+4. **테스트**: 봇과 대화 시작 (`/start` 전송) 후 에이전트 실행
+
+**상세 문서**: [src/notifications/README.md](src/notifications/README.md)
+
+### 알림 종류
+
+- 🟢 거래 체결 알림 (BUY/SELL + 신뢰도)
+- 🚨 서킷 브레이커 발동 (자동 거래 중단)
+- ⚠️ 팻 핑거 차단 (과도한 주문 차단)
+- ℹ️ 장 시작/종료 알림
+- 📝 시스템 시작/종료 상태
+
+### 양방향 명령어
+
+`TELEGRAM_COMMANDS_ENABLED=true` (기본값) 설정 시 9개 대화형 명령어 지원:
+
+| 명령어 | 설명 |
+|--------|------|
+| `/help` | 사용 가능한 명령어 목록 |
+| `/status` | 거래 상태 (모드, 시장, P&L) |
+| `/positions` | 계좌 요약 (잔고, 현금, P&L) |
+| `/report` | 일일 요약 (거래 수, P&L, 승률) |
+| `/scenarios` | 오늘의 플레이북 시나리오 |
+| `/review` | 최근 스코어카드 (L6_DAILY) |
+| `/dashboard` | 대시보드 URL 표시 |
+| `/stop` | 거래 일시 정지 |
+| `/resume` | 거래 재개 |
+
+**안전장치**: 알림 실패해도 거래는 계속 진행됩니다.
+
 ## 테스트
 
-35개 테스트가 TDD 방식으로 구현 전에 먼저 작성되었습니다.
+551개 테스트가 25개 파일에 걸쳐 구현되어 있습니다. 최소 커버리지 80%.
 
 ```
-tests/test_risk.py   — 서킷 브레이커, 팻 핑거, 통합 검증 (11개)
-tests/test_broker.py — 토큰 관리, 타임아웃, HTTP 에러, 해시키 (6개)
-tests/test_brain.py  — JSON 파싱, 신뢰도 임계값, 비정상 응답 처리 (15개)
+tests/test_scenario_engine.py    — 시나리오 매칭 (44개)
+tests/test_data_integration.py   — 외부 데이터 연동 (38개)
+tests/test_pre_market_planner.py — 플레이북 생성 (37개)
+tests/test_main.py               — 거래 루프 통합 (37개)
+tests/test_token_efficiency.py   — 토큰 최적화 (34개)
+tests/test_strategy_models.py    — 전략 모델 검증 (33개)
+tests/test_telegram_commands.py  — 텔레그램 명령어 (31개)
+tests/test_latency_control.py    — 지연시간 제어 (30개)
+tests/test_telegram.py           — 텔레그램 알림 (25개)
+... 외 16개 파일
 ```
 
+**상세**: [docs/testing.md](docs/testing.md)
+
 ## 기술 스택
 
 - **언어**: Python 3.11+ (asyncio 기반)
-- **브로커**: KIS Open API (REST)
+- **브로커**: KIS Open API (REST, 국내+해외)
 - **AI**: Google Gemini Pro
-- **DB**: SQLite
-- **검증**: pytest + coverage
+- **DB**: SQLite (5개 테이블: trades, contexts, decision_logs, playbooks, context_metadata)
+- **대시보드**: FastAPI + uvicorn
+- **검증**: pytest + coverage (551 tests)
 - **CI/CD**: GitHub Actions
 - **배포**: Docker + Docker Compose
 
@@ -99,26 +185,50 @@ tests/test_brain.py  — JSON 파싱, 신뢰도 임계값, 비정상 응답 처
 
 ```
 The-Ouroboros/
-├── .github/workflows/ci.yml     # CI 파이프라인
 ├── docs/
-│   ├── agents.md                # AI 에이전트 페르소나 정의
-│   └── skills.md                # 사용 가능한 도구 목록
+│   ├── architecture.md          # 시스템 아키텍처
+│   ├── testing.md               # 테스트 가이드
+│   ├── commands.md              # 명령어 레퍼런스
+│   ├── context-tree.md          # L1-L7 메모리 시스템
+│   ├── workflow.md              # Git 워크플로우
+│   ├── agents.md                # 에이전트 정책
+│   ├── skills.md                # 도구 목록
+│   ├── disaster_recovery.md     # 백업/복구
+│   └── requirements-log.md      # 요구사항 기록
 ├── src/
+│   ├── analysis/                # 기술적 분석 (RSI, ATR, Smart Scanner)
+│   ├── backup/                  # 백업 (스케줄러, S3, 무결성 검증)
+│   ├── brain/                   # Gemini 의사결정 (프롬프트 최적화, 컨텍스트 선택)
+│   ├── broker/                  # KIS API (국내 + 해외)
+│   ├── context/                 # L1-L7 계층 메모리
+│   ├── core/                    # 리스크 관리 (READ-ONLY)
+│   ├── dashboard/               # FastAPI 모니터링 대시보드
+│   ├── data/                    # 외부 데이터 연동
+│   ├── evolution/               # 전략 진화 + Daily Review
+│   ├── logging/                 # 의사결정 감사 추적
+│   ├── markets/                 # 시장 스케줄 + 타임존
+│   ├── notifications/           # 텔레그램 알림 + 명령어
+│   ├── strategy/                # 플레이북 (Planner, Scenario Engine)
 │   ├── config.py                # Pydantic 설정
-│   ├── logging_config.py        # JSON 구조화 로깅
-│   ├── db.py                    # SQLite 거래 기록
-│   ├── main.py                  # 비동기 거래 루프
-│   ├── broker/kis_api.py        # KIS API 클라이언트
-│   ├── brain/gemini_client.py   # Gemini 의사결정 엔진
-│   ├── core/risk_manager.py     # 리스크 관리
-│   ├── evolution/optimizer.py   # 전략 진화 엔진
-│   └── strategies/base.py       # 전략 베이스 클래스
-├── tests/                       # TDD 테스트 스위트
+│   ├── db.py                    # SQLite 데이터베이스
+│   └── main.py                  # 비동기 거래 루프
+├── tests/                       # 551개 테스트 (25개 파일)
 ├── Dockerfile                   # 멀티스테이지 빌드
 ├── docker-compose.yml           # 서비스 오케스트레이션
 └── pyproject.toml               # 의존성 및 도구 설정
 ```
 
+## 문서
+
+- **[아키텍처](docs/architecture.md)** — 시스템 설계, 컴포넌트, 데이터 흐름
+- **[테스트](docs/testing.md)** — 테스트 구조, 커버리지, 작성 가이드
+- **[명령어](docs/commands.md)** — CLI, Dashboard, Telegram 명령어
+- **[컨텍스트 트리](docs/context-tree.md)** — L1-L7 계층 메모리
+- **[워크플로우](docs/workflow.md)** — Git 워크플로우 정책
+- **[에이전트 정책](docs/agents.md)** — 안전 제약, 금지 행위
+- **[백업/복구](docs/disaster_recovery.md)** — 재해 복구 절차
+- **[요구사항](docs/requirements-log.md)** — 사용자 요구사항 추적
+
 ## 라이선스
 
 이 프로젝트의 라이선스는 [LICENSE](LICENSE) 파일을 참조하세요.

docs/agent-constraints.md

@@ -0,0 +1,45 @@
+# Agent Constraints
+
+This document records **persistent behavioral constraints** for agents working on this repository.
+It is distinct from `docs/requirements-log.md`, which records **project/product requirements**.
+
+## Scope
+
+- Applies to all AI agents and automation that modify this repo.
+- Supplements (does not replace) `docs/agents.md` and `docs/workflow.md`.
+
+## Persistent Rules
+
+1. **Workflow enforcement**
+   - Follow `docs/workflow.md` for all changes.
+   - Create a Gitea issue before any code or documentation change.
+   - Work on a feature branch `feature/issue-{N}-{short-description}` and open a PR.
+   - Never commit directly to `main`.
+
+2. **Document-first routing**
+   - When performing work, consult relevant `docs/` files *before* making changes.
+   - Route decisions to the documented policy whenever applicable.
+   - If guidance conflicts, prefer the stricter/safety-first rule and note it in the PR.
+
+3. **Docs with code**
+   - Any code change must be accompanied by relevant documentation updates.
+   - If no doc update is needed, state the reason explicitly in the PR.
+
+4. **Session-persistent user constraints**
+   - If the user requests that a behavior should persist across sessions, record it here
+     (or in a dedicated policy doc) and reference it when working.
+   - Keep entries short and concrete, with dates.
+
+## Change Control
+
+- Changes to this file follow the same workflow as code changes.
+- Keep the history chronological and minimize rewording of existing entries.
+
+## History
+
+### 2026-02-08
+
+- Always enforce Gitea workflow: issue -> feature branch -> PR before changes.
+- When work requires guidance, consult the relevant `docs/` policies first.
+- Any code change must be accompanied by relevant documentation updates.
+- Persist user constraints across sessions by recording them in this document.

docs/architecture.md

@@ -0,0 +1,643 @@
+# System Architecture
+
+## Overview
+
+Self-evolving AI trading agent for global stock markets via KIS (Korea Investment & Securities) API. The main loop in `src/main.py` orchestrates components across multiple markets with two trading modes: daily (batch API calls) or realtime (per-stock decisions).
+
+**v2 Proactive Playbook Architecture**: The system uses a "plan once, execute locally" approach. Pre-market, the AI generates a playbook of scenarios (one Gemini API call per market per day). During trading hours, a local scenario engine matches live market data against these pre-computed scenarios — no additional AI calls needed. This dramatically reduces API costs and latency.
+
+## Trading Modes
+
+The system supports two trading frequency modes controlled by the `TRADE_MODE` environment variable:
+
+### Daily Mode (default)
+
+Optimized for Gemini Free tier API limits (20 calls/day):
+
+- **Batch decisions**: 1 API call per market per session
+- **Fixed schedule**: 4 sessions per day at 6-hour intervals (configurable)
+- **API efficiency**: Processes all stocks in a market simultaneously
+- **Use case**: Free tier users, cost-conscious deployments
+- **Configuration**:
+  ```bash
+  TRADE_MODE=daily
+  DAILY_SESSIONS=4        # Sessions per day (1-10)
+  SESSION_INTERVAL_HOURS=6  # Hours between sessions (1-24)
+  ```
+
+**Example**: With 2 markets (US, KR) and 4 sessions/day = 8 API calls/day (within 20 call limit)
+
+### Realtime Mode
+
+High-frequency trading with individual stock analysis:
+
+- **Per-stock decisions**: 1 API call per stock per cycle
+- **60-second interval**: Continuous monitoring
+- **Use case**: Production deployments with Gemini paid tier
+- **Configuration**:
+  ```bash
+  TRADE_MODE=realtime
+  ```
+
+**Note**: Realtime mode requires Gemini API subscription due to high call volume.
+
+## Core Components
+
+### 1. Broker (`src/broker/`)
+
+**KISBroker** (`kis_api.py`) — Async KIS API client for domestic Korean market
+
+- Automatic OAuth token refresh (valid for 24 hours)
+- Leaky-bucket rate limiter (configurable RPS, default 2.0)
+- POST body hash-key signing for order authentication
+- Custom SSL context with disabled hostname verification for VTS (virtual trading) endpoint due to known certificate mismatch
+- `fetch_market_rankings()` — Fetch volume surge rankings from KIS API
+- `get_daily_prices()` — Fetch OHLCV history for technical analysis
+
+**OverseasBroker** (`overseas.py`) — KIS overseas stock API wrapper
+
+- Reuses KISBroker infrastructure (session, token, rate limiter) via composition
+- Supports 9 global markets: US (NASDAQ/NYSE/AMEX), Japan, Hong Kong, China (Shanghai/Shenzhen), Vietnam (Hanoi/HCM)
+- Different API endpoints for overseas price/balance/order operations
+
+**Market Schedule** (`src/markets/schedule.py`) — Timezone-aware market management
+
+- `MarketInfo` dataclass with timezone, trading hours, lunch breaks
+- Automatic DST handling via `zoneinfo.ZoneInfo`
+- `is_market_open()` checks weekends, trading hours, lunch breaks
+- `get_open_markets()` returns currently active markets
+- `get_next_market_open()` finds next market to open and when
+- 10 global markets defined (KR, US_NASDAQ, US_NYSE, US_AMEX, JP, HK, CN_SHA, CN_SZA, VN_HNX, VN_HSX)
+
+**Overseas Ranking API Methods** (added in v0.10.x):
+- `fetch_overseas_rankings()` — Fetch overseas ranking universe (fluctuation / volume)
+- Ranking endpoint paths and TR_IDs are configurable via environment variables
+
+### 2. Analysis (`src/analysis/`)
+
+**VolatilityAnalyzer** (`volatility.py`) — Technical indicator calculations
+
+- ATR (Average True Range) for volatility measurement
+- RSI (Relative Strength Index) using Wilder's smoothing method
+- Price change percentages across multiple timeframes
+- Volume surge ratios and price-volume divergence
+- Momentum scoring (0-100 scale)
+- Breakout/breakdown pattern detection
+
+**SmartVolatilityScanner** (`smart_scanner.py`) — Python-first filtering pipeline
+
+- **Domestic (KR)**:
+  - **Step 1**: Fetch domestic fluctuation ranking as primary universe
+  - **Step 2**: Fetch domestic volume ranking for liquidity bonus
+  - **Step 3**: Compute volatility-first score (max of daily change% and intraday range%)
+  - **Step 4**: Apply liquidity bonus and return top N candidates
+- **Overseas (US/JP/HK/CN/VN)**:
+  - **Step 1**: Fetch overseas ranking universe (fluctuation rank + volume rank bonus)
+  - **Step 2**: Compute volatility-first score (max of daily change% and intraday range%)
+  - **Step 3**: Apply liquidity bonus from volume ranking
+  - **Step 4**: Return top N candidates (default 3)
+- **Fallback (overseas only)**: If ranking API is unavailable, uses dynamic universe
+  from runtime active symbols + recent traded symbols + current holdings (no static watchlist)
+- **Realtime mode only**: Daily mode uses batch processing for API efficiency
+
+**Benefits:**
+- Reduces Gemini API calls from 20-30 stocks to 1-3 qualified candidates
+- Fast Python-based filtering before expensive AI judgment
+- Logs selection context (RSI-compatible proxy, volume_ratio, signal, score) for Evolution system
+
+### 3. Brain (`src/brain/`)
+
+**GeminiClient** (`gemini_client.py`) — AI decision engine powered by Google Gemini
+
+- Constructs structured prompts from market data
+- Parses JSON responses into `TradeDecision` objects (`action`, `confidence`, `rationale`)
+- Forces HOLD when confidence < threshold (default 80)
+- Falls back to safe HOLD on any parse/API error
+- Handles markdown-wrapped JSON, malformed responses, invalid actions
+
+**PromptOptimizer** (`prompt_optimizer.py`) — Token efficiency optimization
+
+- Reduces prompt size while preserving decision quality
+- Caches optimized prompts
+
+**ContextSelector** (`context_selector.py`) — Relevant context selection for prompts
+
+- Selects appropriate context layers for current market conditions
+
+### 4. Risk Manager (`src/core/risk_manager.py`)
+
+**RiskManager** — Safety circuit breaker and order validation
+
+> **READ-ONLY by policy** (see [`docs/agents.md`](./agents.md))
+
+- **Circuit Breaker**: Halts all trading via `SystemExit` when daily P&L drops below -3.0%
+  - Threshold may only be made stricter, never relaxed
+  - Calculated as `(total_eval - purchase_total) / purchase_total * 100`
+- **Fat-Finger Protection**: Rejects orders exceeding 30% of available cash
+  - Must always be enforced, cannot be disabled
+
+### 5. Strategy (`src/strategy/`)
+
+**Pre-Market Planner** (`pre_market_planner.py`) — AI playbook generation
+
+- Runs before market open (configurable `PRE_MARKET_MINUTES`, default 30)
+- Generates scenario-based playbooks via single Gemini API call per market
+- Handles timeout (`PLANNER_TIMEOUT_SECONDS`, default 60) with defensive playbook fallback
+- Persists playbooks to database for audit trail
+
+**Scenario Engine** (`scenario_engine.py`) — Local scenario matching
+
+- Matches live market data against pre-computed playbook scenarios
+- No AI calls during trading hours — pure Python matching logic
+- Returns matched scenarios with confidence scores
+- Configurable `MAX_SCENARIOS_PER_STOCK` (default 5)
+- Periodic rescan at `RESCAN_INTERVAL_SECONDS` (default 300)
+
+**Playbook Store** (`playbook_store.py`) — Playbook persistence
+
+- SQLite-backed storage for daily playbooks
+- Date and market-based retrieval
+- Status tracking (generated, active, expired)
+
+**Models** (`models.py`) — Pydantic data models
+
+- Scenario, Playbook, MatchResult, and related type definitions
+
+### 6. Context System (`src/context/`)
+
+**Context Store** (`store.py`) — L1-L7 hierarchical memory
+
+- 7-layer context system (see [docs/context-tree.md](./context-tree.md)):
+  - L1: Tick-level (real-time price)
+  - L2: Intraday (session summary)
+  - L3: Daily (end-of-day)
+  - L4: Weekly (trend analysis)
+  - L5: Monthly (strategy review)
+  - L6: Daily Review (scorecard)
+  - L7: Evolution (long-term learning)
+- Key-value storage with timeframe tagging
+- SQLite persistence in `contexts` table
+
+**Context Scheduler** (`scheduler.py`) — Periodic aggregation
+
+- Scheduled summarization from lower to higher layers
+- Configurable aggregation intervals
+
+**Context Summarizer** (`summarizer.py`) — Layer summarization
+
+- Aggregates lower-layer data into higher-layer summaries
+
+### 7. Dashboard (`src/dashboard/`)
+
+**FastAPI App** (`app.py`) — Read-only monitoring dashboard
+
+- Runs as daemon thread when enabled (`--dashboard` CLI flag or `DASHBOARD_ENABLED=true`)
+- Configurable host/port (`DASHBOARD_HOST`, `DASHBOARD_PORT`, default `127.0.0.1:8080`)
+- Serves static HTML frontend
+
+**8 API Endpoints:**
+
+| Endpoint | Method | Description |
+|----------|--------|-------------|
+| `/` | GET | Static HTML dashboard |
+| `/api/status` | GET | Daily trading status by market |
+| `/api/playbook/{date}` | GET | Playbook for specific date and market |
+| `/api/scorecard/{date}` | GET | Daily scorecard from L6_DAILY context |
+| `/api/performance` | GET | Trading performance metrics (by market + combined) |
+| `/api/context/{layer}` | GET | Query context by layer (L1-L7) |
+| `/api/decisions` | GET | Decision log entries with outcomes |
+| `/api/scenarios/active` | GET | Today's matched scenarios |
+
+### 8. Notifications (`src/notifications/telegram_client.py`)
+
+**TelegramClient** — Real-time event notifications via Telegram Bot API
+
+- Sends alerts for trades, circuit breakers, fat-finger rejections, system events
+- Non-blocking: failures are logged but never crash trading
+- Rate-limited: 1 message/second default to respect Telegram API limits
+- Auto-disabled when credentials missing
+
+**TelegramCommandHandler** — Bidirectional command interface
+
+- Long polling from Telegram API (configurable `TELEGRAM_POLLING_INTERVAL`)
+- 9 interactive commands: `/help`, `/status`, `/positions`, `/report`, `/scenarios`, `/review`, `/dashboard`, `/stop`, `/resume`
+- Authorization filtering by `TELEGRAM_CHAT_ID`
+- Enable/disable via `TELEGRAM_COMMANDS_ENABLED` (default: true)
+
+**Notification Types:**
+- Trade execution (BUY/SELL with confidence)
+- Circuit breaker trips (critical alert)
+- Fat-finger protection triggers (order rejection)
+- Market open/close events
+- System startup/shutdown status
+- Playbook generation results
+- Stop-loss monitoring alerts
+
+### 9. Evolution (`src/evolution/`)
+
+**StrategyOptimizer** (`optimizer.py`) — Self-improvement loop
+
+- Analyzes high-confidence losing trades from SQLite
+- Asks Gemini to generate new `BaseStrategy` subclasses
+- Validates generated strategies by running full pytest suite
+- Simulates PR creation for human review
+- Only activates strategies that pass all tests
+
+**DailyReview** (`daily_review.py`) — End-of-day review
+
+- Generates comprehensive trade performance summary
+- Stores results in L6_DAILY context layer
+- Tracks win rate, P&L, confidence accuracy
+
+**DailyScorecard** (`scorecard.py`) — Performance scoring
+
+- Calculates daily metrics (trades, P&L, win rate, avg confidence)
+- Enables trend tracking across days
+
+**Stop-Loss Monitoring** — Real-time position protection
+
+- Monitors positions against stop-loss levels from playbook scenarios
+- Sends Telegram alerts when thresholds approached or breached
+
+### 10. Decision Logger (`src/logging/decision_logger.py`)
+
+**DecisionLogger** — Comprehensive audit trail
+
+- Logs every trading decision with full context snapshot
+- Captures input data, rationale, confidence, and outcomes
+- Supports outcome tracking (P&L, accuracy) for post-analysis
+- Stored in `decision_logs` table with indexed queries
+- Review workflow support (reviewed flag, review notes)
+
+### 11. Data Integration (`src/data/`)
+
+**External Data Sources** (optional):
+
+- `news_api.py` — News sentiment data
+- `market_data.py` — Extended market data
+- `economic_calendar.py` — Economic event calendar
+
+### 12. Backup (`src/backup/`)
+
+**Disaster Recovery** (see [docs/disaster_recovery.md](./disaster_recovery.md)):
+
+- `scheduler.py` — Automated backup scheduling
+- `exporter.py` — Data export to various formats
+- `cloud_storage.py` — S3-compatible cloud backup
+- `health_monitor.py` — Backup integrity verification
+
+## Data Flow
+
+### Playbook Mode (Daily — Primary v2 Flow)
+
+```
+┌─────────────────────────────────────────────────────────────┐
+│ Pre-Market Phase (before market open)                       │
+└─────────────────────────────────────────────────────────────┘
+                           │
+                           ▼
+        ┌──────────────────────────────────┐
+        │ Pre-Market Planner               │
+        │ - 1 Gemini API call per market   │
+        │ - Generate scenario playbook     │
+        │ - Store in playbooks table       │
+        └──────────────────┬───────────────┘
+                           │
+                           ▼
+┌─────────────────────────────────────────────────────────────┐
+│ Trading Hours (market open → close)                         │
+└─────────────────────────────────────────────────────────────┘
+                           │
+                           ▼
+        ┌──────────────────────────────────┐
+        │ Market Schedule Check            │
+        │ - Get open markets               │
+        │ - Filter by enabled markets      │
+        └──────────────────┬───────────────┘
+                           │
+                           ▼
+        ┌──────────────────────────────────┐
+        │ Scenario Engine (local)          │
+        │ - Match live data vs playbook    │
+        │ - No AI calls needed             │
+        │ - Return matched scenarios       │
+        └──────────────────┬───────────────┘
+                           │
+                           ▼
+        ┌──────────────────────────────────┐
+        │ Risk Manager: Validate Order     │
+        │ - Check circuit breaker          │
+        │ - Check fat-finger limit         │
+        └──────────────────┬───────────────┘
+                           │
+                           ▼
+        ┌──────────────────────────────────┐
+        │ Broker: Execute Order            │
+        │ - Domestic: send_order()         │
+        │ - Overseas: send_overseas_order()│
+        └──────────────────┬───────────────┘
+                           │
+                           ▼
+        ┌──────────────────────────────────┐
+        │ Decision Logger + DB             │
+        │ - Full audit trail               │
+        │ - Context snapshot               │
+        │ - Telegram notification          │
+        └──────────────────┬───────────────┘
+                           │
+                           ▼
+┌─────────────────────────────────────────────────────────────┐
+│ Post-Market Phase                                           │
+└─────────────────────────────────────────────────────────────┘
+                           │
+                           ▼
+        ┌──────────────────────────────────┐
+        │ Daily Review + Scorecard         │
+        │ - Performance summary            │
+        │ - Store in L6_DAILY context      │
+        │ - Evolution learning             │
+        └──────────────────────────────────┘
+```
+
+### Realtime Mode (with Smart Scanner)
+
+```
+┌─────────────────────────────────────────────────────────────┐
+│ Main Loop (60s cycle per market)                            │
+└─────────────────────────────────────────────────────────────┘
+                           │
+                           ▼
+        ┌──────────────────────────────────┐
+        │ Market Schedule Check            │
+        │ - Get open markets               │
+        │ - Filter by enabled markets      │
+        │ - Wait if all closed             │
+        └──────────────────┬───────────────┘
+                           │
+                           ▼
+        ┌──────────────────────────────────┐
+        │ Smart Scanner (Python-first)      │
+        │ - Domestic: fluctuation rank     │
+        │   + volume rank bonus            │
+        │   + volatility-first scoring     │
+        │ - Overseas: ranking universe     │
+        │   + volatility-first scoring     │
+        │ - Fallback: dynamic universe     │
+        │ - Return top 3 qualified stocks  │
+        └──────────────────┬───────────────┘
+                           │
+                           ▼
+        ┌──────────────────────────────────┐
+        │ For Each Qualified Candidate     │
+        └──────────────────┬───────────────┘
+                           │
+                           ▼
+        ┌──────────────────────────────────┐
+        │ Broker: Fetch Market Data        │
+        │ - Domestic: orderbook + balance  │
+        │ - Overseas: price + balance      │
+        └──────────────────┬───────────────┘
+                           │
+                           ▼
+        ┌──────────────────────────────────┐
+        │ Brain: Get Decision (AI)         │
+        │ - Build prompt with market data  │
+        │ - Call Gemini API                │
+        │ - Parse JSON response            │
+        │ - Return TradeDecision           │
+        └──────────────────┬───────────────┘
+                           │
+                           ▼
+        ┌──────────────────────────────────┐
+        │ Risk Manager: Validate Order     │
+        │ - Check circuit breaker          │
+        │ - Check fat-finger limit         │
+        └──────────────────┬───────────────┘
+                           │
+                           ▼
+        ┌──────────────────────────────────┐
+        │ Broker: Execute Order            │
+        │ - Domestic: send_order()         │
+        │ - Overseas: send_overseas_order()│
+        └──────────────────┬───────────────┘
+                           │
+                           ▼
+        ┌──────────────────────────────────┐
+        │ Decision Logger + Notifications  │
+        │ - Log trade to SQLite            │
+        │ - selection_context (JSON)       │
+        │ - Telegram notification          │
+        └──────────────────────────────────┘
+```
+
+## Database Schema
+
+**SQLite** (`src/db.py`) — Database: `data/trades.db`
+
+### trades
+```sql
+CREATE TABLE trades (
+    id INTEGER PRIMARY KEY AUTOINCREMENT,
+    timestamp TEXT NOT NULL,
+    stock_code TEXT NOT NULL,
+    action TEXT NOT NULL,          -- BUY | SELL | HOLD
+    confidence INTEGER NOT NULL,   -- 0-100
+    rationale TEXT,
+    quantity INTEGER,
+    price REAL,
+    pnl REAL DEFAULT 0.0,
+    market TEXT DEFAULT 'KR',
+    exchange_code TEXT DEFAULT 'KRX',
+    selection_context TEXT,        -- JSON: {rsi, volume_ratio, signal, score}
+    decision_id TEXT              -- Links to decision_logs
+);
+```
+
+### contexts
+```sql
+CREATE TABLE contexts (
+    id INTEGER PRIMARY KEY AUTOINCREMENT,
+    layer TEXT NOT NULL,           -- L1 through L7
+    timeframe TEXT,
+    key TEXT NOT NULL,
+    value TEXT NOT NULL,           -- JSON data
+    created_at TEXT NOT NULL,
+    updated_at TEXT NOT NULL
+);
+-- Indices: idx_contexts_layer, idx_contexts_timeframe, idx_contexts_updated
+```
+
+### decision_logs
+```sql
+CREATE TABLE decision_logs (
+    decision_id TEXT PRIMARY KEY,
+    timestamp TEXT NOT NULL,
+    stock_code TEXT,
+    market TEXT,
+    exchange_code TEXT,
+    action TEXT,
+    confidence INTEGER,
+    rationale TEXT,
+    context_snapshot TEXT,         -- JSON: full context at decision time
+    input_data TEXT,              -- JSON: market data used
+    outcome_pnl REAL,
+    outcome_accuracy REAL,
+    reviewed INTEGER DEFAULT 0,
+    review_notes TEXT
+);
+-- Indices: idx_decision_logs_timestamp, idx_decision_logs_reviewed, idx_decision_logs_confidence
+```
+
+### playbooks
+```sql
+CREATE TABLE playbooks (
+    id INTEGER PRIMARY KEY AUTOINCREMENT,
+    date TEXT NOT NULL,
+    market TEXT NOT NULL,
+    status TEXT DEFAULT 'generated',
+    playbook_json TEXT NOT NULL,   -- Full playbook with scenarios
+    generated_at TEXT NOT NULL,
+    token_count INTEGER,
+    scenario_count INTEGER,
+    match_count INTEGER DEFAULT 0
+);
+-- Indices: idx_playbooks_date, idx_playbooks_market
+```
+
+### context_metadata
+```sql
+CREATE TABLE context_metadata (
+    layer TEXT PRIMARY KEY,
+    description TEXT,
+    retention_days INTEGER,
+    aggregation_source TEXT
+);
+```
+
+## Configuration
+
+**Pydantic Settings** (`src/config.py`)
+
+Loaded from `.env` file:
+
+```bash
+# Required
+KIS_APP_KEY=your_app_key
+KIS_APP_SECRET=your_app_secret
+KIS_ACCOUNT_NO=XXXXXXXX-XX
+GEMINI_API_KEY=your_gemini_key
+
+# Optional — Trading Mode
+MODE=paper                    # paper | live
+TRADE_MODE=daily              # daily | realtime
+DAILY_SESSIONS=4              # Sessions per day (daily mode only)
+SESSION_INTERVAL_HOURS=6      # Hours between sessions (daily mode only)
+
+# Optional — Database
+DB_PATH=data/trades.db
+
+# Optional — Risk
+CONFIDENCE_THRESHOLD=80
+MAX_LOSS_PCT=3.0
+MAX_ORDER_PCT=30.0
+
+# Optional — Markets
+ENABLED_MARKETS=KR,US         # Comma-separated market codes
+RATE_LIMIT_RPS=2.0            # KIS API requests per second
+
+# Optional — Pre-Market Planner (v2)
+PRE_MARKET_MINUTES=30         # Minutes before market open to generate playbook
+MAX_SCENARIOS_PER_STOCK=5     # Max scenarios per stock in playbook
+PLANNER_TIMEOUT_SECONDS=60    # Timeout for playbook generation
+DEFENSIVE_PLAYBOOK_ON_FAILURE=true  # Fallback on AI failure
+RESCAN_INTERVAL_SECONDS=300   # Scenario rescan interval during trading
+
+# Optional — Smart Scanner (realtime mode only)
+RSI_OVERSOLD_THRESHOLD=30     # 0-50, oversold threshold
+RSI_MOMENTUM_THRESHOLD=70     # 50-100, momentum threshold
+VOL_MULTIPLIER=2.0            # Minimum volume ratio (2.0 = 200%)
+SCANNER_TOP_N=3               # Max qualified candidates per scan
+
+# Optional — Dashboard
+DASHBOARD_ENABLED=false       # Enable FastAPI dashboard
+DASHBOARD_HOST=127.0.0.1      # Dashboard bind address
+DASHBOARD_PORT=8080           # Dashboard port (1-65535)
+
+# Optional — Telegram
+TELEGRAM_BOT_TOKEN=1234567890:ABCdefGHIjklMNOpqrsTUVwxyz
+TELEGRAM_CHAT_ID=123456789
+TELEGRAM_ENABLED=true
+TELEGRAM_COMMANDS_ENABLED=true   # Enable bidirectional commands
+TELEGRAM_POLLING_INTERVAL=1.0    # Command polling interval (seconds)
+
+# Optional — Backup
+BACKUP_ENABLED=false
+BACKUP_DIR=data/backups
+S3_ENDPOINT_URL=...
+S3_ACCESS_KEY=...
+S3_SECRET_KEY=...
+S3_BUCKET_NAME=...
+S3_REGION=...
+
+# Optional — External Data
+NEWS_API_KEY=...
+NEWS_API_PROVIDER=...
+MARKET_DATA_API_KEY=...
+
+# Position Sizing (optional)
+POSITION_SIZING_ENABLED=true
+POSITION_BASE_ALLOCATION_PCT=5.0
+POSITION_MIN_ALLOCATION_PCT=1.0
+POSITION_MAX_ALLOCATION_PCT=10.0
+POSITION_VOLATILITY_TARGET_SCORE=50.0
+
+# Legacy/compat scanner thresholds (kept for backward compatibility)
+RSI_OVERSOLD_THRESHOLD=30
+RSI_MOMENTUM_THRESHOLD=70
+VOL_MULTIPLIER=2.0
+
+# Overseas Ranking API (optional override; account-dependent)
+OVERSEAS_RANKING_ENABLED=true
+OVERSEAS_RANKING_FLUCT_TR_ID=HHDFS76200100
+OVERSEAS_RANKING_VOLUME_TR_ID=HHDFS76200200
+OVERSEAS_RANKING_FLUCT_PATH=/uapi/overseas-price/v1/quotations/inquire-updown-rank
+OVERSEAS_RANKING_VOLUME_PATH=/uapi/overseas-price/v1/quotations/inquire-volume-rank
+```
+
+Tests use in-memory SQLite (`DB_PATH=":memory:"`) and dummy credentials via `tests/conftest.py`.
+
+## Error Handling
+
+### Connection Errors (Broker API)
+- Retry with exponential backoff (2^attempt seconds)
+- Max 3 retries per stock
+- After exhaustion, skip stock and continue with next
+
+### API Quota Errors (Gemini)
+- Return safe HOLD decision with confidence=0
+- Log error but don't crash
+- Agent continues trading on next cycle
+
+### Circuit Breaker Tripped
+- Immediately halt via `SystemExit`
+- Log critical message
+- Requires manual intervention to restart
+
+### Market Closed
+- Wait until next market opens
+- Use `get_next_market_open()` to calculate wait time
+- Sleep until market open time
+
+### Telegram API Errors
+- Log warning but continue trading
+- Missing credentials → auto-disable notifications
+- Network timeout → skip notification, no retry
+- Invalid token → log error, trading unaffected
+- Rate limit exceeded → queued via rate limiter
+
+### Playbook Generation Failure
+- Timeout → fall back to defensive playbook (`DEFENSIVE_PLAYBOOK_ON_FAILURE`)
+- API error → use previous day's playbook if available
+- No playbook → skip pre-market phase, fall back to direct AI calls
+
+**Guarantee**: Notification and dashboard failures never interrupt trading operations.

docs/commands.md

@@ -0,0 +1,227 @@
+# Command Reference
+
+## Common Command Failures
+
+**Critical: Learn from failures. Never repeat the same failed command without modification.**
+
+### tea CLI (Gitea Command Line Tool)
+
+#### ❌ TTY Error - Interactive Confirmation Fails
+```bash
+~/bin/tea issues create --repo X --title "Y" --description "Z"
+# Error: huh: could not open a new TTY: open /dev/tty: no such device or address
+```
+**💡 Reason:** tea tries to open `/dev/tty` for interactive confirmation prompts, which is unavailable in non-interactive environments.
+
+**✅ Solution:** Use `YES=""` environment variable to bypass confirmation
+```bash
+YES="" ~/bin/tea issues create --repo jihoson/The-Ouroboros --title "Title" --description "Body"
+YES="" ~/bin/tea issues edit <number> --repo jihoson/The-Ouroboros --description "Updated body"
+YES="" ~/bin/tea pulls create --repo jihoson/The-Ouroboros --head feature-branch --base main --title "Title" --description "Body"
+```
+
+**📝 Notes:**
+- Always set default login: `~/bin/tea login default local`
+- Use `--repo jihoson/The-Ouroboros` when outside repo directory
+- tea is preferred over direct Gitea API calls for consistency
+
+#### ❌ Wrong Parameter Name
+```bash
+tea issues create --body "text"
+# Error: flag provided but not defined: -body
+```
+**💡 Reason:** Parameter is `--description`, not `--body`.
+
+**✅ Solution:** Use correct parameter name
+```bash
+YES="" ~/bin/tea issues create --description "text"
+```
+
+### Gitea API (Direct HTTP Calls)
+
+#### ❌ Wrong Hostname
+```bash
+curl http://gitea.local:3000/api/v1/...
+# Error: Could not resolve host: gitea.local
+```
+**💡 Reason:** Gitea instance runs on `localhost:3000`, not `gitea.local`.
+
+**✅ Solution:** Use correct hostname (but prefer tea CLI)
+```bash
+curl http://localhost:3000/api/v1/repos/jihoson/The-Ouroboros/issues \
+  -H "Authorization: token $GITEA_TOKEN" \
+  -H "Content-Type: application/json" \
+  -d '{"title":"...", "body":"..."}'
+```
+
+**📝 Notes:**
+- Prefer `tea` CLI over direct API calls
+- Only use curl for operations tea doesn't support
+
+### Git Commands
+
+#### ❌ User Not Configured
+```bash
+git commit -m "message"
+# Error: Author identity unknown
+```
+**💡 Reason:** Git user.name and user.email not set.
+
+**✅ Solution:** Configure git user
+```bash
+git config user.name "agentson"
+git config user.email "agentson@localhost"
+```
+
+#### ❌ Permission Denied on Push
+```bash
+git push origin branch
+# Error: User permission denied for writing
+```
+**💡 Reason:** Repository access token lacks write permissions or user lacks repo write access.
+
+**✅ Solution:**
+1. Verify user has write access to repository (admin grants this)
+2. Ensure git credential has correct token with `write:repository` scope
+3. Check remote URL uses correct authentication
+
+### Python/Pytest
+
+#### ❌ Module Import Error
+```bash
+pytest tests/test_foo.py
+# ModuleNotFoundError: No module named 'src'
+```
+**💡 Reason:** Package not installed in development mode.
+
+**✅ Solution:** Install package with dev dependencies
+```bash
+pip install -e ".[dev]"
+```
+
+#### ❌ Async Test Hangs
+```python
+async def test_something():  # Hangs forever
+    result = await async_function()
+```
+**💡 Reason:** Missing pytest-asyncio or wrong configuration.
+
+**✅ Solution:** Already configured in pyproject.toml
+```toml
+[tool.pytest.ini_options]
+asyncio_mode = "auto"
+```
+No decorator needed for async tests.
+
+## Build & Test Commands
+
+```bash
+# Install all dependencies (production + dev)
+pip install -e ".[dev]"
+
+# Run full test suite with coverage (551 tests across 25 files)
+pytest -v --cov=src --cov-report=term-missing
+
+# Run a single test file
+pytest tests/test_risk.py -v
+
+# Run a single test by name
+pytest tests/test_brain.py -k "test_parse_valid_json" -v
+
+# Lint
+ruff check src/ tests/
+
+# Type check (strict mode, non-blocking in CI)
+mypy src/ --strict
+
+# Run the trading agent
+python -m src.main --mode=paper
+
+# Run with dashboard enabled
+python -m src.main --mode=paper --dashboard
+
+# Docker
+docker compose up -d ouroboros          # Run agent
+docker compose --profile test up test   # Run tests in container
+```
+
+## Dashboard
+
+The FastAPI dashboard provides read-only monitoring of the trading system.
+
+### Starting the Dashboard
+
+```bash
+# Via CLI flag
+python -m src.main --mode=paper --dashboard
+
+# Via environment variable
+DASHBOARD_ENABLED=true python -m src.main --mode=paper
+```
+
+Dashboard runs as a daemon thread on `DASHBOARD_HOST:DASHBOARD_PORT` (default: `127.0.0.1:8080`).
+
+### API Endpoints
+
+| Endpoint | Description |
+|----------|-------------|
+| `GET /` | HTML dashboard UI |
+| `GET /api/status` | Daily trading status by market |
+| `GET /api/playbook/{date}` | Playbook for specific date (query: `market`) |
+| `GET /api/scorecard/{date}` | Daily scorecard from L6_DAILY context |
+| `GET /api/performance` | Performance metrics by market and combined |
+| `GET /api/context/{layer}` | Context data by layer L1-L7 (query: `timeframe`) |
+| `GET /api/decisions` | Decision log entries (query: `limit`, `market`) |
+| `GET /api/scenarios/active` | Today's matched scenarios |
+
+## Telegram Commands
+
+When `TELEGRAM_COMMANDS_ENABLED=true` (default), the bot accepts these interactive commands:
+
+| Command | Description |
+|---------|-------------|
+| `/help` | List available commands |
+| `/status` | Show trading status (mode, markets, P&L) |
+| `/positions` | Display account summary (balance, cash, P&L) |
+| `/report` | Daily summary metrics (trades, P&L, win rate) |
+| `/scenarios` | Show today's playbook scenarios |
+| `/review` | Display recent scorecards (L6_DAILY layer) |
+| `/dashboard` | Show dashboard URL if enabled |
+| `/stop` | Pause trading |
+| `/resume` | Resume trading |
+
+Commands are only processed from the authorized `TELEGRAM_CHAT_ID`.
+
+## KIS API TR_ID 참조 문서
+
+**TR_ID를 추가하거나 수정할 때 반드시 공식 문서를 먼저 확인할 것.**
+
+공식 문서: `docs/한국투자증권_오픈API_전체문서_20260221_030000.xlsx`
+
+> ⚠️ 커뮤니티 블로그, GitHub 예제 등 비공식 자료의 TR_ID는 오래되거나 틀릴 수 있음.
+> 실제로 `VTTT1006U`(미국 매도 — 잘못됨)가 오랫동안 코드에 남아있던 사례가 있음 (Issue #189).
+
+### 주요 TR_ID 목록
+
+| 구분 | 모의투자 TR_ID | 실전투자 TR_ID | 시트명 |
+|------|---------------|---------------|--------|
+| 해외주식 매수 (미국) | `VTTT1002U` | `TTTT1002U` | 해외주식 주문 |
+| 해외주식 매도 (미국) | `VTTT1001U` | `TTTT1006U` | 해외주식 주문 |
+
+새로운 TR_ID가 필요할 때:
+1. 위 xlsx 파일에서 해당 거래 유형의 시트를 찾는다.
+2. 모의투자(`VTTT`) / 실전투자(`TTTT`) 컬럼을 구분하여 정확한 값을 사용한다.
+3. 코드에 출처 주석을 남긴다: `# Source: 한국투자증권_오픈API_전체문서 — '<시트명>' 시트`
+
+## Environment Setup
+
+```bash
+# Create .env file from example
+cp .env.example .env
+
+# Edit .env with your credentials
+# Required: KIS_APP_KEY, KIS_APP_SECRET, KIS_ACCOUNT_NO, GEMINI_API_KEY
+
+# Verify configuration
+python -c "from src.config import Settings; print(Settings())"
+```

docs/context-tree.md

@@ -0,0 +1,338 @@
+# Context Tree: Multi-Layered Memory Management
+
+The context tree implements **Pillar 2** of The Ouroboros: hierarchical memory management across 7 time horizons, from real-time market data to generational trading wisdom.
+
+## Overview
+
+Instead of a flat memory structure, The Ouroboros maintains a **7-tier context tree** where each layer represents a different time horizon and level of abstraction:
+
+```
+L1 (Legacy)      ←  Cumulative wisdom across generations
+  ↑
+L2 (Annual)      ←  Yearly performance metrics
+  ↑
+L3 (Quarterly)   ←  Quarterly strategy adjustments
+  ↑
+L4 (Monthly)     ←  Monthly portfolio rebalancing
+  ↑
+L5 (Weekly)      ←  Weekly stock selection
+  ↑
+L6 (Daily)       ←  Daily trade logs
+  ↑
+L7 (Real-time)   ←  Live market data
+```
+
+Data flows **bottom-up**: real-time trades aggregate into daily summaries, which roll up to weekly, then monthly, quarterly, annual, and finally into permanent legacy knowledge.
+
+## The 7 Layers
+
+### L7: Real-time
+**Retention**: 7 days
+**Timeframe format**: `YYYY-MM-DD` (same-day)
+**Content**: Current positions, live quotes, orderbook snapshots, tick-by-tick volatility
+
+**Use cases**:
+- Immediate execution decisions
+- Stop-loss triggers
+- Real-time P&L tracking
+
+**Example keys**:
+- `current_position_{stock_code}`: Current holdings
+- `live_price_{stock_code}`: Latest quote
+- `volatility_5m_{stock_code}`: 5-minute rolling volatility
+
+### L6: Daily
+**Retention**: 90 days
+**Timeframe format**: `YYYY-MM-DD`
+**Content**: Daily trade logs, end-of-day P&L, market summaries, decision accuracy
+
+**Use cases**:
+- Daily performance review
+- Identify patterns in recent trading
+- Backtest strategy adjustments
+
+**Example keys**:
+- `total_pnl`: Daily profit/loss
+- `trade_count`: Number of trades
+- `win_rate`: Percentage of profitable trades
+- `avg_confidence`: Average Gemini confidence
+
+### L5: Weekly
+**Retention**: 1 year
+**Timeframe format**: `YYYY-Www` (ISO week, e.g., `2026-W06`)
+**Content**: Weekly stock selection, sector rotation, volatility regime classification
+
+**Use cases**:
+- Weekly strategy adjustment
+- Sector momentum tracking
+- Identify hot/cold markets
+
+**Example keys**:
+- `weekly_pnl`: Week's total P&L
+- `top_performers`: Best-performing stocks
+- `sector_focus`: Dominant sectors
+- `avg_confidence`: Weekly average confidence
+
+### L4: Monthly
+**Retention**: 2 years
+**Timeframe format**: `YYYY-MM`
+**Content**: Monthly portfolio rebalancing, risk exposure analysis, drawdown recovery
+
+**Use cases**:
+- Monthly performance reporting
+- Risk exposure adjustment
+- Correlation analysis
+
+**Example keys**:
+- `monthly_pnl`: Month's total P&L
+- `sharpe_ratio`: Risk-adjusted return
+- `max_drawdown`: Largest peak-to-trough decline
+- `rebalancing_notes`: Manual insights
+
+### L3: Quarterly
+**Retention**: 3 years
+**Timeframe format**: `YYYY-Qn` (e.g., `2026-Q1`)
+**Content**: Quarterly strategy pivots, market phase detection (bull/bear/sideways), macro regime changes
+
+**Use cases**:
+- Strategic pivots (e.g., growth → value)
+- Macro regime classification
+- Long-term pattern recognition
+
+**Example keys**:
+- `quarterly_pnl`: Quarter's total P&L
+- `market_phase`: Bull/Bear/Sideways
+- `strategy_adjustments`: Major changes made
+- `lessons_learned`: Key insights
+
+### L2: Annual
+**Retention**: 10 years
+**Timeframe format**: `YYYY`
+**Content**: Yearly returns, Sharpe ratio, max drawdown, win rate, strategy effectiveness
+
+**Use cases**:
+- Annual performance review
+- Multi-year trend analysis
+- Strategy benchmarking
+
+**Example keys**:
+- `annual_pnl`: Year's total P&L
+- `sharpe_ratio`: Annual risk-adjusted return
+- `win_rate`: Yearly win percentage
+- `best_strategy`: Most successful strategy
+- `worst_mistake`: Biggest lesson learned
+
+### L1: Legacy
+**Retention**: Forever
+**Timeframe format**: `LEGACY` (single timeframe)
+**Content**: Cumulative trading history, core principles, generational wisdom
+
+**Use cases**:
+- Long-term philosophy
+- Foundational rules
+- Lessons that transcend market cycles
+
+**Example keys**:
+- `total_pnl`: All-time profit/loss
+- `years_traded`: Trading longevity
+- `avg_annual_pnl`: Long-term average return
+- `core_principles`: Immutable trading rules
+- `greatest_trades`: Hall of fame
+- `never_again`: Permanent warnings
+
+## Usage
+
+### Setting Context
+
+```python
+from src.context import ContextLayer, ContextStore
+from src.db import init_db
+
+conn = init_db("data/ouroboros.db")
+store = ContextStore(conn)
+
+# Store daily P&L
+store.set_context(
+    layer=ContextLayer.L6_DAILY,
+    timeframe="2026-02-04",
+    key="total_pnl",
+    value=1234.56
+)
+
+# Store weekly insight
+store.set_context(
+    layer=ContextLayer.L5_WEEKLY,
+    timeframe="2026-W06",
+    key="top_performers",
+    value=["005930", "000660", "035720"]  # JSON-serializable
+)
+
+# Store legacy wisdom
+store.set_context(
+    layer=ContextLayer.L1_LEGACY,
+    timeframe="LEGACY",
+    key="core_principles",
+    value=[
+        "Cut losses fast",
+        "Let winners run",
+        "Never average down on losing positions"
+    ]
+)
+```
+
+### Retrieving Context
+
+```python
+# Get a specific value
+pnl = store.get_context(ContextLayer.L6_DAILY, "2026-02-04", "total_pnl")
+# Returns: 1234.56
+
+# Get all keys for a timeframe
+daily_summary = store.get_all_contexts(ContextLayer.L6_DAILY, "2026-02-04")
+# Returns: {"total_pnl": 1234.56, "trade_count": 10, "win_rate": 60.0, ...}
+
+# Get all data for a layer (any timeframe)
+all_daily = store.get_all_contexts(ContextLayer.L6_DAILY)
+# Returns: {"total_pnl": 1234.56, "trade_count": 10, ...} (latest timeframes first)
+
+# Get the latest timeframe
+latest = store.get_latest_timeframe(ContextLayer.L6_DAILY)
+# Returns: "2026-02-04"
+```
+
+### Automatic Aggregation
+
+The `ContextAggregator` rolls up data from lower to higher layers:
+
+```python
+from src.context.aggregator import ContextAggregator
+
+aggregator = ContextAggregator(conn)
+
+# Aggregate daily metrics from trades
+aggregator.aggregate_daily_from_trades("2026-02-04")
+
+# Roll up weekly from daily
+aggregator.aggregate_weekly_from_daily("2026-W06")
+
+# Roll up all layers at once (bottom-up)
+aggregator.run_all_aggregations()
+```
+
+**Aggregation schedule** (recommended):
+- **L7 → L6**: Every midnight (daily rollup)
+- **L6 → L5**: Every Sunday (weekly rollup)
+- **L5 → L4**: First day of each month (monthly rollup)
+- **L4 → L3**: First day of quarter (quarterly rollup)
+- **L3 → L2**: January 1st (annual rollup)
+- **L2 → L1**: On demand (major milestones)
+
+### Context Cleanup
+
+Expired contexts are automatically deleted based on retention policies:
+
+```python
+# Manual cleanup
+deleted = store.cleanup_expired_contexts()
+# Returns: {ContextLayer.L7_REALTIME: 42, ContextLayer.L6_DAILY: 15, ...}
+```
+
+**Retention policies** (defined in `src/context/layer.py`):
+- L1: Forever
+- L2: 10 years
+- L3: 3 years
+- L4: 2 years
+- L5: 1 year
+- L6: 90 days
+- L7: 7 days
+
+## Integration with Gemini Brain
+
+The context tree provides hierarchical memory for decision-making:
+
+```python
+from src.brain.gemini_client import GeminiClient
+
+# Build prompt with multi-layer context
+def build_enhanced_prompt(stock_code: str, store: ContextStore) -> str:
+    # L7: Real-time data
+    current_price = store.get_context(ContextLayer.L7_REALTIME, "2026-02-04", f"live_price_{stock_code}")
+
+    # L6: Recent daily performance
+    yesterday_pnl = store.get_context(ContextLayer.L6_DAILY, "2026-02-03", "total_pnl")
+
+    # L5: Weekly trend
+    weekly_data = store.get_all_contexts(ContextLayer.L5_WEEKLY, "2026-W06")
+
+    # L1: Core principles
+    principles = store.get_context(ContextLayer.L1_LEGACY, "LEGACY", "core_principles")
+
+    return f"""
+    Analyze {stock_code} for trading decision.
+
+    Current price: {current_price}
+    Yesterday's P&L: {yesterday_pnl}
+    This week: {weekly_data}
+
+    Core principles:
+    {chr(10).join(f'- {p}' for p in principles)}
+
+    Decision (BUY/SELL/HOLD):
+    """
+```
+
+## Database Schema
+
+```sql
+-- Context storage
+CREATE TABLE contexts (
+    id INTEGER PRIMARY KEY AUTOINCREMENT,
+    layer TEXT NOT NULL,          -- L1_LEGACY, L2_ANNUAL, ..., L7_REALTIME
+    timeframe TEXT NOT NULL,      -- "LEGACY", "2026", "2026-Q1", "2026-02", "2026-W06", "2026-02-04"
+    key TEXT NOT NULL,            -- "total_pnl", "win_rate", "core_principles", etc.
+    value TEXT NOT NULL,          -- JSON-serialized value
+    created_at TEXT NOT NULL,     -- ISO 8601 timestamp
+    updated_at TEXT NOT NULL,     -- ISO 8601 timestamp
+    UNIQUE(layer, timeframe, key)
+);
+
+-- Layer metadata
+CREATE TABLE context_metadata (
+    layer TEXT PRIMARY KEY,
+    description TEXT NOT NULL,
+    retention_days INTEGER,       -- NULL = keep forever
+    aggregation_source TEXT       -- Parent layer for rollup
+);
+
+-- Indices for fast queries
+CREATE INDEX idx_contexts_layer ON contexts(layer);
+CREATE INDEX idx_contexts_timeframe ON contexts(timeframe);
+CREATE INDEX idx_contexts_updated ON contexts(updated_at);
+```
+
+## Best Practices
+
+1. **Write to leaf layers only** — Never manually write to L1-L5; let aggregation populate them
+2. **Aggregate regularly** — Schedule aggregation jobs to keep higher layers fresh
+3. **Query specific timeframes** — Use `get_context(layer, timeframe, key)` for precise retrieval
+4. **Clean up periodically** — Run `cleanup_expired_contexts()` weekly to free space
+5. **Preserve L1 forever** — Legacy wisdom should never expire
+6. **Use JSON-serializable values** — Store dicts, lists, strings, numbers (not custom objects)
+
+## Testing
+
+See `tests/test_context.py` for comprehensive test coverage (18 tests, 100% coverage on context modules).
+
+```bash
+pytest tests/test_context.py -v
+```
+
+## References
+
+- **Implementation**: `src/context/`
+  - `layer.py`: Layer definitions and metadata
+  - `store.py`: CRUD operations
+  - `aggregator.py`: Bottom-up aggregation logic
+- **Database**: `src/db.py` (table initialization)
+- **Tests**: `tests/test_context.py`
+- **Related**: Pillar 2 (Multi-layered Context Management)

docs/disaster_recovery.md

@@ -0,0 +1,348 @@
+# Disaster Recovery Guide
+
+Complete guide for backing up and restoring The Ouroboros trading system.
+
+## Table of Contents
+
+- [Backup Strategy](#backup-strategy)
+- [Creating Backups](#creating-backups)
+- [Restoring from Backup](#restoring-from-backup)
+- [Health Monitoring](#health-monitoring)
+- [Export Formats](#export-formats)
+- [RTO/RPO](#rtorpo)
+- [Testing Recovery](#testing-recovery)
+
+## Backup Strategy
+
+The system implements a 3-tier backup retention policy:
+
+| Policy | Frequency | Retention | Purpose |
+|--------|-----------|-----------|---------|
+| **Daily** | Every day | 30 days | Quick recovery from recent issues |
+| **Weekly** | Sunday | 1 year | Medium-term historical analysis |
+| **Monthly** | 1st of month | Forever | Long-term archival |
+
+### Storage Structure
+
+```
+data/backups/
+├── daily/          # Last 30 days
+├── weekly/         # Last 52 weeks
+└── monthly/        # Forever (cold storage)
+```
+
+## Creating Backups
+
+### Automated Backups (Recommended)
+
+Set up a cron job to run daily:
+
+```bash
+# Edit crontab
+crontab -e
+
+# Run backup at 2 AM every day
+0 2 * * * cd /path/to/The-Ouroboros && ./scripts/backup.sh >> logs/backup.log 2>&1
+```
+
+### Manual Backups
+
+```bash
+# Run backup script
+./scripts/backup.sh
+
+# Or use Python directly
+python3 -c "
+from pathlib import Path
+from src.backup.scheduler import BackupScheduler, BackupPolicy
+
+scheduler = BackupScheduler('data/trade_logs.db', Path('data/backups'))
+metadata = scheduler.create_backup(BackupPolicy.DAILY, verify=True)
+print(f'Backup created: {metadata.file_path}')
+"
+```
+
+### Export to Other Formats
+
+```bash
+python3 -c "
+from pathlib import Path
+from src.backup.exporter import BackupExporter, ExportFormat
+
+exporter = BackupExporter('data/trade_logs.db')
+results = exporter.export_all(
+    Path('exports'),
+    formats=[ExportFormat.JSON, ExportFormat.CSV],
+    compress=True
+)
+"
+```
+
+## Restoring from Backup
+
+### Interactive Restoration
+
+```bash
+./scripts/restore.sh
+```
+
+The script will:
+1. List available backups
+2. Ask you to select one
+3. Create a safety backup of current database
+4. Restore the selected backup
+5. Verify database integrity
+
+### Manual Restoration
+
+```python
+from pathlib import Path
+from src.backup.scheduler import BackupScheduler
+
+scheduler = BackupScheduler('data/trade_logs.db', Path('data/backups'))
+
+# List backups
+backups = scheduler.list_backups()
+for backup in backups:
+    print(f"{backup.timestamp}: {backup.file_path}")
+
+# Restore specific backup
+scheduler.restore_backup(backups[0], verify=True)
+```
+
+## Health Monitoring
+
+### Check System Health
+
+```python
+from pathlib import Path
+from src.backup.health_monitor import HealthMonitor
+
+monitor = HealthMonitor('data/trade_logs.db', Path('data/backups'))
+
+# Run all checks
+report = monitor.get_health_report()
+print(f"Overall status: {report['overall_status']}")
+
+# Individual checks
+checks = monitor.run_all_checks()
+for name, result in checks.items():
+    print(f"{name}: {result.status.value} - {result.message}")
+```
+
+### Health Checks
+
+The system monitors:
+
+- **Database Health**: Accessibility, integrity, size
+- **Disk Space**: Available storage (alerts if < 10 GB)
+- **Backup Recency**: Ensures backups are < 25 hours old
+
+### Health Status Levels
+
+- **HEALTHY**: All systems operational
+- **DEGRADED**: Warning condition (e.g., low disk space)
+- **UNHEALTHY**: Critical issue (e.g., database corrupted, no backups)
+
+## Export Formats
+
+### JSON (Human-Readable)
+
+```json
+{
+  "export_timestamp": "2024-01-15T10:30:00Z",
+  "record_count": 150,
+  "trades": [
+    {
+      "timestamp": "2024-01-15T09:00:00Z",
+      "stock_code": "005930",
+      "action": "BUY",
+      "quantity": 10,
+      "price": 70000.0,
+      "confidence": 85,
+      "rationale": "Strong momentum",
+      "pnl": 0.0
+    }
+  ]
+}
+```
+
+### CSV (Analysis Tools)
+
+Compatible with Excel, pandas, R:
+
+```csv
+timestamp,stock_code,action,quantity,price,confidence,rationale,pnl
+2024-01-15T09:00:00Z,005930,BUY,10,70000.0,85,Strong momentum,0.0
+```
+
+### Parquet (Big Data)
+
+Columnar format for Spark, DuckDB:
+
+```python
+import pandas as pd
+df = pd.read_parquet('exports/trades_20240115.parquet')
+```
+
+## RTO/RPO
+
+### Recovery Time Objective (RTO)
+
+**Target: < 5 minutes**
+
+Time to restore trading operations:
+1. Identify backup to restore (1 min)
+2. Run restore script (2 min)
+3. Verify database integrity (1 min)
+4. Restart trading system (1 min)
+
+### Recovery Point Objective (RPO)
+
+**Target: < 24 hours**
+
+Maximum acceptable data loss:
+- Daily backups ensure ≤ 24-hour data loss
+- For critical periods, run backups more frequently
+
+## Testing Recovery
+
+### Quarterly Recovery Test
+
+Perform full disaster recovery test every quarter:
+
+1. **Create test backup**
+   ```bash
+   ./scripts/backup.sh
+   ```
+
+2. **Simulate disaster** (use test database)
+   ```bash
+   cp data/trade_logs.db data/trade_logs_test.db
+   rm data/trade_logs_test.db  # Simulate data loss
+   ```
+
+3. **Restore from backup**
+   ```bash
+   DB_PATH=data/trade_logs_test.db ./scripts/restore.sh
+   ```
+
+4. **Verify data integrity**
+   ```python
+   import sqlite3
+   conn = sqlite3.connect('data/trade_logs_test.db')
+   cursor = conn.execute('SELECT COUNT(*) FROM trades')
+   print(f"Restored {cursor.fetchone()[0]} trades")
+   ```
+
+5. **Document results** in `logs/recovery_test_YYYYMMDD.md`
+
+### Backup Verification
+
+Always verify backups after creation:
+
+```python
+from pathlib import Path
+from src.backup.scheduler import BackupScheduler
+
+scheduler = BackupScheduler('data/trade_logs.db', Path('data/backups'))
+
+# Create and verify
+metadata = scheduler.create_backup(BackupPolicy.DAILY, verify=True)
+print(f"Checksum: {metadata.checksum}")  # Should not be None
+```
+
+## Emergency Procedures
+
+### Database Corrupted
+
+1. Stop trading system immediately
+2. Check most recent backup age: `ls -lht data/backups/daily/`
+3. Restore: `./scripts/restore.sh`
+4. Verify: Run health check
+5. Resume trading
+
+### Disk Full
+
+1. Check disk space: `df -h`
+2. Clean old backups: Run cleanup manually
+   ```python
+   from pathlib import Path
+   from src.backup.scheduler import BackupScheduler
+   scheduler = BackupScheduler('data/trade_logs.db', Path('data/backups'))
+   scheduler.cleanup_old_backups()
+   ```
+3. Consider archiving old monthly backups to external storage
+4. Increase disk space if needed
+
+### Lost All Backups
+
+If local backups are lost:
+1. Check if exports exist in `exports/` directory
+2. Reconstruct database from CSV/JSON exports
+3. If no exports: Check broker API for trade history
+4. Manual reconstruction as last resort
+
+## Best Practices
+
+1. **Test Restores Regularly**: Don't wait for disaster
+2. **Monitor Disk Space**: Set up alerts at 80% usage
+3. **Keep Multiple Generations**: Never delete all backups at once
+4. **Verify Checksums**: Always verify backup integrity
+5. **Document Changes**: Update this guide when backup strategy changes
+6. **Off-Site Storage**: Consider external backup for monthly archives
+
+## Troubleshooting
+
+### Backup Script Fails
+
+```bash
+# Check database file permissions
+ls -l data/trade_logs.db
+
+# Check disk space
+df -h data/
+
+# Run backup manually with debug
+python3 -c "
+import logging
+logging.basicConfig(level=logging.DEBUG)
+from pathlib import Path
+from src.backup.scheduler import BackupScheduler, BackupPolicy
+scheduler = BackupScheduler('data/trade_logs.db', Path('data/backups'))
+scheduler.create_backup(BackupPolicy.DAILY, verify=True)
+"
+```
+
+### Restore Fails Verification
+
+```bash
+# Check backup file integrity
+python3 -c "
+import sqlite3
+conn = sqlite3.connect('data/backups/daily/trade_logs_daily_20240115.db')
+cursor = conn.execute('PRAGMA integrity_check')
+print(cursor.fetchone()[0])
+"
+```
+
+### Health Check Fails
+
+```python
+from pathlib import Path
+from src.backup.health_monitor import HealthMonitor
+
+monitor = HealthMonitor('data/trade_logs.db', Path('data/backups'))
+
+# Check each component individually
+print("Database:", monitor.check_database_health())
+print("Disk Space:", monitor.check_disk_space())
+print("Backup Recency:", monitor.check_backup_recency())
+```
+
+## Contact
+
+For backup/recovery issues:
+- Check logs: `logs/backup.log`
+- Review health status: Run health monitor
+- Raise issue on GitHub if automated recovery fails

docs/live-trading-checklist.md

@@ -0,0 +1,131 @@
+# 실전 전환 체크리스트
+
+모의 거래(paper)에서 실전(live)으로 전환하기 전에 아래 항목을 **순서대로** 모두 확인하세요.
+
+---
+
+## 1. 사전 조건
+
+### 1-1. KIS OpenAPI 실전 계좌 준비
+- [ ] 한국투자증권 계좌 개설 완료 (일반 위탁 계좌)
+- [ ] OpenAPI 실전 사용 신청 (KIS 홈페이지 → Open API → 서비스 신청)
+- [ ] 실전용 APP_KEY / APP_SECRET 발급 완료
+- [ ] KIS_ACCOUNT_NO 형식 확인: `XXXXXXXX-XX` (8자리-2자리)
+
+### 1-2. 리스크 파라미터 검토
+- [ ] `CIRCUIT_BREAKER_PCT` 확인: 기본값 -3.0% (더 엄격하게 조정 권장)
+- [ ] `FAT_FINGER_PCT` 확인: 기본값 30.0% (1회 주문 최대 잔고 대비 %)
+- [ ] `CONFIDENCE_THRESHOLD` 확인: BEARISH ≥ 90, NEUTRAL ≥ 80, BULLISH ≥ 75
+- [ ] 초기 투자금 결정 및 해외 주식 운용 한도 설정
+
+### 1-3. 시스템 요건
+- [ ] 커버리지 80% 이상 유지 확인: `pytest --cov=src`
+- [ ] 타입 체크 통과: `mypy src/ --strict`
+- [ ] Lint 통과: `ruff check src/ tests/`
+
+---
+
+## 2. 환경 설정
+
+### 2-1. `.env` 파일 수정
+
+```bash
+# 1. KIS 실전 URL로 변경 (모의: openapivts 포트 29443)
+KIS_BASE_URL=https://openapi.koreainvestment.com:9443
+
+# 2. 실전 APP_KEY / APP_SECRET으로 교체
+KIS_APP_KEY=<실전_APP_KEY>
+KIS_APP_SECRET=<실전_APP_SECRET>
+KIS_ACCOUNT_NO=<실전_계좌번호>
+
+# 3. 모드를 live로 변경
+MODE=live
+
+# 4. PAPER_OVERSEAS_CASH 비활성화 (live 모드에선 무시되지만 명시적으로 0 설정)
+PAPER_OVERSEAS_CASH=0
+```
+
+> ⚠️ `KIS_BASE_URL` 포트 주의:
+> - **모의(VTS)**: `https://openapivts.koreainvestment.com:29443`
+> - **실전**: `https://openapi.koreainvestment.com:9443`
+
+### 2-2. TR_ID 자동 분기 확인
+
+아래 TR_ID는 `MODE` 값에 따라 코드에서 **자동으로 선택**됩니다.
+별도 설정 불필요하나, 문제 발생 시 아래 표를 참조하세요.
+
+| 구분 | 모의 TR_ID | 실전 TR_ID |
+|------|-----------|-----------|
+| 국내 잔고 조회 | `VTTC8434R` | `TTTC8434R` |
+| 국내 현금 매수 | `VTTC0012U` | `TTTC0012U` |
+| 국내 현금 매도 | `VTTC0011U` | `TTTC0011U` |
+| 해외 잔고 조회 | `VTTS3012R` | `TTTS3012R` |
+| 해외 매수 | `VTTT1002U` | `TTTT1002U` |
+| 해외 매도 | `VTTT1001U` | `TTTT1006U` |
+
+> **출처**: `docs/한국투자증권_오픈API_전체문서_20260221_030000.xlsx` (공식 문서 기준)
+
+---
+
+## 3. 최종 확인
+
+### 3-1. 실전 시작 전 점검
+- [ ] DB 백업 완료: `data/trade_logs.db` → `data/backups/`
+- [ ] Telegram 알림 설정 확인 (실전에서는 알림이 더욱 중요)
+- [ ] 소액으로 첫 거래 진행 후 TR_ID/계좌 정상 동작 확인
+
+### 3-2. 실행 명령
+
+```bash
+# 실전 모드로 실행
+python -m src.main --mode=live
+
+# 대시보드 함께 실행 (별도 터미널에서 모니터링)
+python -m src.main --mode=live --dashboard
+```
+
+### 3-3. 실전 시작 직후 확인 사항
+- [ ] 로그에 `MODE=live` 출력 확인
+- [ ] 첫 잔고 조회 성공 (ConnectionError 없음)
+- [ ] Telegram 알림 수신 확인 ("System started")
+- [ ] 첫 주문 후 KIS 앱에서 체결 내역 확인
+
+---
+
+## 4. 비상 정지 방법
+
+### 즉각 정지
+```bash
+# 터미널에서 Ctrl+C (정상 종료 트리거)
+# 또는 Telegram 봇 명령:
+/stop
+```
+
+### Circuit Breaker 발동 시
+- CB가 발동되면 자동으로 거래 중단 및 Telegram 알림 전송
+- CB 임계값: `CIRCUIT_BREAKER_PCT` (기본 -3.0%)
+- **임계값은 엄격하게만 조정 가능** (더 낮은 음수 값으로만 변경)
+
+---
+
+## 5. 롤백 절차
+
+실전 전환 후 문제 발생 시:
+
+```bash
+# 1. 즉시 .env에서 MODE=paper로 복원
+# 2. 재시작
+python -m src.main --mode=paper
+
+# 3. DB에서 최근 거래 확인
+sqlite3 data/trade_logs.db "SELECT * FROM trades ORDER BY id DESC LIMIT 20;"
+```
+
+---
+
+## 관련 문서
+
+- [시스템 아키텍처](architecture.md)
+- [워크플로우 가이드](workflow.md)
+- [재해 복구](disaster_recovery.md)
+- [Agent 제약 조건](agents.md)

docs/ouroboros/00_validation_system.md

@@ -0,0 +1,56 @@
+<!--
+Doc-ID: DOC-VAL-001
+Version: 1.0.0
+Status: active
+Owner: strategy
+Updated: 2026-02-26
+-->
+
+# 문서 검증 시스템
+
+본 문서는 문서 간 허위 내용, 수치 충돌, 구현 불가능 지시를 사전에 제거하기 위한 검증 규칙이다.
+
+## 검증 목표
+
+- 단일 진실원장 기준으로 모든 지시서의 수치/규칙 정합성 보장
+- 설계 문장과 코드 작업 지시 간 추적성 보장
+- 테스트 미정의 상태에서 구현 착수 금지
+
+## 불일치 유형 정의
+
+- `RULE-DOC-001`: 정의되지 않은 요구사항 ID 사용
+- `RULE-DOC-002`: 동일 요구사항 ID에 상충되는 값(예: 슬리피지 수치) 기술
+- `RULE-DOC-003`: 시간대 미표기 또는 KST/UTC 혼용 지시
+- `RULE-DOC-004`: 주문 정책과 리스크 정책 충돌(예: 저유동 세션 시장가 허용)
+- `RULE-DOC-005`: 구현 태스크에 테스트 ID 미연결
+- `RULE-DOC-006`: 문서 라우팅 링크 깨짐
+
+## 검증 파이프라인
+
+1. 정적 검사 (자동)
+- 대상: `docs/ouroboros/*.md`
+- 검사: 메타데이터, 링크 유효성, ID 정의/참조 일치, REQ-추적성 매핑
+- 도구: `scripts/validate_ouroboros_docs.py`
+
+2. 추적성 검사 (자동 + 수동)
+- 자동: `REQ-*`가 최소 1개 `TASK-*`와 1개 `TEST-*`에 연결되었는지 확인
+- 수동: 정책 충돌 후보를 PR 체크리스트로 검토
+
+3. 도메인 무결성 검사 (수동)
+- KIS 점검시간 회피, 주문 유형 강제, Kill Switch 순서, 환율 정책이 동시에 존재하는지 점검
+- 백테스트 체결가가 보수 가정인지 점검
+
+## 변경 통제 규칙
+
+- `REQ-*` 추가/수정 시 반드시 `01_requirements_registry.md` 먼저 변경
+- `TASK-*` 수정 시 반드시 `40_acceptance_and_test_plan.md`의 대응 테스트를 동시 수정
+- 충돌 발생 시 우선순위: `requirements_registry > phase execution > code work order`
+
+적용 룰셋:
+- `RULE-DOC-001` `RULE-DOC-002` `RULE-DOC-003` `RULE-DOC-004` `RULE-DOC-005` `RULE-DOC-006`
+
+## PR 게이트
+
+- `python3 scripts/validate_ouroboros_docs.py` 성공
+- 신규/변경 `REQ-*`가 테스트 기준(`TEST-*`)과 연결됨
+- 원본 계획(v2/v3)과 모순 없음

docs/ouroboros/01_requirements_registry.md

@@ -0,0 +1,39 @@
+<!--
+Doc-ID: DOC-REQ-001
+Version: 1.0.0
+Status: active
+Owner: strategy
+Updated: 2026-02-26
+-->
+
+# 요구사항 원장 (Single Source of Truth)
+
+이 문서의 ID가 계획/구현/테스트 전 문서에서 참조되는 유일한 요구사항 집합이다.
+
+## v2 핵심 요구사항
+
+- `REQ-V2-001`: 상태는 `HOLDING`, `BE_LOCK`, `ARMED`, `EXITED` 4단계여야 한다.
+- `REQ-V2-002`: 상태 전이는 매 틱/바 평가 시 최상위 상태로 즉시 승격되어야 한다.
+- `REQ-V2-003`: `EXITED` 조건은 모든 상태보다 우선 평가되어야 한다.
+- `REQ-V2-004`: 청산 로직은 Hard Stop, BE Lock, ATR Trailing, 모델 확률 보조 트리거를 포함해야 한다.
+- `REQ-V2-005`: 라벨링은 Triple Barrier(Upper/Lower/Time) 방식이어야 한다.
+- `REQ-V2-006`: 검증은 Walk-forward + Purge/Embargo를 강제한다.
+- `REQ-V2-007`: 백테스트는 비용/슬리피지/체결실패를 반영하지 않으면 채택 불가다.
+- `REQ-V2-008`: Kill Switch는 신규주문차단 -> 미체결취소 -> 재조회 -> 리스크축소 -> 스냅샷 순서다.
+
+## v3 핵심 요구사항
+
+- `REQ-V3-001`: 모든 신호/주문/로그는 `session_id`를 포함해야 한다.
+- `REQ-V3-002`: 세션 전환 시 리스크 파라미터 재로딩이 수행되어야 한다.
+- `REQ-V3-003`: 브로커 블랙아웃 시간대에는 신규 주문이 금지되어야 한다.
+- `REQ-V3-004`: 블랙아웃 중 신호는 Queue에 적재되고, 복구 후 유효성 재검증을 거친다.
+- `REQ-V3-005`: 저유동 세션(`NXT_AFTER`, `US_PRE`, `US_DAY`, `US_AFTER`)은 시장가 주문 금지다.
+- `REQ-V3-006`: 백테스트 체결가는 불리한 방향 체결 가정을 기본으로 한다.
+- `REQ-V3-007`: US 운용은 환율 손익 분리 추적과 통화 버퍼 정책을 포함해야 한다.
+- `REQ-V3-008`: 마감/오버나잇 규칙은 Kill Switch와 충돌 없이 연동되어야 한다.
+
+## 공통 운영 요구사항
+
+- `REQ-OPS-001`: 타임존은 모든 시간 필드에 명시(KST/UTC)되어야 한다.
+- `REQ-OPS-002`: 문서의 수치 정책은 원장에서만 변경한다.
+- `REQ-OPS-003`: 구현 태스크는 반드시 테스트 태스크를 동반한다.

docs/ouroboros/10_phase_v2_execution.md

@@ -0,0 +1,63 @@
+<!--
+Doc-ID: DOC-PHASE-V2-001
+Version: 1.0.0
+Status: active
+Owner: strategy
+Updated: 2026-02-26
+-->
+
+# v2 실행 지시서 (설계 -> 코드)
+
+참조 요구사항: `REQ-V2-001` `REQ-V2-002` `REQ-V2-003` `REQ-V2-004` `REQ-V2-005` `REQ-V2-006` `REQ-V2-007` `REQ-V2-008` `REQ-OPS-001` `REQ-OPS-002` `REQ-OPS-003`
+
+## 단계 1: 도메인 모델 확정
+
+- `TASK-V2-001`: 상태머신 enum/전이 이벤트/전이 사유 스키마 설계
+- `TASK-V2-002`: `position_state` 스냅샷 구조(현재상태, peak, stops, last_reason) 정의
+- `TASK-V2-003`: 청산 판단 입력 DTO(가격, ATR, pred_prob, liquidity_signal) 정의
+
+완료 기준:
+- 상태와 전이 사유가 로그/DB에서 재현 가능
+- `REQ-V2-001`~`003`을 코드 타입 수준에서 강제
+
+## 단계 2: 청산 엔진 구현
+
+- `TASK-V2-004`: 우선순위 기반 전이 함수 구현(`evaluate_exit_first` -> `promote_state`)
+- `TASK-V2-005`: Hard Stop/BE Lock/ATR Trailing 결합 로직 구현
+- `TASK-V2-006`: 모델 확률 신호를 보조 트리거로 결합(단독 청산 금지)
+
+완료 기준:
+- 갭 상황에서 다중 조건 동시 충족 시 최상위 상태로 단번 전이
+- `REQ-V2-004` 준수
+
+## 단계 3: 라벨링/학습 데이터 파이프라인
+
+- `TASK-V2-007`: Triple Barrier 라벨러 구현(장벽 선터치 우선)
+- `TASK-V2-008`: 피처 구간/라벨 구간 분리 검증 유틸 구현
+- `TASK-V2-009`: 라벨 생성 로그(진입시각, 터치장벽, 만기장벽) 기록
+
+완료 기준:
+- look-ahead 차단 증빙 로그 확보
+- `REQ-V2-005` 충족
+
+## 단계 4: 검증 프레임워크
+
+- `TASK-V2-010`: Walk-forward split + Purge/Embargo 분할기 구현
+- `TASK-V2-011`: 베이스라인(`B0`,`B1`,`M1`) 비교 리포트 포맷 구현
+- `TASK-V2-012`: 체결 비용/슬리피지/실패 반영 백테스트 옵션 강제
+
+완료 기준:
+- `REQ-V2-006`, `REQ-V2-007` 충족
+
+## 단계 5: Kill Switch 통합
+
+- `TASK-V2-013`: Kill Switch 순차 실행 오케스트레이터 구현 (`src/core/risk_manager.py` 수정 금지)
+- `TASK-V2-014`: 주문 차단 플래그/미체결 취소/재조회 재시도 로직 구현
+- `TASK-V2-015`: 스냅샷/알림/복구 진입 절차 구현
+
+완료 기준:
+- `REQ-V2-008` 순서 일치
+
+라우팅:
+- 코드 지시 상세: [30_code_level_work_orders.md](./30_code_level_work_orders.md)
+- 테스트 상세: [40_acceptance_and_test_plan.md](./40_acceptance_and_test_plan.md)

docs/ouroboros/20_phase_v3_execution.md

@@ -0,0 +1,60 @@
+<!--
+Doc-ID: DOC-PHASE-V3-001
+Version: 1.0.0
+Status: active
+Owner: strategy
+Updated: 2026-02-26
+-->
+
+# v3 실행 지시서 (세션 확장)
+
+참조 요구사항: `REQ-V3-001` `REQ-V3-002` `REQ-V3-003` `REQ-V3-004` `REQ-V3-005` `REQ-V3-006` `REQ-V3-007` `REQ-V3-008` `REQ-OPS-001` `REQ-OPS-002` `REQ-OPS-003`
+
+## 단계 1: 세션 엔진
+
+- `TASK-V3-001`: `session_id` 분류기 구현(KR/US 확장 세션)
+- `TASK-V3-002`: 세션 전환 훅에서 리스크 파라미터 재로딩 구현
+- `TASK-V3-003`: 로그/DB 스키마에 `session_id` 필드 강제
+
+완료 기준:
+- `REQ-V3-001`, `REQ-V3-002` 충족
+
+## 단계 2: 블랙아웃/복구 제어
+
+- `TASK-V3-004`: 블랙아웃 윈도우 정책 로더 구현(설정 기반)
+- `TASK-V3-005`: 블랙아웃 중 신규 주문 차단 + 의도 큐 적재 구현
+- `TASK-V3-006`: 복구 시 동기화(잔고/미체결/체결) 후 큐 재검증 실행
+
+완료 기준:
+- `REQ-V3-003`, `REQ-V3-004` 충족
+
+## 단계 3: 주문 정책 강화
+
+- `TASK-V3-007`: 세션별 주문 타입 매트릭스 구현
+- `TASK-V3-008`: 저유동 세션 시장가 주문 하드 차단
+- `TASK-V3-009`: 재호가 간격/횟수 제한 및 주문 철회 조건 구현
+
+완료 기준:
+- `REQ-V3-005` 충족
+
+## 단계 4: 비용/체결 모델 정교화
+
+- `TASK-V3-010`: 세션별 슬리피지/비용 테이블 엔진 반영
+- `TASK-V3-011`: 불리한 체결 가정(상대 호가 방향) 체결기 구현
+- `TASK-V3-012`: 시나리오별 체결 실패/부분체결 모델 반영
+
+완료 기준:
+- `REQ-V3-006` 충족
+
+## 단계 5: 환율/오버나잇/Kill Switch 연동
+
+- `TASK-V3-013`: 전략 PnL과 FX PnL 분리 회계 구현
+- `TASK-V3-014`: USD/KRW 버퍼 규칙 위반 시 신규 진입 제한 구현
+- `TASK-V3-015`: 오버나잇 예외와 Kill Switch 우선순위 통합
+
+완료 기준:
+- `REQ-V3-007`, `REQ-V3-008` 충족
+
+라우팅:
+- 코드 지시 상세: [30_code_level_work_orders.md](./30_code_level_work_orders.md)
+- 테스트 상세: [40_acceptance_and_test_plan.md](./40_acceptance_and_test_plan.md)

docs/ouroboros/30_code_level_work_orders.md

@@ -0,0 +1,59 @@
+<!--
+Doc-ID: DOC-CODE-001
+Version: 1.0.0
+Status: active
+Owner: strategy
+Updated: 2026-02-26
+-->
+
+# 코드 레벨 작업 지시서
+
+본 문서는 파일 단위 구현 지시서다. 모든 작업은 요구사항 ID와 테스트 ID를 포함해야 한다.
+
+제약:
+- `src/core/risk_manager.py`는 READ-ONLY로 간주하고 수정하지 않는다.
+- Kill Switch는 별도 모듈(예: `src/core/kill_switch.py`)로 추가하고 상위 실행 루프에서 연동한다.
+
+## 구현 단위 A: 상태기계/청산
+
+- `TASK-CODE-001` (`REQ-V2-001`,`REQ-V2-002`,`REQ-V2-003`): `src/strategy/`에 상태기계 모듈 추가
+- `TASK-CODE-002` (`REQ-V2-004`): ATR/BE/Hard Stop 결합 청산 함수 추가
+- `TASK-CODE-003` (`REQ-V2-008`): Kill Switch 오케스트레이터를 `src/core/kill_switch.py`에 추가
+- `TEST-CODE-001`: 갭 점프 시 최고상태 승격 테스트
+- `TEST-CODE-002`: EXIT 우선순위 테스트
+
+## 구현 단위 B: 라벨링/검증
+
+- `TASK-CODE-004` (`REQ-V2-005`): Triple Barrier 라벨러 모듈 추가(`src/analysis/` 또는 `src/strategy/`)
+- `TASK-CODE-005` (`REQ-V2-006`): Walk-forward + Purge/Embargo 분할 유틸 추가
+- `TASK-CODE-006` (`REQ-V2-007`): 백테스트 실행기에서 비용/슬리피지 옵션 필수화
+- `TEST-CODE-003`: 라벨 선터치 우선 테스트
+- `TEST-CODE-004`: 누수 차단 테스트
+
+## 구현 단위 C: 세션/주문 정책
+
+- `TASK-CODE-007` (`REQ-V3-001`,`REQ-V3-002`): 세션 분류/전환 훅을 `src/markets/schedule.py` 연동
+- `TASK-CODE-008` (`REQ-V3-003`,`REQ-V3-004`): 블랙아웃 큐 처리기를 `src/broker/`에 추가
+- `TASK-CODE-009` (`REQ-V3-005`): 세션별 주문 타입 검증기 추가
+- `TEST-CODE-005`: 블랙아웃 신규주문 차단 테스트
+- `TEST-CODE-006`: 저유동 세션 시장가 거부 테스트
+
+## 구현 단위 D: 체결/환율/오버나잇
+
+- `TASK-CODE-010` (`REQ-V3-006`): 불리한 체결가 모델을 백테스트 체결기로 구현
+- `TASK-CODE-011` (`REQ-V3-007`): FX PnL 분리 회계 테이블/컬럼 추가
+- `TASK-CODE-012` (`REQ-V3-008`): 오버나잇 예외와 Kill Switch 충돌 해소 로직 구현
+- `TEST-CODE-007`: 불리한 체결가 모델 테스트
+- `TEST-CODE-008`: FX 버퍼 위반 시 신규진입 제한 테스트
+
+## 구현 단위 E: 운영/문서 거버넌스
+
+- `TASK-OPS-001` (`REQ-OPS-001`): 시간 필드/로그 스키마의 타임존 표기 강제 규칙 구현
+- `TASK-OPS-002` (`REQ-OPS-002`): 정책 수치 변경 시 `01_requirements_registry.md` 선수정 CI 체크 추가
+- `TASK-OPS-003` (`REQ-OPS-003`): `TASK-*` 없는 `REQ-*` 또는 `TEST-*` 없는 `REQ-*`를 차단하는 문서 검증 게이트 유지
+
+## 커밋 규칙
+
+- 커밋 메시지에 `TASK-*` 포함
+- PR 본문에 `REQ-*`, `TEST-*` 매핑 표 포함
+- 변경 파일마다 최소 1개 테스트 연결

docs/ouroboros/40_acceptance_and_test_plan.md

@@ -0,0 +1,63 @@
+<!--
+Doc-ID: DOC-TEST-001
+Version: 1.0.0
+Status: active
+Owner: strategy
+Updated: 2026-02-26
+-->
+
+# 수용 기준 및 테스트 계획
+
+## 수용 기준
+
+- `TEST-ACC-000` (`REQ-V2-001`): 상태 enum은 4개(`HOLDING`,`BE_LOCK`,`ARMED`,`EXITED`)만 허용한다.
+- `TEST-ACC-001` (`REQ-V2-002`): 상태 전이는 순차 if-else가 아닌 우선순위 승격으로 동작한다.
+- `TEST-ACC-010` (`REQ-V2-003`): `EXITED` 조건은 어떤 상태보다 먼저 평가된다.
+- `TEST-ACC-011` (`REQ-V2-004`): 청산 판단은 Hard Stop/BE Lock/ATR/모델보조 4요소를 모두 포함한다.
+- `TEST-ACC-012` (`REQ-V2-005`): Triple Barrier 라벨은 first-touch 규칙으로 결정된다.
+- `TEST-ACC-013` (`REQ-V2-006`): 학습/검증 분할은 Walk-forward + Purge/Embargo를 적용한다.
+- `TEST-ACC-014` (`REQ-V2-007`): 비용/슬리피지/체결실패 옵션 비활성 시 백테스트 실행을 거부한다.
+- `TEST-ACC-002` (`REQ-V2-008`): Kill Switch 실행 순서가 고정 순서를 위반하지 않는다.
+- `TEST-ACC-015` (`REQ-V3-001`): 모든 주문/로그 레코드에 `session_id`가 저장된다.
+- `TEST-ACC-016` (`REQ-V3-002`): 세션 전환 이벤트 시 리스크 파라미터가 재로딩된다.
+- `TEST-ACC-003` (`REQ-V3-003`): 블랙아웃 중 신규 주문 API 호출이 발생하지 않는다.
+- `TEST-ACC-017` (`REQ-V3-004`): 블랙아웃 큐는 복구 후 재검증을 통과한 주문만 실행한다.
+- `TEST-ACC-004` (`REQ-V3-005`): 저유동 세션 시장가 주문은 항상 거부된다.
+- `TEST-ACC-005` (`REQ-V3-006`): 백테스트 체결가가 단순 종가 체결보다 보수적 손익을 낸다.
+- `TEST-ACC-006` (`REQ-V3-007`): 전략 손익과 환율 손익이 별도 집계된다.
+- `TEST-ACC-018` (`REQ-V3-008`): 오버나잇 예외 상태에서도 Kill Switch 우선순위가 유지된다.
+- `TEST-ACC-007` (`REQ-OPS-001`): 시간 관련 필드는 타임존(KST/UTC)이 누락되면 검증 실패한다.
+- `TEST-ACC-008` (`REQ-OPS-002`): 정책 수치 변경이 원장 미반영이면 검증 실패한다.
+- `TEST-ACC-009` (`REQ-OPS-003`): `REQ-*`가 `TASK-*`/`TEST-*` 매핑 없이 존재하면 검증 실패한다.
+
+## 테스트 계층
+
+1. 단위 테스트
+- 상태 전이, 주문타입 검증, 큐 복구 로직, 체결가 모델
+
+2. 통합 테스트
+- 세션 전환 -> 주문 정책 -> 리스크 엔진 연동
+- 블랙아웃 시작/해제 이벤트 연동
+
+3. 회귀 테스트
+- 기존 `tests/` 스위트 전량 실행
+- 신규 기능 플래그 ON/OFF 비교
+
+4. 구동/모니터링 검증 (필수)
+- 개발 완료 후 시스템을 실제 구동해 핵심 경로를 관찰
+- 필수 관찰 항목: 주문 차단 정책, Kill Switch 동작, 경보/예외 로그, 세션 전환 로그
+- Runtime Verifier 코멘트로 증적(실행 명령/요약 로그) 첨부
+
+## 실행 명령
+
+```bash
+pytest -q
+python3 scripts/validate_ouroboros_docs.py
+```
+
+## 실패 처리 규칙
+
+- 문서 검증 실패 시 구현 PR 병합 금지
+- `REQ-*` 변경 후 테스트 매핑 누락 시 병합 금지
+- 회귀 실패 시 원인 모듈 분리 후 재검증
+- 구동/모니터링 증적 누락 시 검증 승인 금지

docs/ouroboros/50_scenario_matrix_and_issue_taxonomy.md

@@ -0,0 +1,68 @@
+<!--
+Doc-ID: DOC-PM-001
+Version: 1.0.0
+Status: active
+Owner: strategy
+Updated: 2026-02-26
+-->
+
+# 실전 시나리오 매트릭스 + 이슈 분류 체계
+
+목표: 운영에서 바로 사용할 수 있는 형태로 Happy Path / Failure Path / Ops Incident를 추적 가능한 ID 체계(`REQ-*`, `TASK-*`, `TEST-*`)에 매핑한다.
+
+## 1) 시나리오 매트릭스
+
+| Scenario ID | Type | Trigger | Expected System Behavior | Primary IDs (REQ/TASK/TEST) | Ticket Priority |
+|---|---|---|---|---|---|
+| `SCN-HAPPY-001` | Happy Path | KR 정규 세션에서 진입 신호 발생, 블랙아웃 아님 | 주문/로그에 `session_id` 저장 후 정책에 맞는 주문 전송 | `REQ-V3-001`, `TASK-V3-001`, `TASK-V3-003`, `TEST-ACC-015` | P1 |
+| `SCN-HAPPY-002` | Happy Path | 보유 포지션에서 BE/ATR/Hard Stop 조건 순차 도달 | 상태가 즉시 상위 단계로 승격, `EXITED` 우선 평가 보장 | `REQ-V2-002`, `REQ-V2-003`, `TASK-V2-004`, `TEST-ACC-001`, `TEST-ACC-010` | P0 |
+| `SCN-HAPPY-003` | Happy Path | 세션 전환(KR->US) 이벤트 발생 | 리스크 파라미터 자동 재로딩, 새 세션 정책으로 즉시 전환 | `REQ-V3-002`, `TASK-V3-002`, `TEST-ACC-016` | P0 |
+| `SCN-HAPPY-004` | Happy Path | 백테스트 실행 요청 | 비용/슬리피지/체결실패 옵션 누락 시 실행 거부, 포함 시 실행 | `REQ-V2-007`, `TASK-V2-012`, `TEST-ACC-014` | P1 |
+| `SCN-FAIL-001` | Failure Path | 블랙아웃 중 신규 주문 신호 발생 | 신규 주문 차단 + 주문 의도 큐 적재, API 직접 호출 금지 | `REQ-V3-003`, `REQ-V3-004`, `TASK-V3-005`, `TEST-ACC-003`, `TEST-ACC-017` | P0 |
+| `SCN-FAIL-002` | Failure Path | 저유동 세션에 시장가 주문 요청 | 시장가 하드 거부, 지정가 대체 또는 주문 취소 | `REQ-V3-005`, `TASK-V3-007`, `TASK-V3-008`, `TEST-ACC-004` | P0 |
+| `SCN-FAIL-003` | Failure Path | Kill Switch 트리거(손실/연결/리스크 한도) | 신규주문차단->미체결취소->재조회->리스크축소->스냅샷 순서 강제 | `REQ-V2-008`, `TASK-V2-013`, `TEST-ACC-002` | P0 |
+| `SCN-FAIL-004` | Failure Path | FX 버퍼 부족 상태에서 US 진입 신호 | 전략 PnL/FX PnL 분리 집계 유지, 신규 진입 제한 | `REQ-V3-007`, `TASK-V3-013`, `TASK-V3-014`, `TEST-ACC-006` | P1 |
+| `SCN-OPS-001` | Ops Incident | 브로커 점검/블랙아웃 종료 직후 | 잔고/미체결/체결 동기화 후 큐 재검증 통과 주문만 집행 | `REQ-V3-004`, `TASK-V3-006`, `TEST-ACC-017` | P0 |
+| `SCN-OPS-002` | Ops Incident | 정책 수치가 코드에만 반영되고 원장 미수정 | 문서 검증에서 실패 처리, PR 병합 차단 | `REQ-OPS-002`, `TASK-OPS-002`, `TEST-ACC-008` | P0 |
+| `SCN-OPS-003` | Ops Incident | 타임존 누락 로그/스케줄 데이터 유입 | KST/UTC 미표기 레코드 검증 실패 처리 | `REQ-OPS-001`, `TASK-OPS-001`, `TEST-ACC-007` | P1 |
+| `SCN-OPS-004` | Ops Incident | 신규 REQ 추가 후 TASK/TEST 누락 | 추적성 게이트 실패, 구현 PR 병합 차단 | `REQ-OPS-003`, `TASK-OPS-003`, `TEST-ACC-009` | P0 |
+| `SCN-OPS-005` | Ops Incident | 배포 후 런타임 이상 동작(주문오류/상태전이오류/정책위반) 탐지 | Runtime Verifier가 즉시 이슈 발행, Dev 수정 후 재관측으로 클로즈 판정 | `REQ-V2-008`, `REQ-V3-003`, `REQ-V3-005`, `TEST-ACC-002`, `TEST-ACC-003`, `TEST-ACC-004` | P0 |
+
+## 2) 이슈 분류 체계 (Issue Taxonomy)
+
+| Taxonomy | Definition | Typical Symptoms | Default Owner | Mapping Baseline |
+|---|---|---|---|---|
+| `EXEC-STATE` | 상태기계/청산 우선순위 위반 | EXIT 우선순위 깨짐, 상태 역행, 갭 대응 실패 | Strategy | `REQ-V2-001`~`REQ-V2-004`, `TASK-V2-004`~`TASK-V2-006`, `TEST-ACC-000`,`001`,`010`,`011` |
+| `EXEC-POLICY` | 세션/주문 정책 위반 | 블랙아웃 주문 전송, 저유동 시장가 허용 | Broker/Execution | `REQ-V3-003`~`REQ-V3-005`, `TASK-V3-004`~`TASK-V3-009`, `TEST-ACC-003`,`004`,`017` |
+| `BACKTEST-MODEL` | 백테스트 현실성/검증 무결성 위반 | 비용 옵션 off로 실행, 체결가 과낙관 | Research | `REQ-V2-006`,`REQ-V2-007`,`REQ-V3-006`, `TASK-V2-010`~`012`, `TASK-V3-010`~`012`, `TEST-ACC-013`,`014`,`005` |
+| `RISK-EMERGENCY` | Kill Switch/리스크 비상 대응 실패 | 순서 위반, 차단 누락, 복구 절차 누락 | Risk | `REQ-V2-008`,`REQ-V3-008`, `TASK-V2-013`~`015`, `TASK-V3-015`, `TEST-ACC-002`,`018` |
+| `FX-ACCOUNTING` | 환율/통화 버퍼 정책 위반 | 전략손익/환차손익 혼합 집계, 버퍼 미적용 | Risk + Data | `REQ-V3-007`, `TASK-V3-013`,`014`, `TEST-ACC-006` |
+| `OPS-GOVERNANCE` | 문서/추적성/타임존 거버넌스 위반 | 원장 미수정, TEST 누락, 타임존 미표기 | PM + QA | `REQ-OPS-001`~`003`, `TASK-OPS-001`~`003`, `TEST-ACC-007`~`009` |
+| `RUNTIME-VERIFY` | 실동작 모니터링 검증 | 배포 후 이상 현상, 간헐 오류, 테스트 미포착 회귀 | Runtime Verifier + TPM | 관련 `REQ/TASK/TEST`와 런타임 로그 증적 필수 |
+
+## 3) 티켓 생성 규칙 (Implementable)
+
+1. 모든 이슈는 `taxonomy + scenario_id`를 제목에 포함한다.  
+   예: `[EXEC-POLICY][SCN-FAIL-001] blackout 주문 차단 누락`
+2. 본문 필수 항목: 재현절차, 기대결과, 실제결과, 영향범위, 롤백/완화책.
+3. 본문에 최소 1개 `REQ-*`, 1개 `TASK-*`, 1개 `TEST-*`를 명시한다.
+4. 우선순위 기준:
+- P0: 실주문 위험, Kill Switch, 블랙아웃/시장가 정책, 추적성 게이트 실패
+- P1: 손익 왜곡 가능성(체결/FX/시간대), 운영 리스크 증가
+- P2: 보고서/관측성 품질 이슈(거래 안전성 영향 없음)
+5. Runtime Verifier가 발행한 `RUNTIME-VERIFY` 이슈는 Main Agent 확인 전 클로즈 금지.
+
+## 4) 즉시 생성 권장 티켓 (초기 백로그)
+
+- `TKT-P0-001`: `[EXEC-POLICY][SCN-FAIL-001]` 블랙아웃 차단 + 큐적재 + 복구 재검증 e2e 점검 (`REQ-V3-003`,`REQ-V3-004`)
+- `TKT-P0-002`: `[RISK-EMERGENCY][SCN-FAIL-003]` Kill Switch 순서 강제 검증 자동화 (`REQ-V2-008`)
+- `TKT-P0-003`: `[OPS-GOVERNANCE][SCN-OPS-004]` REQ/TASK/TEST 누락 시 PR 차단 게이트 상시 점검 (`REQ-OPS-003`)
+- `TKT-P1-001`: `[FX-ACCOUNTING][SCN-FAIL-004]` FX 버퍼 위반 시 진입 제한 회귀 케이스 보강 (`REQ-V3-007`)
+- `TKT-P1-002`: `[BACKTEST-MODEL][SCN-HAPPY-004]` 비용/슬리피지 미설정 백테스트 거부 UX 명확화 (`REQ-V2-007`)
+- `TKT-P0-004`: `[RUNTIME-VERIFY][SCN-OPS-005]` 배포 후 런타임 이상 탐지/재현/클로즈 판정 절차 자동화
+
+## 5) 운영 체크포인트
+
+- 스프린트 계획 시 `P0` 시나리오 100% 테스트 통과를 출발 조건으로 둔다.
+- 배포 승인 시 `SCN-FAIL-*`, `SCN-OPS-*` 관련 `TEST-ACC-*`를 우선 확인한다.
+- 정책 변경 PR은 반드시 원장(`01_requirements_registry.md`) 선수정 후 진행한다.

docs/ouroboros/50_tpm_control_protocol.md

@@ -0,0 +1,201 @@
+<!--
+Doc-ID: DOC-TPM-001
+Version: 1.0.0
+Status: active
+Owner: tpm
+Updated: 2026-02-26
+-->
+
+# TPM Control Protocol (Main <-> PM <-> TPM <-> Dev <-> Verifier <-> Runtime Verifier)
+
+목적:
+- PM 시나리오가 구현 가능한 단위로 분해되고, 개발/검증이 동일 ID 체계(`REQ-*`, `TASK-*`, `TEST-*`)로 닫히도록 강제한다.
+- 각 단계는 Entry/Exit gate를 통과해야 다음 단계로 이동 가능하다.
+- 주요 의사결정 포인트마다 Main Agent의 승인/의견 확인을 강제한다.
+
+## Team Roles
+
+- Main Agent: 최종 취합/우선순위/승인 게이트 오너
+- PM Agent: 시나리오/요구사항/티켓 관리
+- TPM Agent: PM-Dev-검증 간 구현 가능성/달성률 통제, 티켓 등록 및 구현 우선순위 지정 오너
+- Dev Agent: 구현 수행, 블로커 발생 시 재계획 요청
+- Verifier Agent: 문서/코드/테스트 산출물 검증
+- Runtime Verifier Agent: 실제 동작 모니터링, 이상 징후 이슈 발행, 수정 후 이슈 클로즈 판정
+
+Main Agent 아이디에이션 책임:
+- 진행 중 신규 구현 아이디어를 별도 문서에 누적 기록한다.
+- 기록 위치: [70_main_agent_ideation.md](./70_main_agent_ideation.md)
+- 각 항목은 `IDEA-*` 식별자, 배경, 기대효과, 리스크, 후속 티켓 후보를 포함해야 한다.
+
+## Main Decision Checkpoints (Mandatory)
+
+- DCP-01 범위 확정: Phase 0 종료 전 Main Agent 승인 필수
+- DCP-02 요구사항 확정: Phase 1 종료 전 Main Agent 승인 필수
+- DCP-03 구현 착수: Phase 2 종료 전 Main Agent 승인 필수
+- DCP-04 배포 승인: Phase 4 종료 후 Main Agent 최종 승인 필수
+
+## Phase Control Gates
+
+### Phase 0: Scenario Intake and Scope Lock
+
+Entry criteria:
+- PM 시나리오가 사용자 가치, 실패 모드, 우선순위를 포함해 제출됨
+- 영향 범위(모듈/세션/KR-US 시장)가 명시됨
+
+Exit criteria:
+- 시나리오가 `REQ-*` 후보에 1:1 또는 1:N 매핑됨
+- 모호한 표현("개선", "최적화")은 측정 가능한 조건으로 치환됨
+- 비범위 항목(out-of-scope) 명시
+
+Control checks:
+- PM/TPM 합의 완료
+- Main Agent 승인(DCP-01)
+- 산출물: 시나리오 카드, 초기 매핑 메모
+
+### Phase 1: Requirement Registry Gate
+
+Entry criteria:
+- Phase 0 산출물 승인
+- 변경 대상 요구사항 문서 식별 완료
+
+Exit criteria:
+- [01_requirements_registry.md](./01_requirements_registry.md)에 `REQ-*` 정의/수정 반영
+- 각 `REQ-*`가 최소 1개 `TASK-*`, 1개 `TEST-*`와 연결 가능 상태
+- 시간/정책 수치는 원장 단일 소스로 확정(`REQ-OPS-001`,`REQ-OPS-002`)
+
+Control checks:
+- `python3 scripts/validate_ouroboros_docs.py` 통과
+- Main Agent 승인(DCP-02)
+- 산출물: 업데이트된 요구사항 원장
+
+### Phase 2: Design and Work-Order Gate
+
+Entry criteria:
+- 요구사항 원장 갱신 완료
+- 영향 모듈 분석 완료(상태기계, 주문정책, 백테스트, 세션)
+
+Exit criteria:
+- [10_phase_v2_execution.md](./10_phase_v2_execution.md), [20_phase_v3_execution.md](./20_phase_v3_execution.md), [30_code_level_work_orders.md](./30_code_level_work_orders.md)에 작업 분해 완료
+- 각 작업은 구현 위치/제약/완료 조건을 가짐
+- 위험 작업(Kill Switch, blackout, session transition)은 별도 롤백 절차 포함
+
+Control checks:
+- TPM이 `REQ -> TASK` 누락 여부 검토
+- Main Agent 승인(DCP-03)
+- 산출물: 승인된 Work Order 세트
+
+### Phase 3: Implementation Gate
+
+Entry criteria:
+- 승인된 `TASK-*`가 브랜치 작업 단위로 분리됨
+- 변경 범위별 테스트 계획이 PR 본문에 링크됨
+
+Exit criteria:
+- 코드 변경이 `TASK-*`에 대응되어 추적 가능
+- 제약 준수(`src/core/risk_manager.py` 직접 수정 금지 등) 확인
+- 신규 로직마다 최소 1개 테스트 추가 또는 기존 테스트 확장
+
+Control checks:
+- PR 템플릿 내 `REQ-*`/`TASK-*`/`TEST-*` 매핑 확인
+- 산출물: 리뷰 가능한 PR
+
+### Phase 4: Verification and Acceptance Gate
+
+Entry criteria:
+- 구현 PR ready 상태
+- 테스트 케이스/픽스처 준비 완료
+
+Exit criteria:
+- [40_acceptance_and_test_plan.md](./40_acceptance_and_test_plan.md)의 해당 `TEST-ACC-*` 전부 통과
+- 회귀 테스트 통과(`pytest -q`)
+- 문서 검증 통과(`python3 scripts/validate_ouroboros_docs.py`)
+
+Control checks:
+- Verifier가 테스트 증적(로그/리포트/실행 커맨드) 첨부
+- Runtime Verifier가 스테이징/실운영 모니터링 계획 승인
+- 산출물: 수용 승인 레코드
+
+### Phase 5: Release and Post-Release Control
+
+Entry criteria:
+- Phase 4 승인
+- 운영 체크리스트 준비(세션 전환, 블랙아웃, Kill Switch)
+
+Exit criteria:
+- 배포 후 초기 관찰 윈도우에서 치명 경보 없음
+- 신규 시나리오/회귀 이슈는 다음 Cycle의 Phase 0 입력으로 환류
+- 요구사항/테스트 문서 버전 동기화 완료
+
+Control checks:
+- PM/TPM/Dev 3자 종료 확인
+- Runtime Verifier가 운영 모니터링 이슈 상태(신규/진행/해결)를 리포트
+- Main Agent 최종 승인(DCP-04)
+- 산출물: 릴리즈 노트 + 후속 액션 목록
+
+## Replan Protocol (Dev -> TPM)
+
+- 트리거:
+  - 구현 불가능(기술적 제약/외부 API 제약)
+  - 예상 대비 개발 리소스 과다(공수/인력/의존성 급증)
+- 절차:
+  1) Dev Agent가 `REPLAN-REQUEST` 발행(영향 REQ/TASK, 원인, 대안, 추가 공수 포함)
+  2) TPM Agent가 1차 심사(범위 축소/단계 분할/요구사항 조정안)
+  3) Verifier/PM 의견 수렴 후 Main Agent 승인으로 재계획 확정
+- 규칙:
+  - Main Agent 승인 없는 재계획은 실행 금지
+  - 재계획 반영 시 문서(`REQ/TASK/TEST`) 동시 갱신 필수
+
+TPM 티켓 운영 규칙:
+- TPM은 합의된 변경을 이슈로 등록하고 우선순위(`P0/P1/P2`)를 지정한다.
+- PR 본문에는 TPM이 지정한 우선순위와 범위가 그대로 반영되어야 한다.
+- 우선순위 변경은 TPM 제안 + Main Agent 승인으로만 가능하다.
+- PM/TPM/Dev/Reviewer/Verifier/Runtime Verifier는 주요 의사결정 시점마다 PR 코멘트를 남겨 결정 근거를 추적 가능 상태로 유지한다.
+
+브랜치 운영 규칙:
+- TPM은 각 티켓에 대해 `ticket temp branch -> program feature branch` PR 경로를 지정한다.
+- 티켓 머지 대상은 항상 program feature branch이며, `main`은 최종 통합 단계에서만 사용한다.
+
+## Runtime Verification Protocol
+
+- Runtime Verifier는 테스트 통과 이후 실제 동작(스테이징/실운영)을 모니터링한다.
+- 이상 동작/현상 발견 시 즉시 이슈 발행:
+  - 제목 규칙: `[RUNTIME-VERIFY][SCN-*] ...`
+  - 본문 필수: 재현조건, 관측 로그, 영향 범위, 임시 완화책, 관련 `REQ/TASK/TEST`
+- 이슈 클로즈 규칙:
+  - Dev 수정 완료 + Verifier 재검증 통과 + Runtime Verifier 재관측 정상
+  - 최종 클로즈 승인자는 Main Agent
+- 개발 완료 필수 절차:
+  - 시스템 실제 구동(스테이징/로컬 실운영 모드) 실행
+  - 모니터링 체크리스트(핵심 경보/주문 경로/예외 로그) 수행
+  - 결과를 티켓/PR 코멘트에 증적으로 첨부하지 않으면 완료로 간주하지 않음
+
+## Server Reflection Rule
+
+- `ticket temp branch -> program feature branch` 머지는 검증 승인 후 자동/수동 진행 가능하다.
+- `program feature branch -> main` 머지는 사용자 명시 승인 시에만 허용한다.
+- Main 병합 시 Main Agent가 승인 근거를 PR 코멘트에 기록한다.
+
+## Acceptance Matrix (PM Scenario -> Dev Tasks -> Verifier Checks)
+
+| PM Scenario | Requirement Coverage | Dev Tasks (Primary) | Verifier Checks (Must Pass) |
+|---|---|---|---|
+| 갭 급락/급등에서 청산 우선 처리 필요 | `REQ-V2-001`,`REQ-V2-002`,`REQ-V2-003` | `TASK-V2-004`,`TASK-CODE-001` | `TEST-ACC-000`,`TEST-ACC-001`,`TEST-ACC-010`,`TEST-CODE-001`,`TEST-CODE-002` |
+| 하드스탑 + BE락 + ATR + 모델보조를 한 엔진으로 통합 | `REQ-V2-004` | `TASK-V2-005`,`TASK-V2-006`,`TASK-CODE-002` | `TEST-ACC-011` |
+| 라벨 누수 없는 학습데이터 생성 | `REQ-V2-005` | `TASK-V2-007`,`TASK-CODE-004` | `TEST-ACC-012`,`TEST-CODE-003` |
+| 검증 프레임워크를 시계열 누수 방지 구조로 강제 | `REQ-V2-006` | `TASK-V2-010`,`TASK-CODE-005` | `TEST-ACC-013`,`TEST-CODE-004` |
+| 과낙관 백테스트 방지(비용/슬리피지/실패 강제) | `REQ-V2-007` | `TASK-V2-012`,`TASK-CODE-006` | `TEST-ACC-014` |
+| 장애 시 Kill Switch 실행 순서 고정 | `REQ-V2-008` | `TASK-V2-013`,`TASK-V2-014`,`TASK-V2-015`,`TASK-CODE-003` | `TEST-ACC-002`,`TEST-ACC-018` |
+| 세션 전환 단위 리스크/로그 추적 일관화 | `REQ-V3-001`,`REQ-V3-002` | `TASK-V3-001`,`TASK-V3-002`,`TASK-V3-003`,`TASK-CODE-007` | `TEST-ACC-015`,`TEST-ACC-016` |
+| 블랙아웃 중 주문 차단 + 복구 후 재검증 실행 | `REQ-V3-003`,`REQ-V3-004` | `TASK-V3-004`,`TASK-V3-005`,`TASK-V3-006`,`TASK-CODE-008` | `TEST-ACC-003`,`TEST-ACC-017`,`TEST-CODE-005` |
+| 저유동 세션 시장가 주문 금지 | `REQ-V3-005` | `TASK-V3-007`,`TASK-V3-008`,`TASK-CODE-009` | `TEST-ACC-004`,`TEST-CODE-006` |
+| 보수적 체결 모델을 백테스트 기본으로 설정 | `REQ-V3-006` | `TASK-V3-010`,`TASK-V3-011`,`TASK-V3-012`,`TASK-CODE-010` | `TEST-ACC-005`,`TEST-CODE-007` |
+| 전략손익/환율손익 분리 + 통화 버퍼 통제 | `REQ-V3-007` | `TASK-V3-013`,`TASK-V3-014`,`TASK-CODE-011` | `TEST-ACC-006`,`TEST-CODE-008` |
+| 오버나잇 규칙과 Kill Switch 충돌 방지 | `REQ-V3-008` | `TASK-V3-015`,`TASK-CODE-012` | `TEST-ACC-018` |
+| 타임존/정책변경/추적성 문서 거버넌스 | `REQ-OPS-001`,`REQ-OPS-002`,`REQ-OPS-003` | `TASK-OPS-001`,`TASK-OPS-002`,`TASK-OPS-003` | `TEST-ACC-007`,`TEST-ACC-008`,`TEST-ACC-009` |
+
+## 운영 규율 (TPM Enforcement Rules)
+
+- 어떤 PM 시나리오도 `REQ-*` 없는 구현 착수 금지.
+- 어떤 `REQ-*`도 `TASK-*`,`TEST-*` 없는 승인 금지.
+- Verifier는 "코드 리뷰 통과"만으로 승인 불가, 반드시 `TEST-ACC-*` 증적 필요.
+- 배포 승인권자는 Phase 4 체크리스트 미충족 시 릴리즈 보류 권한을 행사해야 한다.

docs/ouroboros/60_repo_enforcement_checklist.md

@@ -0,0 +1,103 @@
+<!--
+Doc-ID: DOC-OPS-002
+Version: 1.0.0
+Status: active
+Owner: tpm
+Updated: 2026-02-26
+-->
+
+# 저장소 강제 설정 체크리스트
+
+목표: "엄격 검증 운영"을 문서가 아니라 저장소 설정으로 강제한다.
+
+## 1) main 브랜치 보호 (필수)
+
+적용 항목:
+- direct push 금지
+- force push 금지
+- branch 삭제 금지
+- merge는 PR 경로만 허용
+
+검증:
+- `main`에 대해 직접 `git push origin main` 시 거부되는지 확인
+
+## 2) 필수 상태 체크 (필수)
+
+필수 CI 항목:
+- `validate_ouroboros_docs` (명령: `python3 scripts/validate_ouroboros_docs.py`)
+- `test` (명령: `pytest -q`)
+
+설정 기준:
+- 위 2개 체크가 `success` 아니면 머지 금지
+- 체크 스킵/중립 상태 허용 금지
+
+## 3) 필수 리뷰어 규칙 (권장 -> 필수)
+
+역할 기반 승인:
+- Verifier 1명 승인 필수
+- TPM 또는 PM 1명 승인 필수
+- Runtime Verifier 관련 변경(PR 본문에 runtime 영향 있음) 시 Runtime Verifier 승인 필수
+
+설정 기준:
+- 최소 승인 수: 2
+- 작성자 self-approval 불가
+- 새 커밋 푸시 시 기존 승인 재검토 요구
+
+## 4) 워크플로우 게이트
+
+병합 전 체크리스트:
+- 이슈 연결(`Closes #N`) 존재
+- PR 본문에 `REQ-*`, `TASK-*`, `TEST-*` 매핑 표 존재
+- `src/core/risk_manager.py` 변경 없음
+- 주요 의사결정 체크포인트(DCP-01~04) 중 해당 단계 Main Agent 확인 기록 존재
+- 주요 의사결정(리뷰 지적/수정 합의/검증 승인)에 대한 에이전트 PR 코멘트 존재
+- 티켓 PR의 base가 `main`이 아닌 program feature branch인지 확인
+
+자동 점검:
+- 문서 검증 스크립트 통과
+- 테스트 통과
+- 개발 완료 시 시스템 구동/모니터링 증적 코멘트 존재
+
+## 5) 감사 추적
+
+필수 보존 증적:
+- CI 실행 로그 링크
+- 검증 실패/복구 기록
+- 머지 승인 코멘트(Verifier/TPM)
+
+분기별 점검:
+- 브랜치 보호 규칙 drift 여부
+- 필수 CI 이름 변경/누락 여부
+
+## 6) 적용 순서 (운영 절차)
+
+1. 브랜치 보호 활성화
+2. 필수 CI 체크 연결
+3. 리뷰어 규칙 적용
+4. 샘플 PR로 거부 시나리오 테스트
+5. 정상 머지 시나리오 테스트
+
+## 7) 실패 시 조치
+
+- 브랜치 보호 미적용 발견 시: 즉시 릴리즈 중지
+- 필수 CI 우회 발견 시: 관리자 권한 점검 및 감사 이슈 발행
+- 리뷰 규칙 무효화 발견 시: 규칙 복구 후 재머지 정책 시행
+- Runtime 이상 이슈 미해결 상태에서 클로즈 시도 발견 시: 즉시 이슈 재오픈 + 릴리즈 중지
+
+## 8) 재계획(Dev Replan) 운영 규칙
+
+- Dev가 `REPLAN-REQUEST` 발행 시 TPM 심사 없이는 스코프/일정 변경 금지
+- `REPLAN-REQUEST`는 Main Agent 승인 전 \"제안\" 상태로 유지
+- 승인된 재계획은 `REQ/TASK/TEST` 문서를 동시 갱신해야 유효
+
+## 9) 서버 반영 규칙
+
+- 티켓 PR(`feature/issue-* -> feature/{stream}`)은 검증 승인 후 머지 가능하다.
+- 최종 통합 PR(`feature/{stream} -> main`)은 사용자 명시 승인 전 `tea pulls merge` 실행 금지.
+- Main 병합 시 승인 근거 코멘트 필수.
+
+## 10) 최종 main 병합 조건
+
+- 모든 티켓이 program feature branch로 병합 완료
+- Runtime Verifier의 구동/모니터링 검증 완료
+- 사용자 최종 승인 코멘트 확인 후에만 `feature -> main` PR 머지 허용

docs/ouroboros/70_main_agent_ideation.md

@@ -0,0 +1,48 @@
+<!--
+Doc-ID: DOC-IDEA-001
+Version: 1.0.0
+Status: active
+Owner: main-agent
+Updated: 2026-02-26
+-->
+
+# 메인 에이전트 아이디에이션 백로그
+
+목적:
+- 구현 진행 중 떠오른 신규 구현 아이디어를 계획 반영 전 임시 저장한다.
+- 본 문서는 사용자 검토 후 다음 계획 포함 여부를 결정하기 위한 검토 큐다.
+
+운영 규칙:
+- 각 아이디어는 `IDEA-*` 식별자를 사용한다.
+- 필수 필드: 배경, 기대효과, 리스크, 후속 티켓 후보.
+- 상태는 `proposed`, `under-review`, `accepted`, `rejected` 중 하나를 사용한다.
+
+## 아이디어 목록
+
+- `IDEA-001` (status: proposed)
+  - 제목: Kill-Switch 전역 상태를 프로세스 단일 전역에서 시장/세션 단위 상태로 분리
+  - 배경: 현재는 전역 block 플래그 기반이라 시장별 분리 제어가 제한될 수 있음
+  - 기대효과: KR/US 병행 운용 시 한 시장 장애가 다른 시장 주문을 불필요하게 막는 리스크 축소
+  - 리스크: 상태 동기화 복잡도 증가, 테스트 케이스 확장 필요
+  - 후속 티켓 후보: `TKT-P1-KS-SCOPE-SPLIT`
+
+- `IDEA-002` (status: proposed)
+  - 제목: Exit Engine 입력 계약(ATR/peak/model_prob/liquidity) 표준 DTO를 데이터 파이프라인에 고정
+  - 배경: 현재 ATR/모델확률 일부가 fallback 기반이라 운영 일관성이 약함
+  - 기대효과: 백테스트-실거래 입력 동형성 강화, 회귀 분석 용이
+  - 리스크: 기존 스캐너/시나리오 엔진 연동 작업량 증가
+  - 후속 티켓 후보: `TKT-P1-EXIT-CONTRACT`
+
+- `IDEA-003` (status: proposed)
+  - 제목: Runtime Verifier 자동 이슈 생성기(로그 패턴 -> 이슈 템플릿 자동화)
+  - 배경: 런타임 이상 리포트가 수동 작성 중심이라 누락 가능성 존재
+  - 기대효과: 이상 탐지 후 이슈 등록 리드타임 단축, 증적 표준화
+  - 리스크: 오탐 이슈 폭증 가능성, 필터링 룰 필요
+  - 후속 티켓 후보: `TKT-P1-RUNTIME-AUTO-ISSUE`
+
+- `IDEA-004` (status: proposed)
+  - 제목: PR 코멘트 워크플로우 자동 점검(리뷰어->개발논의->검증승인 누락 차단)
+  - 배경: 현재 절차는 강력하지만 수행 확인이 수동
+  - 기대효과: 절차 누락 방지, 감사 추적 자동화
+  - 리스크: CLI/API 연동 유지보수 비용
+  - 후속 티켓 후보: `TKT-P0-WORKFLOW-GUARD`

docs/ouroboros/README.md

@@ -0,0 +1,40 @@
+<!--
+Doc-ID: DOC-ROOT-001
+Version: 1.0.0
+Status: active
+Owner: strategy
+Updated: 2026-02-26
+-->
+
+# The Ouroboros 실행 문서 허브
+
+이 폴더는 `ouroboros_plan_v2.txt`, `ouroboros_plan_v3.txt`를 구현 가능한 작업 지시서 수준으로 분해한 문서 허브다.
+
+## 읽기 순서 (Routing)
+
+1. 검증 체계부터 확정: [00_validation_system.md](./00_validation_system.md)
+2. 단일 진실원장(요구사항): [01_requirements_registry.md](./01_requirements_registry.md)
+3. v2 실행 지시서: [10_phase_v2_execution.md](./10_phase_v2_execution.md)
+4. v3 실행 지시서: [20_phase_v3_execution.md](./20_phase_v3_execution.md)
+5. 코드 레벨 작업 지시: [30_code_level_work_orders.md](./30_code_level_work_orders.md)
+6. 수용 기준/테스트 계획: [40_acceptance_and_test_plan.md](./40_acceptance_and_test_plan.md)
+7. PM 시나리오/이슈 분류: [50_scenario_matrix_and_issue_taxonomy.md](./50_scenario_matrix_and_issue_taxonomy.md)
+8. TPM 제어 프로토콜/수용 매트릭스: [50_tpm_control_protocol.md](./50_tpm_control_protocol.md)
+9. 저장소 강제 설정 체크리스트: [60_repo_enforcement_checklist.md](./60_repo_enforcement_checklist.md)
+10. 메인 에이전트 아이디에이션 백로그: [70_main_agent_ideation.md](./70_main_agent_ideation.md)
+
+## 운영 규칙
+
+- 계획 변경은 반드시 `01_requirements_registry.md`의 ID 정의부터 수정한다.
+- 구현 문서는 원장 ID만 참조하고 자체 숫자/정책을 새로 만들지 않는다.
+- 문서 품질 룰셋(`RULE-DOC-001` `RULE-DOC-002` `RULE-DOC-003` `RULE-DOC-004` `RULE-DOC-005` `RULE-DOC-006`)은 [00_validation_system.md](./00_validation_system.md)를 기준으로 적용한다.
+- 문서 병합 전 아래 검증을 통과해야 한다.
+
+```bash
+python3 scripts/validate_ouroboros_docs.py
+```
+
+## 원본 계획 문서
+
+- [v2](/home/agentson/repos/The-Ouroboros/ouroboros_plan_v2.txt)
+- [v3](/home/agentson/repos/The-Ouroboros/ouroboros_plan_v3.txt)

docs/requirements-log.md

@@ -0,0 +1,357 @@
+# Requirements Log
+
+프로젝트 진화를 위한 사용자 요구사항 기록.
+
+이 문서는 시간순으로 사용자와의 대화에서 나온 요구사항과 피드백을 기록합니다.
+새로운 요구사항이 있으면 날짜와 함께 추가하세요.
+
+---
+
+## 2026-02-21
+
+### 거래 상태 확인 중 발견된 버그 (#187)
+
+- 거래 상태 점검 요청 → SELL 주문(손절/익절)이 Fat Finger에 막혀 전혀 실행 안 됨 발견
+- **#187 (Critical)**: SELL 주문에서 Fat Finger 오탐 — `order_amount/total_cash > 30%`가 SELL에도 적용되어 대형 포지션 매도 불가
+  - JELD stop-loss -6.20% → 차단, RXT take-profit +46.13% → 차단
+  - 수정: SELL은 `check_circuit_breaker`만 호출, `validate_order`(Fat Finger 포함) 미호출
+
+---
+
+## 2026-02-20
+
+### 지속적 모니터링 및 개선점 도출 (이슈 #178~#182)
+
+- Dashboard 포함해서 실행하며 간헐적 문제 모니터링 및 개선점 자동 도출 요청
+- 모니터링 결과 발견된 이슈 목록:
+  - **#178**: uvicorn 미설치 → dashboard 미작동 + 오해의 소지 있는 시작 로그 → uvicorn 설치 완료
+  - **#179 (Critical)**: 잔액 부족 주문 실패 후 매 사이클마다 무한 재시도 (MLECW 20분 이상 반복)
+  - **#180**: 다중 인스턴스 실행 시 Telegram 409 충돌
+  - **#181**: implied_rsi 공식 포화 문제 (change_rate≥12.5% → RSI=100)
+  - **#182 (Critical)**: 보유 종목이 SmartScanner 변동성 필터에 걸려 SELL 신호 미생성 → SELL 체결 0건, 잔고 소진
+- 요구사항: 모니터링 자동화 및 주기적 개선점 리포트 도출
+
+---
+
+## 2026-02-05
+
+### API 효율화
+- Gemini API는 귀중한 자원. 종목별 개별 호출 대신 배치 호출 필요
+- Free tier 한도(20 calls/day) 고려하여 일일 몇 차례 거래 모드로 전환
+- 배치 API 호출로 여러 종목을 한 번에 분석
+
+### 거래 모드
+- **Daily Mode**: 하루 4회 거래 세션 (6시간 간격) - Free tier 호환
+- **Realtime Mode**: 60초 간격 실시간 거래 - 유료 구독 필요
+- `TRADE_MODE` 환경변수로 모드 선택
+
+### 진화 시스템
+- 사용자 대화 내용을 문서로 기록하여 향후에도 의도 반영
+- 프롬프트 품질 검증은 별도 이슈로 다룰 예정
+
+### 문서화
+- 시스템 구조, 기능별 설명 등 코드 문서화 항상 신경쓸 것
+- 새로운 기능 추가 시 관련 문서 업데이트 필수
+
+---
+
+## 2026-02-06
+
+### Smart Volatility Scanner (Python-First, AI-Last 파이프라인)
+
+**배경:**
+- 정적 종목 리스트를 순회하는 방식은 비효율적
+- KIS API 거래량 순위를 통해 시장 주도주를 자동 탐지해야 함
+- Gemini API 호출 전에 Python 기반 기술적 분석으로 필터링 필요
+
+**요구사항:**
+1. KIS API 거래량 순위 API 통합 (`fetch_market_rankings`)
+2. 일별 가격 히스토리 API 추가 (`get_daily_prices`)
+3. RSI(14) 계산 기능 구현 (Wilder's smoothing method)
+4. 필터 조건:
+   - 거래량 > 전일 대비 200% (VOL_MULTIPLIER)
+   - RSI < 30 (과매도) OR RSI > 70 (모멘텀)
+5. 상위 1-3개 적격 종목만 Gemini에 전달
+6. 종목 선정 배경(RSI, volume_ratio, signal, score) 데이터베이스 기록
+
+**구현 결과:**
+- `src/analysis/smart_scanner.py`: SmartVolatilityScanner 클래스
+- `src/analysis/volatility.py`: calculate_rsi() 메서드 추가
+- `src/broker/kis_api.py`: 2개 신규 API 메서드
+- `src/db.py`: selection_context 컬럼 추가
+- 설정 가능한 임계값: RSI_OVERSOLD_THRESHOLD, RSI_MOMENTUM_THRESHOLD, VOL_MULTIPLIER, SCANNER_TOP_N
+
+**효과:**
+- Gemini API 호출 20-30개 → 1-3개로 감소
+- Python 기반 빠른 필터링 → 비용 절감
+- 선정 기준 추적 → Evolution 시스템 최적화 가능
+- API 장애 시 정적 watchlist로 자동 전환
+
+**참고:** Realtime 모드 전용. Daily 모드는 배치 효율성을 위해 정적 watchlist 사용.
+
+**이슈/PR:** #76, #77
+
+---
+
+## 2026-02-10
+
+### 코드 리뷰 시 플랜-구현 일치 검증 규칙
+
+**배경:**
+- 코드 리뷰 시 플랜(EnterPlanMode에서 승인된 계획)과 실제 구현이 일치하는지 확인하는 절차가 없었음
+- 플랜과 다른 구현이 리뷰 없이 통과될 위험
+
+**요구사항:**
+1. 모든 PR 리뷰에서 플랜-구현 일치 여부를 필수 체크
+2. 플랜에 없는 변경은 정당한 사유 필요
+3. 플랜 항목이 누락되면 PR 설명에 사유 기록
+4. 스코프가 플랜과 일치하는지 확인
+
+**구현 결과:**
+- `docs/workflow.md`에 Code Review Checklist 섹션 추가
+  - Plan Consistency (필수), Safety & Constraints, Quality, Workflow 4개 카테고리
+
+**이슈/PR:** #114
+
+---
+
+## 2026-02-16
+
+### 문서 v2 동기화 (전체 문서 현행화)
+
+**배경:**
+- v2 기능 구현 완료 후 문서가 실제 코드 상태와 크게 괴리
+- 문서에는 54 tests / 4 files로 기록되었으나 실제로는 551 tests / 25 files
+- v2 핵심 기능(Playbook, Scenario Engine, Dashboard, Telegram Commands, Daily Review, Context System, Backup) 문서화 누락
+
+**요구사항:**
+1. `docs/testing.md` — 551 tests / 25 files 반영, 전체 테스트 파일 설명
+2. `docs/architecture.md` — v2 컴포넌트(Strategy, Context, Dashboard, Decision Logger 등) 추가, Playbook Mode 데이터 플로우, DB 스키마 5개 테이블, v2 환경변수
+3. `docs/commands.md` — Dashboard 실행 명령어, Telegram 명령어 9종 레퍼런스
+4. `CLAUDE.md` — Project Structure 트리 확장, 테스트 수 업데이트, `--dashboard` 플래그
+5. `docs/skills.md` — DB 파일명 `trades.db`로 통일, Dashboard 명령어 추가
+6. 기존에 유효한 트러블슈팅, 코드 예제 등은 유지
+
+**구현 결과:**
+- 6개 문서 파일 업데이트
+- 이전 시도(2개 커밋)는 기존 내용을 과도하게 삭제하여 폐기, main 기준으로 재작업
+
+**이슈/PR:** #131, PR #134
+
+### 해외 스캐너 개선: 랭킹 연동 + 변동성 우선 선별
+
+**배경:**
+- `run_overnight` 실운영에서 미국장 동안 거래가 0건 지속
+- 원인: 해외 시장에서도 국내 랭킹/일봉 API 경로를 사용하던 구조적 불일치
+
+**요구사항:**
+1. 해외 시장도 랭킹 API 기반 유니버스 탐색 지원
+2. 단순 상승률/거래대금 상위가 아니라, **변동성이 큰 종목**을 우선 선별
+3. 고정 티커 fallback 금지
+
+**구현 결과:**
+- `src/broker/overseas.py`
+  - `fetch_overseas_rankings()` 추가 (fluctuation / volume)
+  - 해외 랭킹 API 경로/TR_ID를 설정값으로 오버라이드 가능하게 구현
+- `src/analysis/smart_scanner.py`
+  - market-aware 스캔(국내/해외 분리)
+  - 해외: 랭킹 API 유니버스 + 변동성 우선 점수(일변동률 vs 장중 고저폭)
+  - 거래대금/거래량 랭킹은 유동성 보정 점수로 활용
+  - 랭킹 실패 시에는 동적 유니버스(active/recent/holdings)만 사용
+- `src/config.py`
+  - `OVERSEAS_RANKING_*` 설정 추가
+
+**효과:**
+- 해외 시장에서 스캐너 후보 0개로 정지되는 상황 완화
+- 종목 선정 기준이 단순 상승률 중심에서 변동성 중심으로 개선
+- 고정 티커 없이도 시장 주도 변동 종목 탐지 가능
+
+### 국내 스캐너/주문수량 정렬: 변동성 우선 + 리스크 타기팅
+
+**배경:**
+- 해외만 변동성 우선으로 동작하고, 국내는 RSI/거래량 필터 중심으로 동작해 시장 간 전략 일관성이 낮았음
+- 매수 수량이 고정 1주라서 변동성 구간별 익스포저 관리가 어려웠음
+
+**요구사항:**
+1. 국내 스캐너도 변동성 우선 선별로 해외와 통일
+2. 고변동 종목일수록 포지션 크기를 줄이는 수량 산식 적용
+
+**구현 결과:**
+- `src/analysis/smart_scanner.py`
+  - 국내: `fluctuation ranking + volume ranking bonus` 기반 점수화로 전환
+  - 점수는 `max(abs(change_rate), intraday_range_pct)` 중심으로 계산
+  - 국내 랭킹 응답 스키마 키(`price`, `change_rate`, `volume`) 파싱 보강
+- `src/main.py`
+  - `_determine_order_quantity()` 추가
+  - BUY 시 변동성 점수 기반 동적 수량 산정 적용
+  - `trading_cycle`, `run_daily_session` 경로 모두 동일 수량 로직 사용
+- `src/config.py`
+  - `POSITION_SIZING_*` 설정 추가
+
+**효과:**
+- 국내/해외 스캐너 기준이 변동성 중심으로 일관화
+- 고변동 구간에서 자동 익스포저 축소, 저변동 구간에서 과소진입 완화
+
+## 2026-02-18
+
+### KIS 해외 랭킹 API 404 에러 수정
+
+**배경:**
+- KIS 해외주식 랭킹 API(`fetch_overseas_rankings`)가 모든 거래소에서 HTTP 404를 반환
+- Smart Scanner가 해외 시장 후보 종목을 찾지 못해 거래가 전혀 실행되지 않음
+
+**근본 원인:**
+- TR_ID, API 경로, 거래소 코드가 모두 KIS 공식 문서와 불일치
+
+**구현 결과:**
+- `src/config.py`: TR_ID/Path 기본값을 KIS 공식 스펙으로 수정
+- `src/broker/overseas.py`: 랭킹 API 전용 거래소 코드 매핑 추가 (NASD→NAS, NYSE→NYS, AMEX→AMS), 올바른 API 파라미터 사용
+- `tests/test_overseas_broker.py`: 19개 단위 테스트 추가
+
+**효과:**
+- 해외 시장 랭킹 스캔이 정상 동작하여 Smart Scanner가 후보 종목 탐지 가능
+
+### Gemini prompt_override 미적용 버그 수정
+
+**배경:**
+- `run_overnight` 실행 시 모든 시장에서 Playbook 생성 실패 (`JSONDecodeError`)
+- defensive playbook으로 폴백되어 모든 종목이 HOLD 처리
+
+**근본 원인:**
+- `pre_market_planner.py`가 `market_data["prompt_override"]`에 Playbook 전용 프롬프트를 넣어 `gemini.decide()` 호출
+- `gemini_client.py`의 `decide()` 메서드가 `prompt_override` 키를 전혀 확인하지 않고 항상 일반 트레이드 결정 프롬프트 생성
+- Gemini가 Playbook JSON 대신 일반 트레이드 결정을 반환하여 파싱 실패
+
+**구현 결과:**
+- `src/brain/gemini_client.py`: `decide()` 메서드에서 `prompt_override` 우선 사용 로직 추가
+- `tests/test_brain.py`: 3개 테스트 추가 (override 전달, optimization 우회, 미지정 시 기존 동작 유지)
+
+**이슈/PR:** #143
+
+### 미국장 거래 미실행 근본 원인 분석 및 수정 (자율 실행 세션)
+
+**배경:**
+- 사용자 요청: "미국장 열면 프로그램 돌려서 거래 한 번도 못 한 거 꼭 원인 찾아서 해결해줘"
+- 프로그램을 미국장 개장(9:30 AM EST) 전부터 실행하여 실시간 로그를 분석
+
+**발견된 근본 원인 #1: Defensive Playbook — BUY 조건 없음**
+
+- Gemini free tier (20 RPD) 소진 → `generate_playbook()` 실패 → `_defensive_playbook()` 폴백
+- Defensive playbook은 `price_change_pct_below: -3.0 → SELL` 조건만 존재, BUY 조건 없음
+- ScenarioEngine이 항상 HOLD 반환 → 거래 0건
+
+**수정 #1 (PR #146, Issue #145):**
+- `src/strategy/pre_market_planner.py`: `_smart_fallback_playbook()` 메서드 추가
+  - 스캐너 signal 기반 BUY 조건 생성: `momentum → volume_ratio_above`, `oversold → rsi_below`
+  - 기존 defensive stop-loss SELL 조건 유지
+- Gemini 실패 시 defensive → smart fallback으로 전환
+- 테스트 10개 추가
+
+**발견된 근본 원인 #2: 가격 API 거래소 코드 불일치 + VTS 잔고 API 오류**
+
+실제 로그:
+```
+Scenario matched for MRNX: BUY (confidence=80)  ✓
+Decision for EWUS (NYSE American): BUY (confidence=80)  ✓
+Skip BUY APLZ (NYSE American): no affordable quantity (cash=0.00, price=0.00)  ✗
+```
+
+- `get_overseas_price()`: `NASD`/`NYSE`/`AMEX` 전송 → API가 `NAS`/`NYS`/`AMS` 기대 → 빈 응답 → `price=0`
+- `VTTS3012R` 잔고 API: "ERROR : INPUT INVALID_CHECK_ACNO" → `total_cash=0`
+- 결과: `_determine_order_quantity()` 가 0 반환 → 주문 건너뜀
+
+**수정 #2 (PR #148, Issue #147):**
+- `src/broker/overseas.py`: `_PRICE_EXCHANGE_MAP = _RANKING_EXCHANGE_MAP` 추가, 가격 API에 매핑 적용
+- `src/config.py`: `PAPER_OVERSEAS_CASH: float = Field(default=50000.0)` — paper 모드 시뮬레이션 잔고
+- `src/main.py`: 잔고 0일 때 PAPER_OVERSEAS_CASH 폴백, 가격 0일 때 candidate.price 폴백
+- 테스트 8개 추가
+
+**효과:**
+- BUY 결정 → 실제 주문 전송까지의 파이프라인이 완전히 동작
+- Paper 모드에서 KIS VTS 해외 잔고 API 오류에 관계없이 시뮬레이션 거래 가능
+
+**이슈/PR:** #145, #146, #147, #148
+
+### 해외주식 시장가 주문 거부 수정 (Fix #3, 연속 발견)
+
+**배경:**
+- Fix #147 적용 후 주문 전송 시작 → KIS VTS가 거부: "지정가만 가능한 상품입니다"
+
+**근본 원인:**
+- `trading_cycle()`, `run_daily_session()` 양쪽에서 `send_overseas_order(price=0.0)` 하드코딩
+- `price=0` → `ORD_DVSN="01"` (시장가) 전송 → KIS VTS 거부
+- Fix #147에서 이미 `current_price`를 올바르게 계산했으나 주문 시 미사용
+
+**구현 결과:**
+- `src/main.py`: 두 곳에서 `price=0.0` → `price=current_price`/`price=stock_data["current_price"]`
+- `tests/test_main.py`: 회귀 테스트 `test_overseas_buy_order_uses_limit_price` 추가
+
+**최종 확인 로그:**
+```
+Order result: 모의투자 매수주문이 완료 되었습니다.  ✓
+```
+
+**이슈/PR:** #149, #150
+
+---
+
+## 2026-02-23
+
+### 국내주식 지정가 전환 및 미체결 처리 (#232)
+
+**배경:**
+- 해외주식은 #211에서 지정가로 전환했으나 국내주식은 여전히 `price=0` (시장가)
+- KRX도 지정가 주문 사용 시 동일한 미체결 위험이 존재
+- 지정가 전환 + 미체결 처리를 함께 구현
+
+**구현 내용:**
+
+1. `src/broker/kis_api.py`
+   - `get_domestic_pending_orders()`: 모의 즉시 `[]`, 실전 `TTTC0084R` GET
+   - `cancel_domestic_order()`: 실전 `TTTC0013U` / 모의 `VTTC0013U`, hashkey 필수
+
+2. `src/main.py`
+   - import `kr_round_down` 추가
+   - `trading_cycle`, `run_daily_session` 국내 주문 `price=0` → 지정가:
+     BUY +0.2% / SELL -0.2%, `kr_round_down` KRX 틱 반올림 적용
+   - `handle_domestic_pending_orders` 함수: BUY→취소+쿨다운, SELL→취소+재주문(-0.4%, 최대1회)
+   - daily/realtime 두 모드에서 domestic pending 체크 호출 추가
+
+3. 테스트 14개 추가:
+   - `TestGetDomesticPendingOrders` (3), `TestCancelDomesticOrder` (5)
+   - `TestHandleDomesticPendingOrders` (4), `TestDomesticLimitOrderPrice` (2)
+
+**이슈/PR:** #232, PR #233
+
+---
+
+## 2026-02-24
+
+### 해외잔고 ghost position 수정 — '모의투자 잔고내역이 없습니다' 반복 방지 (#235)
+
+**배경:**
+- 모의투자 실행 시 MLECW, KNRX, NBY, SNSE 등 만료/정지된 종목에 대해
+  `모의투자 잔고내역이 없습니다` 오류가 매 사이클 반복됨
+
+**근본 원인:**
+1. `ovrs_cblc_qty` (해외잔고수량, 총 보유) vs `ord_psbl_qty` (주문가능수량, 실제 매도 가능)
+   - 기존 코드: `ovrs_cblc_qty` 우선 사용 → 만료 Warrant가 `ovrs_cblc_qty=289456`이지만 실제 `ord_psbl_qty=0`
+   - startup sync / build_overseas_symbol_universe가 이 종목들을 포지션으로 기록
+2. SELL 실패 시 DB 포지션이 닫히지 않아 다음 사이클에서도 재시도 (무한 반복)
+
+**구현 내용:**
+
+1. `src/main.py` — `_extract_held_codes_from_balance`, `_extract_held_qty_from_balance`
+   - 해외 잔고 필드 우선순위 변경: `ord_psbl_qty` → `ovrs_cblc_qty` → `hldg_qty` (fallback 유지)
+   - KIS 공식 문서(VTTS3012R) 기준: `ord_psbl_qty`가 실제 매도 가능 수량
+
+2. `src/main.py` — `trading_cycle` ghost-close 처리
+   - 해외 SELL이 `잔고내역이 없습니다`로 실패 시 DB 포지션을 `[ghost-close]` SELL로 종료
+   - exchange code 불일치 등 예외 상황에서 무한 반복 방지
+
+3. 테스트 7개 추가:
+   - `TestExtractHeldQtyFromBalance` 3개: ord_psbl_qty 우선, 0이면 0 반환, fallback
+   - `TestExtractHeldCodesFromBalance` 2개: ord_psbl_qty=0인 종목 제외, fallback
+   - `TestOverseasGhostPositionClose` 2개: ghost-close 로그 확인, 일반 오류 무시
+
+**이슈/PR:** #235, PR #236

docs/skills.md

@@ -34,6 +34,12 @@ python -m src.main --mode=paper
 ```
 Runs the agent in paper-trading mode (no real orders).
 
+### Start Trading Agent with Dashboard
+```bash
+python -m src.main --mode=paper --dashboard
+```
+Runs the agent with FastAPI dashboard on `127.0.0.1:8080` (configurable via `DASHBOARD_HOST`/`DASHBOARD_PORT`).
+
 ### Start Trading Agent (Production)
 ```bash
 docker compose up -d ouroboros
@@ -59,7 +65,7 @@ Analyze the last 30 days of trade logs and generate performance metrics.
 python -m src.evolution.optimizer --evolve
 ```
 Triggers the evolution engine to:
-1. Analyze `trade_logs.db` for failing patterns
+1. Analyze `trades.db` for failing patterns
 2. Ask Gemini to generate a new strategy
 3. Run tests on the new strategy
 4. Create a PR if tests pass
@@ -91,12 +97,12 @@ curl http://localhost:8080/health
 
 ### View Trade Logs
 ```bash
-sqlite3 data/trade_logs.db "SELECT * FROM trades ORDER BY timestamp DESC LIMIT 20;"
+sqlite3 data/trades.db "SELECT * FROM trades ORDER BY timestamp DESC LIMIT 20;"
 ```
 
 ### Export Trade History
 ```bash
-sqlite3 -header -csv data/trade_logs.db "SELECT * FROM trades;" > trades_export.csv
+sqlite3 -header -csv data/trades.db "SELECT * FROM trades;" > trades_export.csv
 ```
 
 ## Safety Checklist (Pre-Deploy)

docs/testing.md

@@ -0,0 +1,287 @@
+# Testing Guidelines
+
+## Test Structure
+
+**551 tests** across **25 files**. `asyncio_mode = "auto"` in pyproject.toml — async tests need no special decorator.
+
+The `settings` fixture in `conftest.py` provides safe defaults with test credentials and in-memory DB.
+
+### Test Files
+
+#### Core Components
+
+##### `tests/test_risk.py` (14 tests)
+- Circuit breaker boundaries and exact threshold triggers
+- Fat-finger edge cases and percentage validation
+- P&L calculation edge cases
+- Order validation logic
+
+##### `tests/test_broker.py` (11 tests)
+- OAuth token lifecycle
+- Rate limiting enforcement
+- Hash key generation
+- Network error handling
+- SSL context configuration
+
+##### `tests/test_brain.py` (24 tests)
+- Valid JSON parsing and markdown-wrapped JSON handling
+- Malformed JSON fallback
+- Missing fields handling
+- Invalid action validation
+- Confidence threshold enforcement
+- Empty response handling
+- Prompt construction for different markets
+
+##### `tests/test_market_schedule.py` (24 tests)
+- Market open/close logic
+- Timezone handling (UTC, Asia/Seoul, America/New_York, etc.)
+- DST (Daylight Saving Time) transitions
+- Weekend handling and lunch break logic
+- Multiple market filtering
+- Next market open calculation
+
+##### `tests/test_db.py` (3 tests)
+- Database initialization and table creation
+- Trade logging with all fields (market, exchange_code, decision_id)
+- Query and retrieval operations
+
+##### `tests/test_main.py` (37 tests)
+- Trading loop orchestration
+- Market iteration and stock processing
+- Dashboard integration (`--dashboard` flag)
+- Telegram command handler wiring
+- Error handling and graceful shutdown
+
+#### Strategy & Playbook (v2)
+
+##### `tests/test_pre_market_planner.py` (37 tests)
+- Pre-market playbook generation
+- Gemini API integration for scenario creation
+- Timeout handling and defensive playbook fallback
+- Multi-market playbook generation
+
+##### `tests/test_scenario_engine.py` (44 tests)
+- Scenario matching against live market data
+- Confidence scoring and threshold filtering
+- Multiple scenario type handling
+- Edge cases (no match, partial match, expired scenarios)
+
+##### `tests/test_playbook_store.py` (23 tests)
+- Playbook persistence to SQLite
+- Date-based retrieval and market filtering
+- Playbook status management (generated, active, expired)
+- JSON serialization/deserialization
+
+##### `tests/test_strategy_models.py` (33 tests)
+- Pydantic model validation for scenarios, playbooks, decisions
+- Field constraints and default values
+- Serialization round-trips
+
+#### Analysis & Scanning
+
+##### `tests/test_volatility.py` (24 tests)
+- ATR and RSI calculation accuracy
+- Volume surge ratio computation
+- Momentum scoring
+- Breakout/breakdown pattern detection
+- Market scanner watchlist management
+
+##### `tests/test_smart_scanner.py` (13 tests)
+- Python-first filtering pipeline
+- RSI and volume ratio filter logic
+- Candidate scoring and ranking
+- Fallback to static watchlist
+
+#### Context & Memory
+
+##### `tests/test_context.py` (18 tests)
+- L1-L7 layer storage and retrieval
+- Context key-value CRUD operations
+- Timeframe-based queries
+- Layer metadata management
+
+##### `tests/test_context_scheduler.py` (5 tests)
+- Periodic context aggregation scheduling
+- Layer summarization triggers
+
+#### Evolution & Review
+
+##### `tests/test_evolution.py` (24 tests)
+- Strategy optimization loop
+- High-confidence losing trade analysis
+- Generated strategy validation
+
+##### `tests/test_daily_review.py` (10 tests)
+- End-of-day review generation
+- Trade performance summarization
+- Context layer (L6_DAILY) integration
+
+##### `tests/test_scorecard.py` (3 tests)
+- Daily scorecard metrics calculation
+- Win rate, P&L, confidence tracking
+
+#### Notifications & Commands
+
+##### `tests/test_telegram.py` (25 tests)
+- Message sending and formatting
+- Rate limiting (leaky bucket)
+- Error handling (network timeout, invalid token)
+- Auto-disable on missing credentials
+- Notification types (trade, circuit breaker, fat-finger, market events)
+
+##### `tests/test_telegram_commands.py` (31 tests)
+- 9 command handlers (/help, /status, /positions, /report, /scenarios, /review, /dashboard, /stop, /resume)
+- Long polling and command dispatch
+- Authorization filtering by chat_id
+- Command response formatting
+
+#### Dashboard
+
+##### `tests/test_dashboard.py` (14 tests)
+- FastAPI endpoint responses (8 API routes)
+- Status, playbook, scorecard, performance, context, decisions, scenarios
+- Query parameter handling (market, date, limit)
+
+#### Performance & Quality
+
+##### `tests/test_token_efficiency.py` (34 tests)
+- Gemini token usage optimization
+- Prompt size reduction verification
+- Cache effectiveness
+
+##### `tests/test_latency_control.py` (30 tests)
+- API call latency measurement
+- Rate limiter timing accuracy
+- Async operation overhead
+
+##### `tests/test_decision_logger.py` (9 tests)
+- Decision audit trail completeness
+- Context snapshot capture
+- Outcome tracking (P&L, accuracy)
+
+##### `tests/test_data_integration.py` (38 tests)
+- External data source integration
+- News API, market data, economic calendar
+- Error handling for API failures
+
+##### `tests/test_backup.py` (23 tests)
+- Backup scheduler and execution
+- Cloud storage (S3) upload
+- Health monitoring
+- Data export functionality
+
+## Coverage Requirements
+
+**Minimum coverage: 80%**
+
+Check coverage:
+```bash
+pytest -v --cov=src --cov-report=term-missing
+```
+
+**Note:** `main.py` has lower coverage as it contains the main loop which is tested via integration/manual testing.
+
+## Test Configuration
+
+### `pyproject.toml`
+```toml
+[tool.pytest.ini_options]
+asyncio_mode = "auto"
+testpaths = ["tests"]
+python_files = ["test_*.py"]
+```
+
+### `tests/conftest.py`
+```python
+@pytest.fixture
+def settings() -> Settings:
+    """Provide test settings with safe defaults."""
+    return Settings(
+        KIS_APP_KEY="test_key",
+        KIS_APP_SECRET="test_secret",
+        KIS_ACCOUNT_NO="12345678-01",
+        GEMINI_API_KEY="test_gemini_key",
+        MODE="paper",
+        DB_PATH=":memory:",  # In-memory SQLite
+        CONFIDENCE_THRESHOLD=80,
+        ENABLED_MARKETS="KR",
+    )
+```
+
+## Writing New Tests
+
+### Naming Convention
+- Test files: `test_<module>.py`
+- Test functions: `test_<feature>_<scenario>()`
+- Use descriptive names that explain what is being tested
+
+### Good Test Example
+```python
+async def test_send_order_with_market_price(broker, settings):
+    """Market orders should use price=0 and ORD_DVSN='01'."""
+    # Arrange
+    stock_code = "005930"
+    order_type = "BUY"
+    quantity = 10
+
+    # Act
+    with patch.object(broker._session, 'post') as mock_post:
+        mock_post.return_value.__aenter__.return_value.status = 200
+        mock_post.return_value.__aenter__.return_value.json = AsyncMock(
+            return_value={"rt_cd": "0", "msg1": "OK"}
+        )
+
+        await broker.send_order(stock_code, order_type, quantity, price=0)
+
+    # Assert
+    call_args = mock_post.call_args
+    body = call_args.kwargs['json']
+    assert body['ORD_DVSN'] == '01'  # Market order
+    assert body['ORD_UNPR'] == '0'   # Price 0
+```
+
+### Test Checklist
+- [ ] Test passes in isolation (`pytest tests/test_foo.py::test_bar -v`)
+- [ ] Test has clear docstring explaining what it tests
+- [ ] Arrange-Act-Assert structure
+- [ ] Uses appropriate fixtures from conftest.py
+- [ ] Mocks external dependencies (API calls, network)
+- [ ] Tests edge cases and error conditions
+- [ ] Doesn't rely on test execution order
+
+## Running Tests
+
+```bash
+# All tests
+pytest -v
+
+# Specific file
+pytest tests/test_risk.py -v
+
+# Specific test
+pytest tests/test_brain.py::test_parse_valid_json -v
+
+# With coverage
+pytest -v --cov=src --cov-report=term-missing
+
+# Stop on first failure
+pytest -x
+
+# Verbose output with print statements
+pytest -v -s
+```
+
+## CI/CD Integration
+
+Tests run automatically on:
+- Every commit to feature branches
+- Every PR to main
+- Scheduled daily runs
+
+**Blocking conditions:**
+- Test failures → PR blocked
+- Coverage < 80% → PR blocked (warning only for main.py)
+
+**Non-blocking:**
+- `mypy --strict` errors (type hints encouraged but not enforced)
+- `ruff check` warnings (must be acknowledged)

docs/workflow.md

@@ -0,0 +1,172 @@
+# Development Workflow
+
+## Git Workflow Policy
+
+**CRITICAL: All code changes MUST follow this workflow. Direct pushes to `main` are ABSOLUTELY PROHIBITED.**
+
+1. **Create Gitea Issue First** — All features, bug fixes, and policy changes require a Gitea issue before any code is written
+2. **Create Program Feature Branch** — Branch from `main` for the whole development stream
+   - Format: `feature/{epic-or-stream-name}`
+3. **Create Ticket Temp Branch** — Branch from the program feature branch per ticket
+   - Format: `feature/issue-{N}-{short-description}`
+4. **Implement Per Ticket** — Write code, tests, and documentation on the ticket temp branch
+5. **Create Pull Request to Program Feature Branch** — `feature/issue-N-* -> feature/{stream}`
+6. **Review/Verify and Merge into Program Feature Branch** — user approval not required
+7. **Final Integration PR to main** — Only after all ticket stages complete and explicit user approval
+
+**Never commit directly to `main`.** This policy applies to all changes, no exceptions.
+
+## Branch Strategy (Mandatory)
+
+- Team operation default branch is the **program feature branch**, not `main`.
+- Ticket-level development happens only on **ticket temp branches** cut from the program feature branch.
+- Ticket PR merges into program feature branch are allowed after verifier approval.
+- Until final user sign-off, `main` merge is prohibited.
+- 각 에이전트는 주요 의사결정(리뷰 지적, 수정 방향, 검증 승인)마다 PR 코멘트를 적극 작성해 의사결정 과정을 남긴다.
+
+## Gitea CLI Formatting Troubleshooting
+
+Issue/PR 본문 작성 시 줄바꿈(`\n`)이 문자열 그대로 저장되는 문제가 반복될 수 있다. 원인은 `-d "...\n..."` 형태에서 쉘/CLI가 이스케이프를 실제 개행으로 해석하지 않기 때문이다.
+
+권장 패턴:
+
+```bash
+ISSUE_BODY=$(cat <<'EOF'
+## Summary
+- 변경 내용 1
+- 변경 내용 2
+
+## Why
+- 배경 1
+- 배경 2
+
+## Scope
+- 포함 범위
+- 제외 범위
+EOF
+)
+
+tea issues create \
+  -t "docs: 제목" \
+  -d "$ISSUE_BODY"
+```
+
+PR도 동일하게 적용:
+
+```bash
+PR_BODY=$(cat <<'EOF'
+## Summary
+- ...
+
+## Validation
+- python3 scripts/validate_ouroboros_docs.py
+EOF
+)
+
+tea pr create \
+  --base main \
+  --head feature/issue-N-something \
+  --title "docs: ... (#N)" \
+  --description "$PR_BODY"
+```
+
+금지 패턴:
+
+- `-d "line1\nline2"` (웹 UI에 `\n` 문자 그대로 노출될 수 있음)
+- 본문에 백틱/괄호를 인라인로 넣고 적절한 quoting 없이 즉시 실행
+
+## Agent Workflow
+
+**Modern AI development leverages specialized agents for concurrent, efficient task execution.**
+
+### Parallel Execution Strategy
+
+Use **git worktree** or **subagents** (via the Task tool) to handle multiple requirements simultaneously:
+
+- Each task runs in independent context
+- Parallel branches for concurrent features
+- Isolated test environments prevent interference
+- Faster iteration with distributed workload
+
+### Specialized Agent Roles
+
+Deploy task-specific agents as needed instead of handling everything in the main conversation:
+
+- **Conversational Agent** (main) — Interface with user, coordinate other agents
+- **Ticket Management Agent** — Create/update Gitea issues, track task status
+- **Design Agent** — Architectural planning, RFC documents, API design
+- **Code Writing Agent** — Implementation following specs
+- **Testing Agent** — Write tests, verify coverage, run test suites
+- **Documentation Agent** — Update docs, docstrings, CLAUDE.md, README
+- **Review Agent** — Code review, lint checks, security audits
+- **Custom Agents** — Created dynamically for specialized tasks (performance analysis, migration scripts, etc.)
+
+### When to Use Agents
+
+**Prefer spawning specialized agents for:**
+
+1. Complex multi-file changes requiring exploration
+2. Tasks with clear, isolated scope (e.g., "write tests for module X")
+3. Parallel work streams (feature A + bugfix B simultaneously)
+4. Long-running analysis (codebase search, dependency audit)
+5. Tasks requiring different contexts (multiple git worktrees)
+
+**Use the main conversation for:**
+
+1. User interaction and clarification
+2. Quick single-file edits
+3. Coordinating agent work
+4. High-level decision making
+
+### Implementation
+
+```python
+# Example: Spawn parallel test and documentation agents
+task_tool(
+    subagent_type="general-purpose",
+    prompt="Write comprehensive tests for src/markets/schedule.py",
+    description="Write schedule tests"
+)
+
+task_tool(
+    subagent_type="general-purpose",
+    prompt="Update README.md with global market feature documentation",
+    description="Update README"
+)
+```
+
+Use `run_in_background=True` for independent tasks that don't block subsequent work.
+
+## Code Review Checklist
+
+**CRITICAL: Every PR review MUST verify plan-implementation consistency.**
+
+Before approving any PR, the reviewer (human or agent) must check ALL of the following:
+
+### 1. Plan Consistency (MANDATORY)
+
+- [ ] **Implementation matches the approved plan** — Compare the actual code changes against the plan created during `EnterPlanMode`. Every item in the plan must be addressed.
+- [ ] **No unplanned changes** — If the implementation includes changes not in the plan, they must be explicitly justified.
+- [ ] **No plan items omitted** — If any planned item was skipped, the reason must be documented in the PR description.
+- [ ] **Scope matches** — The PR does not exceed or fall short of the planned scope.
+
+### 2. Safety & Constraints
+
+- [ ] `src/core/risk_manager.py` is unchanged (READ-ONLY)
+- [ ] Circuit breaker threshold not weakened (only stricter allowed)
+- [ ] Fat-finger protection (30% max order) still enforced
+- [ ] Confidence < 80 still forces HOLD
+- [ ] No hardcoded API keys or secrets
+
+### 3. Quality
+
+- [ ] All new/modified code has corresponding tests
+- [ ] Test coverage >= 80%
+- [ ] `ruff check src/ tests/` passes (no lint errors)
+- [ ] No `assert` statements removed from tests
+
+### 4. Workflow
+
+- [ ] PR references the Gitea issue number
+- [ ] Feature branch follows naming convention (`feature/issue-N-description`)
+- [ ] Commit messages are clear and descriptive

ouroboros_plan_v2.txt

@@ -0,0 +1,165 @@
+[The Ouroboros] 운영/전략 계획서 v2
+작성일: 2026-02-26
+상태: 코드 구현 전 설계안(전략/검증 중심)
+
+==================================================
+0) 목적
+==================================================
+고정 익절(+3%) 중심 로직에서 벗어나, 다음을 만족하는 실전형 청산 체계로 전환한다.
+- 수익 구간 보호 (손익 역전 방지)
+- 변동성 적응형 청산
+- 예측 모델의 확률 신호를 보조적으로 결합
+- 과적합 방지를 최우선으로 한 검증 프레임워크
+
+==================================================
+1) 핵심 설계 원칙
+==================================================
+1. 예측 성능과 전략 성능을 분리 평가
+- 예측 성능: PR-AUC, Brier, Calibration
+- 전략 성능: Net PnL, Sharpe, MDD, Profit Factor, Turnover
+
+2. 시계열 검증 규율 강제
+- Walk-forward 분할
+- Purge/Embargo 적용
+- Random split 금지
+
+3. 실거래 리얼리즘 우선
+- 거래비용/슬리피지/체결실패 반영 없는 백테스트 결과는 채택 금지
+
+==================================================
+2) 매도 상태기계 (State Machine)
+==================================================
+상태:
+- HOLDING
+- BE_LOCK
+- ARMED
+- EXITED
+
+정의:
+- HOLDING: 일반 보유 상태
+- BE_LOCK: 일정 수익권 진입 시 손절선을 본전(또는 비용 반영 본전)으로 상향
+- ARMED: 추세 추적(피크 추적) 기반 청산 준비 상태
+- EXITED: 청산 완료
+
+전이 규칙(개념):
+- HOLDING -> BE_LOCK: unrealized_pnl_pct >= be_arm_pct
+- BE_LOCK -> ARMED: unrealized_pnl_pct >= arm_pct
+- ARMED -> EXITED: 아래 조건 중 하나 충족
+  1) hard stop 도달
+  2) trailing stop 도달 (peak 대비 하락)
+  3) 모델 하락확률 + 유동성 약화 조건 충족
+
+상태 전이 구현 규칙(필수):
+- 매 틱/바 평가 시 "현재 조건이 허용하는 최상위 상태"로 즉시 승격
+- 순차 if-else로 인한 전이 누락 금지 (예: 갭으로 BE_LOCK/ARMED 동시 충족)
+- EXITED 조건은 모든 상태보다 우선 평가
+- 상태 전이 로그에 이전/이후 상태, 전이 사유, 기준 가격/수익률 기록
+
+==================================================
+3) 청산 로직 구성 (4중 안전장치)
+==================================================
+A. Hard Stop
+- 계좌/포지션 보호용 절대 하한
+- 항상 활성화
+
+B. Dynamic Stop (Break-even Lock)
+- BE_LOCK 진입 시 손절선을 본전 이상으로 상향
+- "수익 포지션이 손실로 반전"되는 구조적 리스크 차단
+
+C. ATR 기반 Trailing Stop
+- 고정 trail_pct 대신 변동성 적응형 사용
+- 예시: ExitPrice = PeakPrice - (k * ATR)
+
+D. 모델 확률 신호
+- 하락전환 확률(pred_prob)이 임계값 이상일 때 청산 가중
+- 단독 트리거가 아닌 trailing/리스크 룰 보조 트리거로 사용
+
+==================================================
+4) 라벨링 체계 (Triple Barrier)
+==================================================
+목표:
+고정 H-window 라벨 편향을 줄이고, 금융 시계열의 경로 의존성을 반영한다.
+
+라벨 정의:
+- Upper barrier (익절)
+- Lower barrier (손절)
+- Time barrier (만기)
+
+규칙:
+- 세 장벽 중 "먼저 터치한 장벽"으로 라벨 확정
+- 라벨은 entry 시점 이후 데이터만 사용해 생성
+- 피처 생성 구간과 라벨 구간을 엄격 분리해 look-ahead bias 방지
+
+==================================================
+5) 검증 프레임워크
+==================================================
+5.1 분할 방식
+- Fold 단위 Walk-forward
+- Purge/Embargo로 인접 샘플 누수 차단
+
+5.2 비교군(Baseline) 구조
+- B0: 기존 고정 손절/익절
+- B1: 모델 없는 trailing only
+- M1: trailing + 모델 확률 결합
+
+5.3 채택 기준
+- M1이 B0/B1 대비 OOS(Out-of-sample)에서 일관된 우위
+- 단일 구간 성과가 아닌 fold 분포 기준으로 판단
+
+==================================================
+6) 실행 아키텍처 원칙
+==================================================
+1. 저지연 실행 경로
+- 실시간 청산 판단은 경량 엔진(룰/GBDT) 담당
+- LLM은 레짐 판단/비중 조절/상위 의사결정 보조
+
+2. 체결 현실 반영
+- 세션 유동성에 따른 슬리피지 페널티 차등 적용
+- 미체결/재호가/재접수 시나리오를 백테스트에 반영
+
+==================================================
+7) 운영 리스크 관리
+==================================================
+승격 단계:
+- Offline backtest -> Paper shadow -> Small-capital live
+
+중단(Kill Switch):
+- rolling Sharpe 악화
+- MDD 한도 초과
+- 체결 실패율/슬리피지 급등
+
+Kill Switch 실행 순서(원자적):
+1) 모든 신규 주문 차단 플래그 ON
+2) 모든 미체결 주문 취소 요청
+3) 취소 결과 재조회(실패 건 재시도)
+4) 포지션 리스크 재계산 후 강제 축소/청산 판단
+5) 상태/로그 스냅샷 저장 및 운영 경보 발송
+
+원칙:
+- 모델이 실패해도 hard stop 기반 보수 모드로 즉시 디그레이드 가능해야 함
+
+==================================================
+8) 고정 파라미터(초기안)
+==================================================
+(15분봉 단기 스윙 기준 제안)
+- KR: be_arm_pct=1.2, arm_pct=2.8, atr_period=14, atr_multiplier_k=2.2,
+      time_barrier_bars=26, p_thresh=0.62
+- US: be_arm_pct=1.0, arm_pct=2.4, atr_period=14, atr_multiplier_k=2.0,
+      time_barrier_bars=32, p_thresh=0.60
+
+민감도 범위(초기 탐색):
+- be_arm_pct: KR 0.9~1.8 / US 0.7~1.5
+- arm_pct: KR 2.2~3.8 / US 1.8~3.2
+- atr_multiplier_k: KR 1.8~2.8 / US 1.6~2.4
+- time_barrier_bars: KR 20~36 / US 24~48
+- p_thresh: 0.55~0.70
+
+==================================================
+9) 구현 전 체크리스트
+==================================================
+- 파라미터 튜닝 시 nested leakage 방지
+- 수수료/세금/슬리피지 전부 반영 여부 확인
+- 세션/타임존/DST 처리 일관성 확인
+- 모델 버전/설정 해시/실험 로그 재현성 확보
+
+끝.

ouroboros_plan_v3.txt

@@ -0,0 +1,185 @@
+[The Ouroboros] 운영확장 v3
+작성일: 2026-02-26
+상태: v2 확장판 / 야간·프리마켓 포함 글로벌 세션 운영 설계안
+
+==================================================
+0) 목적
+==================================================
+"24시간 무중단 자산 증식" 비전을 위해 거래 세션 범위를 KR 정규장 중심에서
+NXT/미국 확장 세션까지 확대한다. 핵심은 다음 3가지다.
+- 세션 인지형 의사결정
+- 세션별 리스크/비용 차등 적용
+- 시간장벽의 현실적 재정의
+
+==================================================
+1) 세션 모델 (Session-aware Engine)
+==================================================
+KR 세션:
+- NXT_PRE   : 08:00 ~ 08:50 (KST)
+- KRX_REG   : 09:00 ~ 15:30 (KST)
+- NXT_AFTER : 15:30 ~ 20:00 (KST)
+
+US 세션(KST 관점 운영):
+- US_DAY    : 10:00 ~ 18:00
+- US_PRE    : 18:00 ~ 23:30
+- US_REG    : 23:30 ~ 06:00
+- US_AFTER  : 06:00 ~ 07:00
+
+원칙:
+- 모든 피처/신호/주문/로그에 session_id를 명시적으로 포함
+- 세션 전환 시 상태 업데이트 및 리스크 파라미터 재로딩
+
+==================================================
+2) 캘린더/휴장/DST 고정 소스
+==================================================
+KR:
+- 기본: pykrx 또는 FinanceDataReader (KRX 기준)
+- 예외: 연휴/임시 휴장/NXT 특이 운영은 KIS 공지 기반 보완
+
+US:
+- pandas_market_calendars (NYSE 기준)
+- 2026 DST:
+  - 시작: 2026-03-08
+  - 종료: 2026-11-01
+
+정합성 규칙:
+- 스케줄 충돌 시 "거래소 캘린더 > 로컬 추정" 우선
+- 시장 상태(open/close/half-day)는 주문 엔진 진입 전 최종 검증
+
+KIS 점검시간 회피 정책(필수):
+- 브로커 점검/장애 블랙아웃 윈도우는 운영 설정으로 별도 관리
+- 블랙아웃 구간에는 신규 주문 전송 금지, 취소/정정도 정책적으로 제한
+- 신호는 유지하되 주문 의도는 Queue에 적재, 복구 후 유효성 재검증 뒤 실행
+- 복구 직후에는 잔고/미체결/체결내역을 우선 동기화한 뒤 주문 엔진 재가동
+
+==================================================
+3) 시간장벽 재정의
+==================================================
+v2의 time_barrier_bars 고정값을 v3에서 다음으로 확장:
+- max_holding_minutes (시장별 기본 만기)
+- 봉 개수는 세션 길이/간격으로 동적 계산
+
+기본값:
+- KR: max_holding_minutes = 2160 (약 3거래일, NXT 포함 관점)
+- US: max_holding_minutes = 4320 (약 72시간)
+
+운영 주의:
+- 고정 "일중 청산"보다 "포지션 유지 시간" 기준 만기 적용
+- 세션 종료 강제청산 규칙과 충돌 시 우선순위 명시 필요
+
+==================================================
+4) 세션별 비용/슬리피지 모델 (보수적)
+==================================================
+KRX_REG:
+- 슬리피지: 2~3틱 (약 0.05%)
+- 수수료+세금: 0.20% ~ 0.23%
+
+NXT_AFTER:
+- 슬리피지: 5~8틱 (약 0.15%)
+- 수수료+세금: 0.20% ~ 0.23%
+
+US_REG:
+- 슬리피지: 2~3틱 (약 0.03%)
+- 수수료+기타 비용: 0.07% ~ 0.15%
+
+US_PRE / US_DAY:
+- 슬리피지: 10틱+ (약 0.3% ~ 0.5%)
+- 수수료+기타 비용: 0.07% ~ 0.15%
+
+원칙:
+- 백테스트 체결가는 세션별 보수 가정 적용
+- 저유동 세션은 자동 보수 모드(p_thresh 상향, atr_k 상향) 권장
+- 백테스트 체결가 기본은 "불리한 방향 체결" 가정 (단순 close 체결 금지)
+
+세션별 주문 유형 강제(필수):
+- KRX_REG / US_REG: 지정가 우선, 시장가 제한적 허용
+- NXT_AFTER / US_PRE / US_DAY / US_AFTER: 시장가 금지
+- 저유동 세션은 최우선 지정가 또는 IOC/FOK(가격 보호 한도 포함)만 허용
+- 주문 실패 시 재호가 간격/횟수 상한을 두고, 초과 시 주문 철회
+
+==================================================
+5) 포지션/잔고 통합 규칙 (KIS 특성 반영)
+==================================================
+문제:
+- KRX/NXT 잔고 조회가 venue 단위로 분리되거나 반영 지연 가능
+
+규칙:
+- 종목 식별은 동일 종목코드(또는 ISIN) 기준 통합 포지션으로 관리
+- 다만 주문 가능 수량은 venue별 API 응답을 최종 기준으로 사용
+- 매도 가능 수량 검증은 주문 직전 재조회로 확정
+
+==================================================
+6) 마감 강제청산/오버나잇 예외 규칙
+==================================================
+기본 원칙:
+- 모든 포지션에 대해 세션 종료 10분 전 REDUCE_ALL 검토
+
+오버나잇 예외 허용 (모두 충족 시):
+1) ARMED 상태 (예: +2.8% 이상)
+2) 모델 하락확률 < 0.30
+3) 포트폴리오 현금 비중 >= 50%
+
+갭 리스크 통제:
+- 다음 개장 시 hard stop를 시가 기준으로 재산정
+- 조건 위반 시 즉시 청산 우선
+
+Kill Switch 연동:
+- MDD/실패율 임계치 초과 시 "미체결 전량 취소 -> 신규 주문 차단 -> 리스크 축소" 순서 강제
+
+==================================================
+7) 데이터 저장/용량 정책
+==================================================
+핵심 테이블(계획):
+- feature_snapshots
+- position_states
+- model_predictions
+
+저장 규칙:
+- feature_hash 기반 중복 제거
+- 가격 변화가 작아도 session_id 변경 시 강제 스냅샷
+- 월 단위 DB 로테이션 권장 (예: trading_YYYY_MM.db)
+
+==================================================
+8) 환율/정산 리스크 정책 (US 필수)
+==================================================
+원칙:
+- USD 노출은 전략 손익과 별도로 환율 손익을 분리 추적
+- 원화 주문 서비스 사용 시 가환율 체결/익일 정산 리스크를 예수금 규칙에 반영
+
+운영 규칙:
+- 환전 시점 정책(사전 환전/수시 환전)을 고정하고 로그에 기록
+- 최소 USD 버퍼와 KRW 버퍼를 각각 설정해 주문 가능금 부족 리스크 완화
+- 환율 급변 구간에는 포지션 한도 축소 또는 신규 진입 제한
+
+==================================================
+9) v3 실험 매트릭스 (우선 3선)
+==================================================
+EXP-KR-01:
+- 시장: KR
+- 포커스: NXT 야간 특화
+- 제안: time barrier 확장(예: 48 bars 상당), p_thresh 상향(0.65)
+
+EXP-US-01:
+- 시장: US
+- 포커스: 21h 준연속 운용
+- 제안: time barrier 확장(예: 80 bars 상당), atr_k 상향(2.5)
+
+EXP-HYB-01:
+- 시장: Global
+- 포커스: KR 낮 + US 밤 연계
+- 제안: 레짐 기반 자산배분 자동조절
+
+==================================================
+10) 코드 착수 전 최종 확정 체크
+==================================================
+1) 세션별 공식 캘린더 소스/우선순위
+2) 세션별 슬리피지/비용 테이블 수치
+3) 시장별 max_holding_minutes
+4) 마감 강제청산 예외 조건 임계값
+5) 블랙아웃(점검/장애) 시간대와 주문 큐 처리 규칙
+6) 세션별 허용 주문 유형(시장가 허용 범위 포함)
+7) 환전/정산 정책 및 통화 버퍼 임계값
+
+모든 항목 확정 후 Step 1 구현(코드)로 이동.
+
+끝.

pyproject.toml

@@ -8,6 +8,9 @@ dependencies = [
     "pydantic>=2.5,<3",
     "pydantic-settings>=2.1,<3",
     "google-genai>=1.0,<2",
+    "scipy>=1.11,<2",
+    "fastapi>=0.110,<1",
+    "uvicorn>=0.29,<1",
 ]
 
 [project.optional-dependencies]

scripts/backup.sh

@@ -0,0 +1,96 @@
+#!/usr/bin/env bash
+# Automated backup script for The Ouroboros trading system
+# Runs daily/weekly/monthly backups
+
+set -euo pipefail
+
+# Configuration
+DB_PATH="${DB_PATH:-data/trade_logs.db}"
+BACKUP_DIR="${BACKUP_DIR:-data/backups}"
+PYTHON="${PYTHON:-python3}"
+
+# Colors for output
+GREEN='\033[0;32m'
+YELLOW='\033[1;33m'
+RED='\033[0;31m'
+NC='\033[0m' # No Color
+
+log_info() {
+    echo -e "${GREEN}[INFO]${NC} $1"
+}
+
+log_warn() {
+    echo -e "${YELLOW}[WARN]${NC} $1"
+}
+
+log_error() {
+    echo -e "${RED}[ERROR]${NC} $1"
+}
+
+# Check if database exists
+if [ ! -f "$DB_PATH" ]; then
+    log_error "Database not found: $DB_PATH"
+    exit 1
+fi
+
+# Create backup directory
+mkdir -p "$BACKUP_DIR"
+
+log_info "Starting backup process..."
+log_info "Database: $DB_PATH"
+log_info "Backup directory: $BACKUP_DIR"
+
+# Determine backup policy based on day of week and month
+DAY_OF_WEEK=$(date +%u)  # 1-7 (Monday-Sunday)
+DAY_OF_MONTH=$(date +%d)
+
+if [ "$DAY_OF_MONTH" == "01" ]; then
+    POLICY="monthly"
+    log_info "Running MONTHLY backup (first day of month)"
+elif [ "$DAY_OF_WEEK" == "7" ]; then
+    POLICY="weekly"
+    log_info "Running WEEKLY backup (Sunday)"
+else
+    POLICY="daily"
+    log_info "Running DAILY backup"
+fi
+
+# Run Python backup script
+$PYTHON -c "
+from pathlib import Path
+from src.backup.scheduler import BackupScheduler, BackupPolicy
+from src.backup.health_monitor import HealthMonitor
+
+# Create scheduler
+scheduler = BackupScheduler(
+    db_path='$DB_PATH',
+    backup_dir=Path('$BACKUP_DIR')
+)
+
+# Create backup
+policy = BackupPolicy.$POLICY.upper()
+metadata = scheduler.create_backup(policy, verify=True)
+print(f'Backup created: {metadata.file_path}')
+print(f'Size: {metadata.size_bytes / 1024 / 1024:.2f} MB')
+print(f'Checksum: {metadata.checksum}')
+
+# Cleanup old backups
+removed = scheduler.cleanup_old_backups()
+total_removed = sum(removed.values())
+if total_removed > 0:
+    print(f'Removed {total_removed} old backup(s)')
+
+# Health check
+monitor = HealthMonitor('$DB_PATH', Path('$BACKUP_DIR'))
+status = monitor.get_overall_status()
+print(f'System health: {status.value}')
+"
+
+if [ $? -eq 0 ]; then
+    log_info "Backup completed successfully"
+else
+    log_error "Backup failed"
+    exit 1
+fi
+
+log_info "Backup process finished"

scripts/morning_report.sh

@@ -0,0 +1,54 @@
+#!/usr/bin/env bash
+# Morning summary for overnight run logs.
+
+set -euo pipefail
+
+LOG_DIR="${LOG_DIR:-data/overnight}"
+
+if [ ! -d "$LOG_DIR" ]; then
+    echo "로그 디렉터리가 없습니다: $LOG_DIR"
+    exit 1
+fi
+
+latest_run="$(ls -1t "$LOG_DIR"/run_*.log 2>/dev/null | head -n 1 || true)"
+latest_watchdog="$(ls -1t "$LOG_DIR"/watchdog_*.log 2>/dev/null | head -n 1 || true)"
+
+if [ -z "$latest_run" ]; then
+    echo "run 로그가 없습니다: $LOG_DIR/run_*.log"
+    exit 1
+fi
+
+echo "Overnight report"
+echo "- run log: $latest_run"
+if [ -n "$latest_watchdog" ]; then
+    echo "- watchdog log: $latest_watchdog"
+fi
+
+start_line="$(head -n 1 "$latest_run" || true)"
+end_line="$(tail -n 1 "$latest_run" || true)"
+
+info_count="$(rg -c '"level": "INFO"' "$latest_run" || true)"
+warn_count="$(rg -c '"level": "WARNING"' "$latest_run" || true)"
+error_count="$(rg -c '"level": "ERROR"' "$latest_run" || true)"
+critical_count="$(rg -c '"level": "CRITICAL"' "$latest_run" || true)"
+traceback_count="$(rg -c 'Traceback' "$latest_run" || true)"
+
+echo "- start: ${start_line:-N/A}"
+echo "- end:   ${end_line:-N/A}"
+echo "- INFO: ${info_count:-0}"
+echo "- WARNING: ${warn_count:-0}"
+echo "- ERROR: ${error_count:-0}"
+echo "- CRITICAL: ${critical_count:-0}"
+echo "- Traceback: ${traceback_count:-0}"
+
+if [ -n "$latest_watchdog" ]; then
+    watchdog_errors="$(rg -c '\[ERROR\]' "$latest_watchdog" || true)"
+    echo "- watchdog ERROR: ${watchdog_errors:-0}"
+    echo ""
+    echo "최근 watchdog 로그:"
+    tail -n 5 "$latest_watchdog" || true
+fi
+
+echo ""
+echo "최근 앱 로그:"
+tail -n 20 "$latest_run" || true

scripts/restore.sh

@@ -0,0 +1,111 @@
+#!/usr/bin/env bash
+# Restore script for The Ouroboros trading system
+# Restores database from a backup file
+
+set -euo pipefail
+
+# Configuration
+DB_PATH="${DB_PATH:-data/trade_logs.db}"
+BACKUP_DIR="${BACKUP_DIR:-data/backups}"
+PYTHON="${PYTHON:-python3}"
+
+# Colors for output
+GREEN='\033[0;32m'
+YELLOW='\033[1;33m'
+RED='\033[0;31m'
+NC='\033[0m' # No Color
+
+log_info() {
+    echo -e "${GREEN}[INFO]${NC} $1"
+}
+
+log_warn() {
+    echo -e "${YELLOW}[WARN]${NC} $1"
+}
+
+log_error() {
+    echo -e "${RED}[ERROR]${NC} $1"
+}
+
+# Check if backup directory exists
+if [ ! -d "$BACKUP_DIR" ]; then
+    log_error "Backup directory not found: $BACKUP_DIR"
+    exit 1
+fi
+
+log_info "Available backups:"
+log_info "=================="
+
+# List available backups
+$PYTHON -c "
+from pathlib import Path
+from src.backup.scheduler import BackupScheduler
+
+scheduler = BackupScheduler(
+    db_path='$DB_PATH',
+    backup_dir=Path('$BACKUP_DIR')
+)
+
+backups = scheduler.list_backups()
+
+if not backups:
+    print('No backups found.')
+    exit(1)
+
+for i, backup in enumerate(backups, 1):
+    size_mb = backup.size_bytes / 1024 / 1024
+    print(f'{i}. [{backup.policy.value.upper()}] {backup.file_path.name}')
+    print(f'   Date: {backup.timestamp.strftime(\"%Y-%m-%d %H:%M:%S UTC\")}')
+    print(f'   Size: {size_mb:.2f} MB')
+    print()
+"
+
+# Ask user to select backup
+echo ""
+read -p "Enter backup number to restore (or 'q' to quit): " BACKUP_NUM
+
+if [ "$BACKUP_NUM" == "q" ]; then
+    log_info "Restore cancelled"
+    exit 0
+fi
+
+# Confirm restoration
+log_warn "WARNING: This will replace the current database!"
+log_warn "Current database will be backed up to: ${DB_PATH}.before_restore"
+read -p "Are you sure you want to continue? (yes/no): " CONFIRM
+
+if [ "$CONFIRM" != "yes" ]; then
+    log_info "Restore cancelled"
+    exit 0
+fi
+
+# Perform restoration
+$PYTHON -c "
+from pathlib import Path
+from src.backup.scheduler import BackupScheduler
+
+scheduler = BackupScheduler(
+    db_path='$DB_PATH',
+    backup_dir=Path('$BACKUP_DIR')
+)
+
+backups = scheduler.list_backups()
+backup_index = int('$BACKUP_NUM') - 1
+
+if backup_index < 0 or backup_index >= len(backups):
+    print('Invalid backup number')
+    exit(1)
+
+selected = backups[backup_index]
+print(f'Restoring: {selected.file_path.name}')
+
+scheduler.restore_backup(selected, verify=True)
+print('Restore completed successfully')
+"
+
+if [ $? -eq 0 ]; then
+    log_info "Database restored successfully"
+else
+    log_error "Restore failed"
+    exit 1
+fi

scripts/run_overnight.sh

@@ -0,0 +1,87 @@
+#!/usr/bin/env bash
+# Start The Ouroboros overnight with logs and watchdog.
+
+set -euo pipefail
+
+LOG_DIR="${LOG_DIR:-data/overnight}"
+CHECK_INTERVAL="${CHECK_INTERVAL:-30}"
+TMUX_AUTO="${TMUX_AUTO:-true}"
+TMUX_ATTACH="${TMUX_ATTACH:-true}"
+TMUX_SESSION_PREFIX="${TMUX_SESSION_PREFIX:-ouroboros_overnight}"
+
+if [ -z "${APP_CMD:-}" ]; then
+    if [ -x ".venv/bin/python" ]; then
+        PYTHON_BIN=".venv/bin/python"
+    elif command -v python3 >/dev/null 2>&1; then
+        PYTHON_BIN="python3"
+    elif command -v python >/dev/null 2>&1; then
+        PYTHON_BIN="python"
+    else
+        echo ".venv/bin/python 또는 python3/python 실행 파일을 찾을 수 없습니다."
+        exit 1
+    fi
+
+    dashboard_port="${DASHBOARD_PORT:-8080}"
+
+    APP_CMD="DASHBOARD_PORT=$dashboard_port $PYTHON_BIN -m src.main --mode=live --dashboard"
+fi
+
+mkdir -p "$LOG_DIR"
+
+timestamp="$(date +"%Y%m%d_%H%M%S")"
+RUN_LOG="$LOG_DIR/run_${timestamp}.log"
+WATCHDOG_LOG="$LOG_DIR/watchdog_${timestamp}.log"
+PID_FILE="$LOG_DIR/app.pid"
+WATCHDOG_PID_FILE="$LOG_DIR/watchdog.pid"
+
+if [ -f "$PID_FILE" ]; then
+    old_pid="$(cat "$PID_FILE" || true)"
+    if [ -n "$old_pid" ] && kill -0 "$old_pid" 2>/dev/null; then
+        echo "앱이 이미 실행 중입니다. pid=$old_pid"
+        exit 1
+    fi
+fi
+
+echo "[$(date -u +"%Y-%m-%dT%H:%M:%SZ")] starting: $APP_CMD" | tee -a "$RUN_LOG"
+nohup bash -lc "$APP_CMD" >>"$RUN_LOG" 2>&1 &
+app_pid=$!
+echo "$app_pid" > "$PID_FILE"
+
+echo "[$(date -u +"%Y-%m-%dT%H:%M:%SZ")] app pid=$app_pid" | tee -a "$RUN_LOG"
+
+nohup env PID_FILE="$PID_FILE" LOG_FILE="$WATCHDOG_LOG" CHECK_INTERVAL="$CHECK_INTERVAL" \
+    bash scripts/watchdog.sh >/dev/null 2>&1 &
+watchdog_pid=$!
+echo "$watchdog_pid" > "$WATCHDOG_PID_FILE"
+
+cat <<EOF
+시작 완료
+- app pid: $app_pid
+- watchdog pid: $watchdog_pid
+- app log: $RUN_LOG
+- watchdog log: $WATCHDOG_LOG
+
+실시간 확인:
+tail -f "$RUN_LOG"
+tail -f "$WATCHDOG_LOG"
+EOF
+
+if [ "$TMUX_AUTO" = "true" ]; then
+    if ! command -v tmux >/dev/null 2>&1; then
+        echo "tmux를 찾지 못해 자동 세션 생성은 건너뜁니다."
+        exit 0
+    fi
+
+    session_name="${TMUX_SESSION_PREFIX}_${timestamp}"
+    window_name="overnight"
+    tmux new-session -d -s "$session_name" -n "$window_name" "tail -f '$RUN_LOG'"
+    tmux split-window -t "${session_name}:${window_name}" -v "tail -f '$WATCHDOG_LOG'"
+    tmux select-layout -t "${session_name}:${window_name}" even-vertical
+
+    echo "tmux session 생성: $session_name"
+    echo "수동 접속: tmux attach -t $session_name"
+
+    if [ -z "${TMUX:-}" ] && [ "$TMUX_ATTACH" = "true" ]; then
+        tmux attach -t "$session_name"
+    fi
+fi

scripts/stop_overnight.sh

@@ -0,0 +1,76 @@
+#!/usr/bin/env bash
+# Stop The Ouroboros overnight app/watchdog/tmux session.
+
+set -euo pipefail
+
+LOG_DIR="${LOG_DIR:-data/overnight}"
+PID_FILE="$LOG_DIR/app.pid"
+WATCHDOG_PID_FILE="$LOG_DIR/watchdog.pid"
+TMUX_SESSION_PREFIX="${TMUX_SESSION_PREFIX:-ouroboros_overnight}"
+KILL_TIMEOUT="${KILL_TIMEOUT:-5}"
+
+stop_pid() {
+    local name="$1"
+    local pid="$2"
+
+    if [ -z "$pid" ]; then
+        echo "$name PID가 비어 있습니다."
+        return 1
+    fi
+
+    if ! kill -0 "$pid" 2>/dev/null; then
+        echo "$name 프로세스가 이미 종료됨 (pid=$pid)"
+        return 0
+    fi
+
+    kill "$pid" 2>/dev/null || true
+    for _ in $(seq 1 "$KILL_TIMEOUT"); do
+        if ! kill -0 "$pid" 2>/dev/null; then
+            echo "$name 종료됨 (pid=$pid)"
+            return 0
+        fi
+        sleep 1
+    done
+
+    kill -9 "$pid" 2>/dev/null || true
+    if ! kill -0 "$pid" 2>/dev/null; then
+        echo "$name 강제 종료됨 (pid=$pid)"
+        return 0
+    fi
+
+    echo "$name 종료 실패 (pid=$pid)"
+    return 1
+}
+
+status=0
+
+if [ -f "$WATCHDOG_PID_FILE" ]; then
+    watchdog_pid="$(cat "$WATCHDOG_PID_FILE" || true)"
+    stop_pid "watchdog" "$watchdog_pid" || status=1
+    rm -f "$WATCHDOG_PID_FILE"
+else
+    echo "watchdog pid 파일 없음: $WATCHDOG_PID_FILE"
+fi
+
+if [ -f "$PID_FILE" ]; then
+    app_pid="$(cat "$PID_FILE" || true)"
+    stop_pid "app" "$app_pid" || status=1
+    rm -f "$PID_FILE"
+else
+    echo "app pid 파일 없음: $PID_FILE"
+fi
+
+if command -v tmux >/dev/null 2>&1; then
+    sessions="$(tmux ls 2>/dev/null | awk -F: -v p="$TMUX_SESSION_PREFIX" '$1 ~ "^" p "_" {print $1}')"
+    if [ -n "$sessions" ]; then
+        while IFS= read -r s; do
+            [ -z "$s" ] && continue
+            tmux kill-session -t "$s" 2>/dev/null || true
+            echo "tmux 세션 종료: $s"
+        done <<< "$sessions"
+    else
+        echo "종료할 tmux 세션 없음 (prefix=${TMUX_SESSION_PREFIX}_)"
+    fi
+fi
+
+exit "$status"

scripts/validate_ouroboros_docs.py

@@ -0,0 +1,140 @@
+#!/usr/bin/env python3
+"""Validate Ouroboros planning docs for metadata, links, and ID consistency."""
+
+from __future__ import annotations
+
+import re
+import sys
+from pathlib import Path
+
+DOC_DIR = Path("docs/ouroboros")
+META_PATTERN = re.compile(
+    r"<!--\n"
+    r"Doc-ID: (?P<doc_id>[^\n]+)\n"
+    r"Version: (?P<version>[^\n]+)\n"
+    r"Status: (?P<status>[^\n]+)\n"
+    r"Owner: (?P<owner>[^\n]+)\n"
+    r"Updated: (?P<updated>\d{4}-\d{2}-\d{2})\n"
+    r"-->",
+    re.MULTILINE,
+)
+ID_PATTERN = re.compile(r"\b(?:REQ|RULE|TASK|TEST|DOC)-[A-Z0-9-]+-\d{3}\b")
+DEF_PATTERN = re.compile(r"^-\s+`(?P<id>(?:REQ|RULE|TASK|TEST|DOC)-[A-Z0-9-]+-\d{3})`", re.MULTILINE)
+LINK_PATTERN = re.compile(r"\[[^\]]+\]\((?P<link>[^)]+)\)")
+LINE_DEF_PATTERN = re.compile(r"^-\s+`(?P<id>(?:REQ|RULE|TASK|TEST|DOC)-[A-Z0-9-]+-\d{3})`.*$", re.MULTILINE)
+
+
+def iter_docs() -> list[Path]:
+    return sorted([p for p in DOC_DIR.glob("*.md") if p.is_file()])
+
+
+def validate_metadata(path: Path, text: str, errors: list[str], doc_ids: dict[str, Path]) -> None:
+    match = META_PATTERN.search(text)
+    if not match:
+        errors.append(f"{path}: missing or malformed metadata block")
+        return
+    doc_id = match.group("doc_id").strip()
+    if doc_id in doc_ids:
+        errors.append(f"{path}: duplicate Doc-ID {doc_id} (already in {doc_ids[doc_id]})")
+    else:
+        doc_ids[doc_id] = path
+
+
+def validate_links(path: Path, text: str, errors: list[str]) -> None:
+    for m in LINK_PATTERN.finditer(text):
+        link = m.group("link").strip()
+        if not link or link.startswith("http") or link.startswith("#"):
+            continue
+        if link.startswith("/"):
+            target = Path(link)
+        else:
+            target = (path.parent / link).resolve()
+        if not target.exists():
+            errors.append(f"{path}: broken link -> {link}")
+
+
+def collect_ids(path: Path, text: str, defs: dict[str, Path], refs: dict[str, set[Path]]) -> None:
+    for m in DEF_PATTERN.finditer(text):
+        defs[m.group("id")] = path
+    for m in ID_PATTERN.finditer(text):
+        idv = m.group(0)
+        refs.setdefault(idv, set()).add(path)
+
+
+def collect_req_traceability(text: str, req_to_task: dict[str, set[str]], req_to_test: dict[str, set[str]]) -> None:
+    for m in LINE_DEF_PATTERN.finditer(text):
+        line = m.group(0)
+        item_id = m.group("id")
+        req_ids = [rid for rid in ID_PATTERN.findall(line) if rid.startswith("REQ-")]
+        if item_id.startswith("TASK-"):
+            for req_id in req_ids:
+                req_to_task.setdefault(req_id, set()).add(item_id)
+        if item_id.startswith("TEST-"):
+            for req_id in req_ids:
+                req_to_test.setdefault(req_id, set()).add(item_id)
+
+
+def main() -> int:
+    if not DOC_DIR.exists():
+        print(f"ERROR: missing directory {DOC_DIR}")
+        return 1
+
+    docs = iter_docs()
+    if not docs:
+        print(f"ERROR: no markdown docs found in {DOC_DIR}")
+        return 1
+
+    errors: list[str] = []
+    doc_ids: dict[str, Path] = {}
+    defs: dict[str, Path] = {}
+    refs: dict[str, set[Path]] = {}
+    req_to_task: dict[str, set[str]] = {}
+    req_to_test: dict[str, set[str]] = {}
+
+    for path in docs:
+        text = path.read_text(encoding="utf-8")
+        validate_metadata(path, text, errors, doc_ids)
+        validate_links(path, text, errors)
+        collect_ids(path, text, defs, refs)
+        collect_req_traceability(text, req_to_task, req_to_test)
+
+    for idv, where_used in sorted(refs.items()):
+        if idv.startswith("DOC-"):
+            continue
+        if idv not in defs:
+            files = ", ".join(str(p) for p in sorted(where_used))
+            errors.append(f"undefined ID {idv}, used in: {files}")
+
+    for idv in sorted(defs):
+        if not idv.startswith("REQ-"):
+            continue
+        if idv not in req_to_task:
+            errors.append(f"REQ without TASK mapping: {idv}")
+        if idv not in req_to_test:
+            errors.append(f"REQ without TEST mapping: {idv}")
+
+    warnings: list[str] = []
+    for idv, where_def in sorted(defs.items()):
+        if len(refs.get(idv, set())) <= 1 and (idv.startswith("REQ-") or idv.startswith("RULE-")):
+            warnings.append(f"orphan ID {idv} defined in {where_def} (not referenced elsewhere)")
+
+    if errors:
+        print("[FAIL] Ouroboros docs validation failed")
+        for err in errors:
+            print(f"- {err}")
+        return 1
+
+    print(f"[OK] validated {len(docs)} docs in {DOC_DIR}")
+    print(f"[OK] unique Doc-ID: {len(doc_ids)}")
+    print(f"[OK] definitions: {len(defs)}, references: {len(refs)}")
+    print(f"[OK] req->task mappings: {len(req_to_task)}")
+    print(f"[OK] req->test mappings: {len(req_to_test)}")
+    if warnings:
+        print(f"[WARN] orphan IDs: {len(warnings)}")
+        for w in warnings:
+            print(f"- {w}")
+    return 0
+
+
+if __name__ == "__main__":
+    sys.exit(main())

scripts/watchdog.sh

@@ -0,0 +1,42 @@
+#!/usr/bin/env bash
+# Simple watchdog for The Ouroboros process.
+
+set -euo pipefail
+
+PID_FILE="${PID_FILE:-data/overnight/app.pid}"
+LOG_FILE="${LOG_FILE:-data/overnight/watchdog.log}"
+CHECK_INTERVAL="${CHECK_INTERVAL:-30}"
+STATUS_EVERY="${STATUS_EVERY:-10}"
+
+mkdir -p "$(dirname "$LOG_FILE")"
+
+log() {
+    printf '%s %s\n' "$(date -u +"%Y-%m-%dT%H:%M:%SZ")" "$1" | tee -a "$LOG_FILE"
+}
+
+if [ ! -f "$PID_FILE" ]; then
+    log "[ERROR] pid file not found: $PID_FILE"
+    exit 1
+fi
+
+PID="$(cat "$PID_FILE")"
+if [ -z "$PID" ]; then
+    log "[ERROR] pid file is empty: $PID_FILE"
+    exit 1
+fi
+
+log "[INFO] watchdog started (pid=$PID, interval=${CHECK_INTERVAL}s)"
+
+count=0
+while true; do
+    if kill -0 "$PID" 2>/dev/null; then
+        count=$((count + 1))
+        if [ $((count % STATUS_EVERY)) -eq 0 ]; then
+            log "[INFO] process alive (pid=$PID)"
+        fi
+    else
+        log "[ERROR] process stopped (pid=$PID)"
+        exit 1
+    fi
+    sleep "$CHECK_INTERVAL"
+done

src/analysis/__init__.py

@@ -0,0 +1,9 @@
+"""Technical analysis and market scanning modules."""
+
+from __future__ import annotations
+
+from src.analysis.scanner import MarketScanner
+from src.analysis.smart_scanner import ScanCandidate, SmartVolatilityScanner
+from src.analysis.volatility import VolatilityAnalyzer
+
+__all__ = ["VolatilityAnalyzer", "MarketScanner", "SmartVolatilityScanner", "ScanCandidate"]

src/analysis/scanner.py

@@ -0,0 +1,244 @@
+"""Real-time market scanner for detecting high-momentum stocks.
+
+Scans all available stocks in a market and ranks by volatility/momentum score.
+"""
+
+from __future__ import annotations
+
+import asyncio
+import logging
+from dataclasses import dataclass
+from typing import Any
+
+from src.analysis.volatility import VolatilityAnalyzer, VolatilityMetrics
+from src.broker.kis_api import KISBroker
+from src.broker.overseas import OverseasBroker
+from src.context.layer import ContextLayer
+from src.context.store import ContextStore
+from src.markets.schedule import MarketInfo
+
+logger = logging.getLogger(__name__)
+
+
+@dataclass
+class ScanResult:
+    """Result from a market scan."""
+
+    market_code: str
+    timestamp: str
+    total_scanned: int
+    top_movers: list[VolatilityMetrics]
+    breakouts: list[str]  # Stock codes with breakout patterns
+    breakdowns: list[str]  # Stock codes with breakdown patterns
+
+
+class MarketScanner:
+    """Scans markets for high-volatility, high-momentum stocks."""
+
+    def __init__(
+        self,
+        broker: KISBroker,
+        overseas_broker: OverseasBroker,
+        volatility_analyzer: VolatilityAnalyzer,
+        context_store: ContextStore,
+        top_n: int = 5,
+        max_concurrent_scans: int = 1,
+    ) -> None:
+        """Initialize the market scanner.
+
+        Args:
+            broker: KIS broker instance for domestic market
+            overseas_broker: Overseas broker instance
+            volatility_analyzer: Volatility analyzer instance
+            context_store: Context store for L7 real-time data
+            top_n: Number of top movers to return per market (default 5)
+            max_concurrent_scans: Max concurrent stock scans (default 1, fully serialized)
+        """
+        self.broker = broker
+        self.overseas_broker = overseas_broker
+        self.analyzer = volatility_analyzer
+        self.context_store = context_store
+        self.top_n = top_n
+        self._scan_semaphore = asyncio.Semaphore(max_concurrent_scans)
+
+    async def scan_stock(
+        self,
+        stock_code: str,
+        market: MarketInfo,
+    ) -> VolatilityMetrics | None:
+        """Scan a single stock for volatility metrics.
+
+        Args:
+            stock_code: Stock code to scan
+            market: Market information
+
+        Returns:
+            VolatilityMetrics if successful, None on error
+        """
+        try:
+            if market.is_domestic:
+                orderbook = await self.broker.get_orderbook(stock_code)
+            else:
+                # For overseas, we need to adapt the price data structure
+                price_data = await self.overseas_broker.get_overseas_price(
+                    market.exchange_code, stock_code
+                )
+                # Convert to orderbook-like structure
+                orderbook = {
+                    "output1": {
+                        "stck_prpr": price_data.get("output", {}).get("last", "0") or "0",
+                        "acml_vol": price_data.get("output", {}).get("tvol", "0") or "0",
+                    }
+                }
+
+            # For now, use empty price history (would need real historical data)
+            # In production, this would fetch from a time-series database or API
+            price_history: dict[str, Any] = {
+                "high": [],
+                "low": [],
+                "close": [],
+                "volume": [],
+            }
+
+            metrics = self.analyzer.analyze(stock_code, orderbook, price_history)
+
+            # Store in L7 real-time layer
+            from datetime import UTC, datetime
+            timeframe = datetime.now(UTC).isoformat()
+            self.context_store.set_context(
+                ContextLayer.L7_REALTIME,
+                timeframe,
+                f"volatility_{market.code}_{stock_code}",
+                {
+                    "price": metrics.current_price,
+                    "atr": metrics.atr,
+                    "price_change_1m": metrics.price_change_1m,
+                    "volume_surge": metrics.volume_surge,
+                    "momentum_score": metrics.momentum_score,
+                },
+            )
+
+            return metrics
+
+        except Exception as exc:
+            logger.warning("Failed to scan %s (%s): %s", stock_code, market.code, exc)
+            return None
+
+    async def scan_market(
+        self,
+        market: MarketInfo,
+        stock_codes: list[str],
+    ) -> ScanResult:
+        """Scan all stocks in a market and rank by momentum.
+
+        Args:
+            market: Market to scan
+            stock_codes: List of stock codes to scan
+
+        Returns:
+            ScanResult with ranked stocks
+        """
+        from datetime import UTC, datetime
+
+        logger.info("Scanning %s market (%d stocks)", market.name, len(stock_codes))
+
+        # Scan stocks with bounded concurrency to prevent API rate limit burst
+        async def _bounded_scan(code: str) -> VolatilityMetrics | None:
+            async with self._scan_semaphore:
+                return await self.scan_stock(code, market)
+
+        tasks = [_bounded_scan(code) for code in stock_codes]
+        results = await asyncio.gather(*tasks)
+
+        # Filter out failures and sort by momentum score
+        valid_metrics = [m for m in results if m is not None]
+        valid_metrics.sort(key=lambda m: m.momentum_score, reverse=True)
+
+        # Get top N movers
+        top_movers = valid_metrics[: self.top_n]
+
+        # Detect breakouts and breakdowns
+        breakouts = [
+            m.stock_code for m in valid_metrics if self.analyzer.is_breakout(m)
+        ]
+        breakdowns = [
+            m.stock_code for m in valid_metrics if self.analyzer.is_breakdown(m)
+        ]
+
+        logger.info(
+            "%s scan complete: %d scanned, top momentum=%.1f, %d breakouts, %d breakdowns",
+            market.name,
+            len(valid_metrics),
+            top_movers[0].momentum_score if top_movers else 0.0,
+            len(breakouts),
+            len(breakdowns),
+        )
+
+        # Store scan results in L7
+        timeframe = datetime.now(UTC).isoformat()
+        self.context_store.set_context(
+            ContextLayer.L7_REALTIME,
+            timeframe,
+            f"scan_result_{market.code}",
+            {
+                "total_scanned": len(valid_metrics),
+                "top_movers": [m.stock_code for m in top_movers],
+                "breakouts": breakouts,
+                "breakdowns": breakdowns,
+            },
+        )
+
+        return ScanResult(
+            market_code=market.code,
+            timestamp=timeframe,
+            total_scanned=len(valid_metrics),
+            top_movers=top_movers,
+            breakouts=breakouts,
+            breakdowns=breakdowns,
+        )
+
+    def get_updated_watchlist(
+        self,
+        current_watchlist: list[str],
+        scan_result: ScanResult,
+        max_replacements: int = 2,
+    ) -> list[str]:
+        """Update watchlist by replacing laggards with leaders.
+
+        Args:
+            current_watchlist: Current watchlist
+            scan_result: Recent scan result
+            max_replacements: Maximum stocks to replace per scan
+
+        Returns:
+            Updated watchlist with leaders
+        """
+        # Keep stocks that are in top movers
+        top_codes = [m.stock_code for m in scan_result.top_movers]
+        keepers = [code for code in current_watchlist if code in top_codes]
+
+        # Add new leaders not in current watchlist
+        new_leaders = [code for code in top_codes if code not in current_watchlist]
+
+        # Limit replacements
+        new_leaders = new_leaders[:max_replacements]
+
+        # Create updated watchlist
+        updated = keepers + new_leaders
+
+        # If we removed too many, backfill from current watchlist
+        if len(updated) < len(current_watchlist):
+            backfill = [
+                code for code in current_watchlist
+                if code not in updated
+            ][: len(current_watchlist) - len(updated)]
+            updated.extend(backfill)
+
+        logger.info(
+            "Watchlist updated: %d kept, %d new leaders, %d total",
+            len(keepers),
+            len(new_leaders),
+            len(updated),
+        )
+
+        return updated

src/analysis/smart_scanner.py

@@ -0,0 +1,449 @@
+"""Smart Volatility Scanner with volatility-first market ranking logic."""
+
+from __future__ import annotations
+
+import logging
+from dataclasses import dataclass
+from typing import Any
+
+from src.analysis.volatility import VolatilityAnalyzer
+from src.broker.kis_api import KISBroker
+from src.broker.overseas import OverseasBroker
+from src.config import Settings
+from src.markets.schedule import MarketInfo
+
+logger = logging.getLogger(__name__)
+
+
+@dataclass
+class ScanCandidate:
+    """A qualified candidate from the smart scanner."""
+
+    stock_code: str
+    name: str
+    price: float
+    volume: float
+    volume_ratio: float  # Current volume / previous day volume
+    rsi: float
+    signal: str  # "oversold" or "momentum"
+    score: float  # Composite score for ranking
+
+
+class SmartVolatilityScanner:
+    """Scans market rankings and applies volatility-first filters.
+
+    Flow:
+    1. Fetch fluctuation rankings as primary universe
+    2. Fetch volume rankings for liquidity bonus
+    3. Score by volatility first, liquidity second
+    4. Return top N qualified candidates
+    """
+
+    def __init__(
+        self,
+        broker: KISBroker,
+        overseas_broker: OverseasBroker | None,
+        volatility_analyzer: VolatilityAnalyzer,
+        settings: Settings,
+    ) -> None:
+        """Initialize the smart scanner.
+
+        Args:
+            broker: KIS broker for API calls
+            volatility_analyzer: Analyzer for RSI calculation
+            settings: Application settings
+        """
+        self.broker = broker
+        self.overseas_broker = overseas_broker
+        self.analyzer = volatility_analyzer
+        self.settings = settings
+
+        # Extract scanner settings
+        self.rsi_oversold = settings.RSI_OVERSOLD_THRESHOLD
+        self.rsi_momentum = settings.RSI_MOMENTUM_THRESHOLD
+        self.vol_multiplier = settings.VOL_MULTIPLIER
+        self.top_n = settings.SCANNER_TOP_N
+
+    async def scan(
+        self,
+        market: MarketInfo | None = None,
+        fallback_stocks: list[str] | None = None,
+    ) -> list[ScanCandidate]:
+        """Execute smart scan and return qualified candidates.
+
+        Args:
+            market: Target market info (domestic vs overseas behavior)
+            fallback_stocks: Stock codes to use if ranking API fails
+
+        Returns:
+            List of ScanCandidate, sorted by score, up to top_n items
+        """
+        if market and not market.is_domestic:
+            return await self._scan_overseas(market, fallback_stocks)
+
+        return await self._scan_domestic(fallback_stocks)
+
+    async def _scan_domestic(
+        self,
+        fallback_stocks: list[str] | None = None,
+    ) -> list[ScanCandidate]:
+        """Scan domestic market using volatility-first ranking + liquidity bonus."""
+        # 1) Primary universe from fluctuation ranking.
+        try:
+            fluct_rows = await self.broker.fetch_market_rankings(
+                ranking_type="fluctuation",
+                limit=50,
+            )
+        except ConnectionError as exc:
+            logger.warning("Domestic fluctuation ranking failed: %s", exc)
+            fluct_rows = []
+
+        # 2) Liquidity bonus from volume ranking.
+        try:
+            volume_rows = await self.broker.fetch_market_rankings(
+                ranking_type="volume",
+                limit=50,
+            )
+        except ConnectionError as exc:
+            logger.warning("Domestic volume ranking failed: %s", exc)
+            volume_rows = []
+
+        if not fluct_rows and fallback_stocks:
+            logger.info(
+                "Domestic ranking unavailable; using fallback symbols (%d)",
+                len(fallback_stocks),
+            )
+            fluct_rows = [
+                {
+                    "stock_code": code,
+                    "name": code,
+                    "price": 0.0,
+                    "volume": 0.0,
+                    "change_rate": 0.0,
+                    "volume_increase_rate": 0.0,
+                }
+                for code in fallback_stocks
+            ]
+
+        if not fluct_rows:
+            return []
+
+        volume_rank_bonus: dict[str, float] = {}
+        for idx, row in enumerate(volume_rows):
+            code = _extract_stock_code(row)
+            if not code:
+                continue
+            volume_rank_bonus[code] = max(0.0, 15.0 - idx * 0.3)
+
+        candidates: list[ScanCandidate] = []
+        for stock in fluct_rows:
+            stock_code = _extract_stock_code(stock)
+            if not stock_code:
+                continue
+
+            try:
+                price = _extract_last_price(stock)
+                change_rate = _extract_change_rate_pct(stock)
+                volume = _extract_volume(stock)
+
+                intraday_range_pct = 0.0
+                volume_ratio = _safe_float(stock.get("volume_increase_rate"), 0.0) / 100.0 + 1.0
+
+                # Use daily chart to refine range/volume when available.
+                daily_prices = await self.broker.get_daily_prices(stock_code, days=2)
+                if daily_prices:
+                    latest = daily_prices[-1]
+                    latest_close = _safe_float(latest.get("close"), default=price)
+                    if price <= 0:
+                        price = latest_close
+                    latest_high = _safe_float(latest.get("high"))
+                    latest_low = _safe_float(latest.get("low"))
+                    if latest_close > 0 and latest_high > 0 and latest_low > 0 and latest_high >= latest_low:
+                        intraday_range_pct = (latest_high - latest_low) / latest_close * 100.0
+                    if volume <= 0:
+                        volume = _safe_float(latest.get("volume"))
+                    if len(daily_prices) >= 2:
+                        prev_day_volume = _safe_float(daily_prices[-2].get("volume"))
+                        if prev_day_volume > 0:
+                            volume_ratio = max(volume_ratio, volume / prev_day_volume)
+
+                volatility_pct = max(abs(change_rate), intraday_range_pct)
+                if price <= 0 or volatility_pct < 0.8:
+                    continue
+
+                volatility_score = min(volatility_pct / 10.0, 1.0) * 85.0
+                liquidity_score = volume_rank_bonus.get(stock_code, 0.0)
+                score = min(100.0, volatility_score + liquidity_score)
+                signal = "momentum" if change_rate >= 0 else "oversold"
+                implied_rsi = max(0.0, min(100.0, 50.0 + (change_rate * 2.0)))
+
+                candidates.append(
+                    ScanCandidate(
+                        stock_code=stock_code,
+                        name=stock.get("name", stock_code),
+                        price=price,
+                        volume=volume,
+                        volume_ratio=max(1.0, volume_ratio, volatility_pct / 2.0),
+                        rsi=implied_rsi,
+                        signal=signal,
+                        score=score,
+                    )
+                )
+
+            except ConnectionError as exc:
+                logger.warning("Failed to analyze %s: %s", stock_code, exc)
+                continue
+            except Exception as exc:
+                logger.error("Unexpected error analyzing %s: %s", stock_code, exc)
+                continue
+
+        logger.info("Domestic ranking scan found %d candidates", len(candidates))
+        candidates.sort(key=lambda c: c.score, reverse=True)
+        return candidates[: self.top_n]
+
+    async def _scan_overseas(
+        self,
+        market: MarketInfo,
+        fallback_stocks: list[str] | None = None,
+    ) -> list[ScanCandidate]:
+        """Scan overseas symbols using ranking API first, then fallback universe."""
+        if self.overseas_broker is None:
+            logger.warning(
+                "Overseas scanner unavailable for %s: overseas broker not configured",
+                market.name,
+            )
+            return []
+
+        candidates = await self._scan_overseas_from_rankings(market)
+        if not candidates:
+            candidates = await self._scan_overseas_from_symbols(market, fallback_stocks)
+
+        candidates.sort(key=lambda c: c.score, reverse=True)
+        return candidates[: self.top_n]
+
+    async def _scan_overseas_from_rankings(
+        self,
+        market: MarketInfo,
+    ) -> list[ScanCandidate]:
+        """Build overseas candidates from ranking APIs using volatility-first scoring."""
+        assert self.overseas_broker is not None
+        try:
+            fluct_rows = await self.overseas_broker.fetch_overseas_rankings(
+                exchange_code=market.exchange_code,
+                ranking_type="fluctuation",
+                limit=50,
+            )
+        except Exception as exc:
+            logger.warning(
+                "Overseas fluctuation ranking failed for %s: %s", market.code, exc
+            )
+            fluct_rows = []
+
+        if not fluct_rows:
+            return []
+
+        volume_rank_bonus: dict[str, float] = {}
+        try:
+            volume_rows = await self.overseas_broker.fetch_overseas_rankings(
+                exchange_code=market.exchange_code,
+                ranking_type="volume",
+                limit=50,
+            )
+        except Exception as exc:
+            logger.warning(
+                "Overseas volume ranking failed for %s: %s", market.code, exc
+            )
+            volume_rows = []
+
+        for idx, row in enumerate(volume_rows):
+            code = _extract_stock_code(row)
+            if not code:
+                continue
+            # Top-ranked by traded value/volume gets higher liquidity bonus.
+            volume_rank_bonus[code] = max(0.0, 15.0 - idx * 0.3)
+
+        candidates: list[ScanCandidate] = []
+        for row in fluct_rows:
+            stock_code = _extract_stock_code(row)
+            if not stock_code:
+                continue
+
+            price = _extract_last_price(row)
+            change_rate = _extract_change_rate_pct(row)
+            volume = _extract_volume(row)
+            intraday_range_pct = _extract_intraday_range_pct(row, price)
+            volatility_pct = max(abs(change_rate), intraday_range_pct)
+
+            # Volatility-first filter (not simple gainers/value ranking).
+            if price <= 0 or volatility_pct < 0.8:
+                continue
+
+            volatility_score = min(volatility_pct / 10.0, 1.0) * 85.0
+            liquidity_score = volume_rank_bonus.get(stock_code, 0.0)
+            score = min(100.0, volatility_score + liquidity_score)
+            signal = "momentum" if change_rate >= 0 else "oversold"
+            implied_rsi = max(0.0, min(100.0, 50.0 + (change_rate * 2.0)))
+            candidates.append(
+                ScanCandidate(
+                    stock_code=stock_code,
+                    name=str(row.get("name") or row.get("ovrs_item_name") or stock_code),
+                    price=price,
+                    volume=volume,
+                    volume_ratio=max(1.0, volatility_pct / 2.0),
+                    rsi=implied_rsi,
+                    signal=signal,
+                    score=score,
+                )
+            )
+
+        if candidates:
+            logger.info(
+                "Overseas ranking scan found %d candidates for %s",
+                len(candidates),
+                market.name,
+            )
+        return candidates
+
+    async def _scan_overseas_from_symbols(
+        self,
+        market: MarketInfo,
+        symbols: list[str] | None,
+    ) -> list[ScanCandidate]:
+        """Fallback overseas scan from dynamic symbol universe."""
+        assert self.overseas_broker is not None
+        if not symbols:
+            logger.info("Overseas scanner: no symbol universe for %s", market.name)
+            return []
+
+        logger.info(
+            "Overseas scanner: scanning %d fallback symbols for %s",
+            len(symbols),
+            market.name,
+        )
+        candidates: list[ScanCandidate] = []
+        for stock_code in symbols:
+            try:
+                price_data = await self.overseas_broker.get_overseas_price(
+                    market.exchange_code, stock_code
+                )
+                output = price_data.get("output", {})
+                price = _extract_last_price(output)
+                change_rate = _extract_change_rate_pct(output)
+                volume = _extract_volume(output)
+                intraday_range_pct = _extract_intraday_range_pct(output, price)
+                volatility_pct = max(abs(change_rate), intraday_range_pct)
+
+                if price <= 0 or volatility_pct < 0.8:
+                    continue
+
+                score = min(volatility_pct / 10.0, 1.0) * 100.0
+                signal = "momentum" if change_rate >= 0 else "oversold"
+                implied_rsi = max(0.0, min(100.0, 50.0 + (change_rate * 2.0)))
+                candidates.append(
+                    ScanCandidate(
+                        stock_code=stock_code,
+                        name=stock_code,
+                        price=price,
+                        volume=volume,
+                        volume_ratio=max(1.0, volatility_pct / 2.0),
+                        rsi=implied_rsi,
+                        signal=signal,
+                        score=score,
+                    )
+                )
+            except ConnectionError as exc:
+                logger.warning("Failed to analyze overseas %s: %s", stock_code, exc)
+            except Exception as exc:
+                logger.error("Unexpected error analyzing overseas %s: %s", stock_code, exc)
+        logger.info(
+            "Overseas symbol fallback scan found %d candidates for %s",
+            len(candidates),
+            market.name,
+        )
+        return candidates
+
+    def get_stock_codes(self, candidates: list[ScanCandidate]) -> list[str]:
+        """Extract stock codes from candidates for watchlist update.
+
+        Args:
+            candidates: List of scan candidates
+
+        Returns:
+            List of stock codes
+        """
+        return [c.stock_code for c in candidates]
+
+
+def _safe_float(value: Any, default: float = 0.0) -> float:
+    """Convert arbitrary values to float safely."""
+    if value in (None, ""):
+        return default
+    try:
+        return float(value)
+    except (TypeError, ValueError):
+        return default
+
+
+def _extract_stock_code(row: dict[str, Any]) -> str:
+    """Extract normalized stock code from various API schemas."""
+    return (
+        str(
+            row.get("symb")
+            or row.get("ovrs_pdno")
+            or row.get("stock_code")
+            or row.get("pdno")
+            or ""
+        )
+        .strip()
+        .upper()
+    )
+
+
+def _extract_last_price(row: dict[str, Any]) -> float:
+    """Extract last/close-like price from API schema variants."""
+    return _safe_float(
+        row.get("last")
+        or row.get("ovrs_nmix_prpr")
+        or row.get("stck_prpr")
+        or row.get("price")
+        or row.get("close")
+    )
+
+
+def _extract_change_rate_pct(row: dict[str, Any]) -> float:
+    """Extract daily change rate (%) from API schema variants."""
+    return _safe_float(
+        row.get("rate")
+        or row.get("change_rate")
+        or row.get("prdy_ctrt")
+        or row.get("evlu_pfls_rt")
+        or row.get("chg_rt")
+    )
+
+
+def _extract_volume(row: dict[str, Any]) -> float:
+    """Extract volume/traded-amount proxy from schema variants."""
+    return _safe_float(
+        row.get("tvol") or row.get("acml_vol") or row.get("vol") or row.get("volume")
+    )
+
+
+def _extract_intraday_range_pct(row: dict[str, Any], price: float) -> float:
+    """Estimate intraday range percentage from high/low fields."""
+    if price <= 0:
+        return 0.0
+    high = _safe_float(
+        row.get("high")
+        or row.get("ovrs_hgpr")
+        or row.get("stck_hgpr")
+        or row.get("day_hgpr")
+    )
+    low = _safe_float(
+        row.get("low")
+        or row.get("ovrs_lwpr")
+        or row.get("stck_lwpr")
+        or row.get("day_lwpr")
+    )
+    if high <= 0 or low <= 0 or high < low:
+        return 0.0
+    return (high - low) / price * 100.0

src/analysis/triple_barrier.py

@@ -0,0 +1,111 @@
+"""Triple barrier labeler utilities.
+
+Implements first-touch labeling with upper/lower/time barriers.
+"""
+
+from __future__ import annotations
+
+from dataclasses import dataclass
+from typing import Literal, Sequence
+
+
+TieBreakMode = Literal["stop_first", "take_first"]
+
+
+@dataclass(frozen=True)
+class TripleBarrierSpec:
+    take_profit_pct: float
+    stop_loss_pct: float
+    max_holding_bars: int
+    tie_break: TieBreakMode = "stop_first"
+
+
+@dataclass(frozen=True)
+class TripleBarrierLabel:
+    label: int  # +1 take-profit first, -1 stop-loss first, 0 timeout
+    touched: Literal["take_profit", "stop_loss", "time"]
+    touch_bar: int
+    entry_price: float
+    upper_barrier: float
+    lower_barrier: float
+
+
+def label_with_triple_barrier(
+    *,
+    highs: Sequence[float],
+    lows: Sequence[float],
+    closes: Sequence[float],
+    entry_index: int,
+    side: int,
+    spec: TripleBarrierSpec,
+) -> TripleBarrierLabel:
+    """Label one entry using triple-barrier first-touch rule.
+
+    Args:
+        highs/lows/closes: OHLC components with identical length.
+        entry_index: Entry bar index in the sequences.
+        side: +1 for long, -1 for short.
+        spec: Barrier specification.
+    """
+    if side not in {1, -1}:
+        raise ValueError("side must be +1 or -1")
+    if len(highs) != len(lows) or len(highs) != len(closes):
+        raise ValueError("highs, lows, closes lengths must match")
+    if entry_index < 0 or entry_index >= len(closes):
+        raise IndexError("entry_index out of range")
+    if spec.max_holding_bars <= 0:
+        raise ValueError("max_holding_bars must be positive")
+
+    entry_price = float(closes[entry_index])
+    if entry_price <= 0:
+        raise ValueError("entry price must be positive")
+
+    if side == 1:
+        upper = entry_price * (1.0 + spec.take_profit_pct)
+        lower = entry_price * (1.0 - spec.stop_loss_pct)
+    else:
+        # For short side, favorable move is down.
+        upper = entry_price * (1.0 + spec.stop_loss_pct)
+        lower = entry_price * (1.0 - spec.take_profit_pct)
+
+    last_index = min(len(closes) - 1, entry_index + spec.max_holding_bars)
+    for idx in range(entry_index + 1, last_index + 1):
+        h = float(highs[idx])
+        l = float(lows[idx])
+
+        up_touch = h >= upper
+        down_touch = l <= lower
+        if not up_touch and not down_touch:
+            continue
+
+        if up_touch and down_touch:
+            if spec.tie_break == "stop_first":
+                touched = "stop_loss"
+                label = -1
+            else:
+                touched = "take_profit"
+                label = 1
+        elif up_touch:
+            touched = "take_profit" if side == 1 else "stop_loss"
+            label = 1 if side == 1 else -1
+        else:
+            touched = "stop_loss" if side == 1 else "take_profit"
+            label = -1 if side == 1 else 1
+
+        return TripleBarrierLabel(
+            label=label,
+            touched=touched,
+            touch_bar=idx,
+            entry_price=entry_price,
+            upper_barrier=upper,
+            lower_barrier=lower,
+        )
+
+    return TripleBarrierLabel(
+        label=0,
+        touched="time",
+        touch_bar=last_index,
+        entry_price=entry_price,
+        upper_barrier=upper,
+        lower_barrier=lower,
+    )

src/analysis/volatility.py

@@ -0,0 +1,373 @@
+"""Volatility and momentum analysis for stock selection.
+
+Calculates ATR, price change percentages, volume surges, and price-volume divergence.
+"""
+
+from __future__ import annotations
+
+from dataclasses import dataclass
+from typing import Any
+
+
+@dataclass
+class VolatilityMetrics:
+    """Volatility and momentum metrics for a stock."""
+
+    stock_code: str
+    current_price: float
+    atr: float  # Average True Range (14 periods)
+    price_change_1m: float  # 1-minute price change %
+    price_change_5m: float  # 5-minute price change %
+    price_change_15m: float  # 15-minute price change %
+    volume_surge: float  # Volume vs average (ratio)
+    pv_divergence: float  # Price-volume divergence score
+    momentum_score: float  # Combined momentum score (0-100)
+
+    def __repr__(self) -> str:
+        return (
+            f"VolatilityMetrics({self.stock_code}: "
+            f"price={self.current_price:.2f}, "
+            f"atr={self.atr:.2f}, "
+            f"1m={self.price_change_1m:.2f}%, "
+            f"vol_surge={self.volume_surge:.2f}x, "
+            f"momentum={self.momentum_score:.1f})"
+        )
+
+
+class VolatilityAnalyzer:
+    """Analyzes stock volatility and momentum for leader detection."""
+
+    def __init__(self, min_volume_surge: float = 2.0, min_price_change: float = 1.0) -> None:
+        """Initialize the volatility analyzer.
+
+        Args:
+            min_volume_surge: Minimum volume surge ratio (default 2x average)
+            min_price_change: Minimum price change % for breakout (default 1%)
+        """
+        self.min_volume_surge = min_volume_surge
+        self.min_price_change = min_price_change
+
+    def calculate_atr(
+        self,
+        high_prices: list[float],
+        low_prices: list[float],
+        close_prices: list[float],
+        period: int = 14,
+    ) -> float:
+        """Calculate Average True Range (ATR).
+
+        Args:
+            high_prices: List of high prices (most recent last)
+            low_prices: List of low prices (most recent last)
+            close_prices: List of close prices (most recent last)
+            period: ATR period (default 14)
+
+        Returns:
+            ATR value
+        """
+        if (
+            len(high_prices) < period + 1
+            or len(low_prices) < period + 1
+            or len(close_prices) < period + 1
+        ):
+            return 0.0
+
+        true_ranges: list[float] = []
+        for i in range(1, len(high_prices)):
+            high = high_prices[i]
+            low = low_prices[i]
+            prev_close = close_prices[i - 1]
+
+            tr = max(
+                high - low,
+                abs(high - prev_close),
+                abs(low - prev_close),
+            )
+            true_ranges.append(tr)
+
+        if len(true_ranges) < period:
+            return 0.0
+
+        # Simple Moving Average of True Range
+        recent_tr = true_ranges[-period:]
+        return sum(recent_tr) / len(recent_tr)
+
+    def calculate_price_change(
+        self, current_price: float, past_price: float
+    ) -> float:
+        """Calculate price change percentage.
+
+        Args:
+            current_price: Current price
+            past_price: Past price to compare against
+
+        Returns:
+            Price change percentage
+        """
+        if past_price == 0:
+            return 0.0
+        return ((current_price - past_price) / past_price) * 100
+
+    def calculate_volume_surge(
+        self, current_volume: float, avg_volume: float
+    ) -> float:
+        """Calculate volume surge ratio.
+
+        Args:
+            current_volume: Current volume
+            avg_volume: Average volume
+
+        Returns:
+            Volume surge ratio (current / average)
+        """
+        if avg_volume == 0:
+            return 1.0
+        return current_volume / avg_volume
+
+    def calculate_rsi(
+        self,
+        close_prices: list[float],
+        period: int = 14,
+    ) -> float:
+        """Calculate Relative Strength Index (RSI) using Wilder's smoothing.
+
+        Args:
+            close_prices: List of closing prices (oldest to newest, minimum period+1 values)
+            period: RSI period (default 14)
+
+        Returns:
+            RSI value between 0 and 100, or 50.0 (neutral) if insufficient data
+
+        Examples:
+            >>> analyzer = VolatilityAnalyzer()
+            >>> prices = [100 - i * 0.5 for i in range(20)]  # Downtrend
+            >>> rsi = analyzer.calculate_rsi(prices)
+            >>> assert rsi < 50  # Oversold territory
+        """
+        if len(close_prices) < period + 1:
+            return 50.0  # Neutral RSI if insufficient data
+
+        # Calculate price changes
+        changes = [close_prices[i] - close_prices[i - 1] for i in range(1, len(close_prices))]
+
+        # Separate gains and losses
+        gains = [max(0.0, change) for change in changes]
+        losses = [max(0.0, -change) for change in changes]
+
+        # Calculate initial average gain/loss (simple average for first period)
+        avg_gain = sum(gains[:period]) / period
+        avg_loss = sum(losses[:period]) / period
+
+        # Apply Wilder's smoothing for remaining periods
+        for i in range(period, len(changes)):
+            avg_gain = (avg_gain * (period - 1) + gains[i]) / period
+            avg_loss = (avg_loss * (period - 1) + losses[i]) / period
+
+        # Calculate RS and RSI
+        if avg_loss == 0:
+            return 100.0  # All gains, maximum RSI
+
+        rs = avg_gain / avg_loss
+        rsi = 100 - (100 / (1 + rs))
+
+        return rsi
+
+    def calculate_pv_divergence(
+        self,
+        price_change: float,
+        volume_surge: float,
+    ) -> float:
+        """Calculate price-volume divergence score.
+
+        Positive divergence: Price up + Volume up = bullish
+        Negative divergence: Price up + Volume down = bearish
+        Neutral: Price/volume move together moderately
+
+        Args:
+            price_change: Price change percentage
+            volume_surge: Volume surge ratio
+
+        Returns:
+            Divergence score (-100 to +100)
+        """
+        # Normalize volume surge to -1 to +1 scale (1.0 = neutral)
+        volume_signal = (volume_surge - 1.0) * 10  # Scale for sensitivity
+
+        # Calculate divergence
+        # Positive: price and volume move in same direction
+        # Negative: price and volume move in opposite directions
+        if price_change > 0 and volume_surge > 1.0:
+            # Bullish: price up, volume up
+            return min(100.0, price_change * volume_signal)
+        elif price_change < 0 and volume_surge < 1.0:
+            # Bearish confirmation: price down, volume down
+            return max(-100.0, price_change * volume_signal)
+        elif price_change > 0 and volume_surge < 1.0:
+            # Bearish divergence: price up but volume low (weak rally)
+            return -abs(price_change) * 0.5
+        elif price_change < 0 and volume_surge > 1.0:
+            # Selling pressure: price down, volume up
+            return price_change * volume_signal
+        else:
+            return 0.0
+
+    def calculate_momentum_score(
+        self,
+        price_change_1m: float,
+        price_change_5m: float,
+        price_change_15m: float,
+        volume_surge: float,
+        atr: float,
+        current_price: float,
+    ) -> float:
+        """Calculate combined momentum score (0-100).
+
+        Weights:
+        - 1m change: 40%
+        - 5m change: 30%
+        - 15m change: 20%
+        - Volume surge: 10%
+
+        Args:
+            price_change_1m: 1-minute price change %
+            price_change_5m: 5-minute price change %
+            price_change_15m: 15-minute price change %
+            volume_surge: Volume surge ratio
+            atr: Average True Range
+            current_price: Current price
+
+        Returns:
+            Momentum score (0-100)
+        """
+        # Weight recent changes more heavily
+        weighted_change = (
+            price_change_1m * 0.4 +
+            price_change_5m * 0.3 +
+            price_change_15m * 0.2
+        )
+
+        # Volume contribution (normalized to 0-10 scale)
+        volume_contribution = min(10.0, (volume_surge - 1.0) * 5.0)
+
+        # Volatility bonus: higher ATR = higher potential (normalized)
+        volatility_bonus = 0.0
+        if current_price > 0:
+            atr_pct = (atr / current_price) * 100
+            volatility_bonus = min(10.0, atr_pct)
+
+        # Combine scores
+        raw_score = weighted_change + volume_contribution + volatility_bonus
+
+        # Normalize to 0-100 scale
+        # Assume typical momentum range is -10 to +30
+        normalized = ((raw_score + 10) / 40) * 100
+
+        return max(0.0, min(100.0, normalized))
+
+    def analyze(
+        self,
+        stock_code: str,
+        orderbook_data: dict[str, Any],
+        price_history: dict[str, Any],
+    ) -> VolatilityMetrics:
+        """Analyze volatility and momentum for a stock.
+
+        Args:
+            stock_code: Stock code
+            orderbook_data: Current orderbook/quote data
+            price_history: Historical price and volume data
+
+        Returns:
+            VolatilityMetrics with calculated indicators
+        """
+        # Extract current data from orderbook
+        output1 = orderbook_data.get("output1", {})
+        current_price = float(output1.get("stck_prpr", 0))
+        current_volume = float(output1.get("acml_vol", 0))
+
+        # Extract historical data
+        high_prices = price_history.get("high", [])
+        low_prices = price_history.get("low", [])
+        close_prices = price_history.get("close", [])
+        volumes = price_history.get("volume", [])
+
+        # Calculate ATR
+        atr = self.calculate_atr(high_prices, low_prices, close_prices)
+
+        # Calculate price changes (use historical data if available)
+        price_change_1m = 0.0
+        price_change_5m = 0.0
+        price_change_15m = 0.0
+
+        if len(close_prices) > 0:
+            if len(close_prices) >= 1:
+                price_change_1m = self.calculate_price_change(
+                    current_price, close_prices[-1]
+                )
+            if len(close_prices) >= 5:
+                price_change_5m = self.calculate_price_change(
+                    current_price, close_prices[-5]
+                )
+            if len(close_prices) >= 15:
+                price_change_15m = self.calculate_price_change(
+                    current_price, close_prices[-15]
+                )
+
+        # Calculate volume surge
+        avg_volume = sum(volumes) / len(volumes) if volumes else current_volume
+        volume_surge = self.calculate_volume_surge(current_volume, avg_volume)
+
+        # Calculate price-volume divergence
+        pv_divergence = self.calculate_pv_divergence(price_change_1m, volume_surge)
+
+        # Calculate momentum score
+        momentum_score = self.calculate_momentum_score(
+            price_change_1m,
+            price_change_5m,
+            price_change_15m,
+            volume_surge,
+            atr,
+            current_price,
+        )
+
+        return VolatilityMetrics(
+            stock_code=stock_code,
+            current_price=current_price,
+            atr=atr,
+            price_change_1m=price_change_1m,
+            price_change_5m=price_change_5m,
+            price_change_15m=price_change_15m,
+            volume_surge=volume_surge,
+            pv_divergence=pv_divergence,
+            momentum_score=momentum_score,
+        )
+
+    def is_breakout(self, metrics: VolatilityMetrics) -> bool:
+        """Determine if a stock is experiencing a breakout.
+
+        Args:
+            metrics: Volatility metrics for the stock
+
+        Returns:
+            True if breakout conditions are met
+        """
+        return (
+            metrics.price_change_1m >= self.min_price_change
+            and metrics.volume_surge >= self.min_volume_surge
+            and metrics.pv_divergence > 0  # Bullish divergence
+        )
+
+    def is_breakdown(self, metrics: VolatilityMetrics) -> bool:
+        """Determine if a stock is experiencing a breakdown.
+
+        Args:
+            metrics: Volatility metrics for the stock
+
+        Returns:
+            True if breakdown conditions are met
+        """
+        return (
+            metrics.price_change_1m <= -self.min_price_change
+            and metrics.volume_surge >= self.min_volume_surge
+            and metrics.pv_divergence < 0  # Bearish divergence
+        )

src/analysis/walk_forward_split.py

@@ -0,0 +1,74 @@
+"""Walk-forward splitter with purge/embargo controls."""
+
+from __future__ import annotations
+
+from dataclasses import dataclass
+
+
+@dataclass(frozen=True)
+class WalkForwardFold:
+    train_indices: list[int]
+    test_indices: list[int]
+
+    @property
+    def train_size(self) -> int:
+        return len(self.train_indices)
+
+    @property
+    def test_size(self) -> int:
+        return len(self.test_indices)
+
+
+def generate_walk_forward_splits(
+    *,
+    n_samples: int,
+    train_size: int,
+    test_size: int,
+    step_size: int | None = None,
+    purge_size: int = 0,
+    embargo_size: int = 0,
+    min_train_size: int = 1,
+) -> list[WalkForwardFold]:
+    """Generate chronological folds with purge/embargo leakage controls."""
+    if n_samples <= 0:
+        raise ValueError("n_samples must be positive")
+    if train_size <= 0 or test_size <= 0:
+        raise ValueError("train_size and test_size must be positive")
+    if purge_size < 0 or embargo_size < 0:
+        raise ValueError("purge_size and embargo_size must be >= 0")
+    if min_train_size <= 0:
+        raise ValueError("min_train_size must be positive")
+
+    step = step_size if step_size is not None else test_size
+    if step <= 0:
+        raise ValueError("step_size must be positive")
+
+    folds: list[WalkForwardFold] = []
+    prev_test_end: int | None = None
+    test_start = train_size + purge_size
+
+    while test_start + test_size <= n_samples:
+        test_end = test_start + test_size - 1
+        train_end = test_start - purge_size - 1
+        if train_end < 0:
+            break
+
+        train_start = max(0, train_end - train_size + 1)
+        train_indices = list(range(train_start, train_end + 1))
+
+        if prev_test_end is not None and embargo_size > 0:
+            emb_from = prev_test_end + 1
+            emb_to = prev_test_end + embargo_size
+            train_indices = [i for i in train_indices if i < emb_from or i > emb_to]
+
+        if len(train_indices) >= min_train_size:
+            folds.append(
+                WalkForwardFold(
+                    train_indices=train_indices,
+                    test_indices=list(range(test_start, test_end + 1)),
+                )
+            )
+            prev_test_end = test_end
+        test_start += step
+
+    return folds

src/backup/__init__.py

@@ -0,0 +1,21 @@
+"""Backup and disaster recovery system for long-term sustainability.
+
+This module provides:
+- Automated database backups (daily, weekly, monthly)
+- Multi-format exports (JSON, CSV, Parquet)
+- Cloud storage integration (S3-compatible)
+- Health monitoring and alerts
+"""
+
+from src.backup.exporter import BackupExporter, ExportFormat
+from src.backup.scheduler import BackupScheduler, BackupPolicy
+from src.backup.cloud_storage import CloudStorage, S3Config
+
+__all__ = [
+    "BackupExporter",
+    "ExportFormat",
+    "BackupScheduler",
+    "BackupPolicy",
+    "CloudStorage",
+    "S3Config",
+]

src/backup/cloud_storage.py

@@ -0,0 +1,274 @@
+"""Cloud storage integration for off-site backups.
+
+Supports S3-compatible storage providers:
+- AWS S3
+- MinIO
+- Backblaze B2
+- DigitalOcean Spaces
+- Cloudflare R2
+"""
+
+from __future__ import annotations
+
+import logging
+from dataclasses import dataclass
+from pathlib import Path
+from typing import Any
+
+logger = logging.getLogger(__name__)
+
+
+@dataclass
+class S3Config:
+    """Configuration for S3-compatible storage."""
+
+    endpoint_url: str | None  # None for AWS S3, custom URL for others
+    access_key: str
+    secret_key: str
+    bucket_name: str
+    region: str = "us-east-1"
+    use_ssl: bool = True
+
+
+class CloudStorage:
+    """Upload backups to S3-compatible cloud storage."""
+
+    def __init__(self, config: S3Config) -> None:
+        """Initialize cloud storage client.
+
+        Args:
+            config: S3 configuration
+
+        Raises:
+            ImportError: If boto3 is not installed
+        """
+        try:
+            import boto3
+        except ImportError:
+            raise ImportError(
+                "boto3 is required for cloud storage. Install with: pip install boto3"
+            )
+
+        self.config = config
+        self.client = boto3.client(
+            "s3",
+            endpoint_url=config.endpoint_url,
+            aws_access_key_id=config.access_key,
+            aws_secret_access_key=config.secret_key,
+            region_name=config.region,
+            use_ssl=config.use_ssl,
+        )
+
+    def upload_file(
+        self,
+        file_path: Path,
+        object_key: str | None = None,
+        metadata: dict[str, str] | None = None,
+    ) -> str:
+        """Upload a file to cloud storage.
+
+        Args:
+            file_path: Local file to upload
+            object_key: S3 object key (default: filename)
+            metadata: Optional metadata to attach
+
+        Returns:
+            S3 object key
+
+        Raises:
+            FileNotFoundError: If file doesn't exist
+            Exception: If upload fails
+        """
+        if not file_path.exists():
+            raise FileNotFoundError(f"File not found: {file_path}")
+
+        if object_key is None:
+            object_key = file_path.name
+
+        extra_args: dict[str, Any] = {}
+
+        # Add server-side encryption
+        extra_args["ServerSideEncryption"] = "AES256"
+
+        # Add metadata if provided
+        if metadata:
+            extra_args["Metadata"] = metadata
+
+        logger.info("Uploading %s to s3://%s/%s", file_path.name, self.config.bucket_name, object_key)
+
+        try:
+            self.client.upload_file(
+                str(file_path),
+                self.config.bucket_name,
+                object_key,
+                ExtraArgs=extra_args,
+            )
+            logger.info("Upload successful: %s", object_key)
+            return object_key
+        except Exception as exc:
+            logger.error("Upload failed: %s", exc)
+            raise
+
+    def download_file(self, object_key: str, local_path: Path) -> Path:
+        """Download a file from cloud storage.
+
+        Args:
+            object_key: S3 object key
+            local_path: Local destination path
+
+        Returns:
+            Path to downloaded file
+
+        Raises:
+            Exception: If download fails
+        """
+        local_path.parent.mkdir(parents=True, exist_ok=True)
+
+        logger.info("Downloading s3://%s/%s to %s", self.config.bucket_name, object_key, local_path)
+
+        try:
+            self.client.download_file(
+                self.config.bucket_name,
+                object_key,
+                str(local_path),
+            )
+            logger.info("Download successful: %s", local_path)
+            return local_path
+        except Exception as exc:
+            logger.error("Download failed: %s", exc)
+            raise
+
+    def list_files(self, prefix: str = "") -> list[dict[str, Any]]:
+        """List files in cloud storage.
+
+        Args:
+            prefix: Filter by object key prefix
+
+        Returns:
+            List of file metadata dictionaries
+        """
+        try:
+            response = self.client.list_objects_v2(
+                Bucket=self.config.bucket_name,
+                Prefix=prefix,
+            )
+
+            if "Contents" not in response:
+                return []
+
+            files = []
+            for obj in response["Contents"]:
+                files.append(
+                    {
+                        "key": obj["Key"],
+                        "size_bytes": obj["Size"],
+                        "last_modified": obj["LastModified"],
+                        "etag": obj["ETag"],
+                    }
+                )
+
+            return files
+        except Exception as exc:
+            logger.error("Failed to list files: %s", exc)
+            raise
+
+    def delete_file(self, object_key: str) -> None:
+        """Delete a file from cloud storage.
+
+        Args:
+            object_key: S3 object key
+
+        Raises:
+            Exception: If deletion fails
+        """
+        logger.info("Deleting s3://%s/%s", self.config.bucket_name, object_key)
+
+        try:
+            self.client.delete_object(
+                Bucket=self.config.bucket_name,
+                Key=object_key,
+            )
+            logger.info("Deletion successful: %s", object_key)
+        except Exception as exc:
+            logger.error("Deletion failed: %s", exc)
+            raise
+
+    def get_storage_stats(self) -> dict[str, Any]:
+        """Get cloud storage statistics.
+
+        Returns:
+            Dictionary with storage stats
+        """
+        try:
+            files = self.list_files()
+
+            total_size = sum(f["size_bytes"] for f in files)
+            total_count = len(files)
+
+            return {
+                "total_files": total_count,
+                "total_size_bytes": total_size,
+                "total_size_mb": total_size / 1024 / 1024,
+                "total_size_gb": total_size / 1024 / 1024 / 1024,
+            }
+        except Exception as exc:
+            logger.error("Failed to get storage stats: %s", exc)
+            return {
+                "error": str(exc),
+                "total_files": 0,
+                "total_size_bytes": 0,
+            }
+
+    def verify_connection(self) -> bool:
+        """Verify connection to cloud storage.
+
+        Returns:
+            True if connection is successful
+        """
+        try:
+            self.client.head_bucket(Bucket=self.config.bucket_name)
+            logger.info("Cloud storage connection verified")
+            return True
+        except Exception as exc:
+            logger.error("Cloud storage connection failed: %s", exc)
+            return False
+
+    def create_bucket_if_not_exists(self) -> None:
+        """Create storage bucket if it doesn't exist.
+
+        Raises:
+            Exception: If bucket creation fails
+        """
+        try:
+            self.client.head_bucket(Bucket=self.config.bucket_name)
+            logger.info("Bucket already exists: %s", self.config.bucket_name)
+        except self.client.exceptions.NoSuchBucket:
+            logger.info("Creating bucket: %s", self.config.bucket_name)
+            if self.config.region == "us-east-1":
+                # us-east-1 requires special handling
+                self.client.create_bucket(Bucket=self.config.bucket_name)
+            else:
+                self.client.create_bucket(
+                    Bucket=self.config.bucket_name,
+                    CreateBucketConfiguration={"LocationConstraint": self.config.region},
+                )
+            logger.info("Bucket created successfully")
+        except Exception as exc:
+            logger.error("Failed to verify/create bucket: %s", exc)
+            raise
+
+    def enable_versioning(self) -> None:
+        """Enable versioning on the bucket.
+
+        Raises:
+            Exception: If versioning enablement fails
+        """
+        try:
+            self.client.put_bucket_versioning(
+                Bucket=self.config.bucket_name,
+                VersioningConfiguration={"Status": "Enabled"},
+            )
+            logger.info("Versioning enabled for bucket: %s", self.config.bucket_name)
+        except Exception as exc:
+            logger.error("Failed to enable versioning: %s", exc)
+            raise

src/backup/exporter.py

@@ -0,0 +1,326 @@
+"""Multi-format database exporter for backups.
+
+Supports JSON, CSV, and Parquet formats for different use cases:
+- JSON: Human-readable, easy to inspect
+- CSV: Analysis tools (Excel, pandas)
+- Parquet: Big data tools (Spark, DuckDB)
+"""
+
+from __future__ import annotations
+
+import csv
+import gzip
+import json
+import logging
+import sqlite3
+from datetime import UTC, datetime
+from enum import Enum
+from pathlib import Path
+from typing import Any
+
+logger = logging.getLogger(__name__)
+
+
+class ExportFormat(str, Enum):
+    """Supported export formats."""
+
+    JSON = "json"
+    CSV = "csv"
+    PARQUET = "parquet"
+
+
+class BackupExporter:
+    """Export database to multiple formats."""
+
+    def __init__(self, db_path: str) -> None:
+        """Initialize the exporter.
+
+        Args:
+            db_path: Path to SQLite database
+        """
+        self.db_path = db_path
+
+    def export_all(
+        self,
+        output_dir: Path,
+        formats: list[ExportFormat] | None = None,
+        compress: bool = True,
+        incremental_since: datetime | None = None,
+    ) -> dict[ExportFormat, Path]:
+        """Export database to multiple formats.
+
+        Args:
+            output_dir: Directory to write export files
+            formats: List of formats to export (default: all)
+            compress: Whether to gzip compress exports
+            incremental_since: Only export records after this timestamp
+
+        Returns:
+            Dictionary mapping format to output file path
+        """
+        if formats is None:
+            formats = [ExportFormat.JSON, ExportFormat.CSV, ExportFormat.PARQUET]
+
+        output_dir.mkdir(parents=True, exist_ok=True)
+        timestamp = datetime.now(UTC).strftime("%Y%m%d_%H%M%S")
+
+        results: dict[ExportFormat, Path] = {}
+
+        for fmt in formats:
+            try:
+                output_file = self._export_format(
+                    fmt, output_dir, timestamp, compress, incremental_since
+                )
+                results[fmt] = output_file
+                logger.info("Exported to %s: %s", fmt.value, output_file)
+            except Exception as exc:
+                logger.error("Failed to export to %s: %s", fmt.value, exc)
+
+        return results
+
+    def _export_format(
+        self,
+        fmt: ExportFormat,
+        output_dir: Path,
+        timestamp: str,
+        compress: bool,
+        incremental_since: datetime | None,
+    ) -> Path:
+        """Export to a specific format.
+
+        Args:
+            fmt: Export format
+            output_dir: Output directory
+            timestamp: Timestamp string for filename
+            compress: Whether to compress
+            incremental_since: Incremental export cutoff
+
+        Returns:
+            Path to output file
+        """
+        if fmt == ExportFormat.JSON:
+            return self._export_json(output_dir, timestamp, compress, incremental_since)
+        elif fmt == ExportFormat.CSV:
+            return self._export_csv(output_dir, timestamp, compress, incremental_since)
+        elif fmt == ExportFormat.PARQUET:
+            return self._export_parquet(
+                output_dir, timestamp, compress, incremental_since
+            )
+        else:
+            raise ValueError(f"Unsupported format: {fmt}")
+
+    def _get_trades(
+        self, incremental_since: datetime | None = None
+    ) -> list[dict[str, Any]]:
+        """Fetch trades from database.
+
+        Args:
+            incremental_since: Only fetch trades after this timestamp
+
+        Returns:
+            List of trade records
+        """
+        conn = sqlite3.connect(self.db_path)
+        conn.row_factory = sqlite3.Row
+
+        if incremental_since:
+            cursor = conn.execute(
+                "SELECT * FROM trades WHERE timestamp > ?",
+                (incremental_since.isoformat(),),
+            )
+        else:
+            cursor = conn.execute("SELECT * FROM trades")
+
+        trades = [dict(row) for row in cursor.fetchall()]
+        conn.close()
+
+        return trades
+
+    def _export_json(
+        self,
+        output_dir: Path,
+        timestamp: str,
+        compress: bool,
+        incremental_since: datetime | None,
+    ) -> Path:
+        """Export to JSON format.
+
+        Args:
+            output_dir: Output directory
+            timestamp: Timestamp for filename
+            compress: Whether to gzip
+            incremental_since: Incremental cutoff
+
+        Returns:
+            Path to output file
+        """
+        trades = self._get_trades(incremental_since)
+
+        filename = f"trades_{timestamp}.json"
+        if compress:
+            filename += ".gz"
+
+        output_file = output_dir / filename
+
+        data = {
+            "export_timestamp": datetime.now(UTC).isoformat(),
+            "incremental_since": (
+                incremental_since.isoformat() if incremental_since else None
+            ),
+            "record_count": len(trades),
+            "trades": trades,
+        }
+
+        if compress:
+            with gzip.open(output_file, "wt", encoding="utf-8") as f:
+                json.dump(data, f, indent=2, ensure_ascii=False)
+        else:
+            with open(output_file, "w", encoding="utf-8") as f:
+                json.dump(data, f, indent=2, ensure_ascii=False)
+
+        return output_file
+
+    def _export_csv(
+        self,
+        output_dir: Path,
+        timestamp: str,
+        compress: bool,
+        incremental_since: datetime | None,
+    ) -> Path:
+        """Export to CSV format.
+
+        Args:
+            output_dir: Output directory
+            timestamp: Timestamp for filename
+            compress: Whether to gzip
+            incremental_since: Incremental cutoff
+
+        Returns:
+            Path to output file
+        """
+        trades = self._get_trades(incremental_since)
+
+        filename = f"trades_{timestamp}.csv"
+        if compress:
+            filename += ".gz"
+
+        output_file = output_dir / filename
+
+        if not trades:
+            # Write empty CSV with headers
+            if compress:
+                with gzip.open(output_file, "wt", encoding="utf-8", newline="") as f:
+                    writer = csv.writer(f)
+                    writer.writerow(
+                        [
+                            "timestamp",
+                            "stock_code",
+                            "action",
+                            "quantity",
+                            "price",
+                            "confidence",
+                            "rationale",
+                            "pnl",
+                        ]
+                    )
+            else:
+                with open(output_file, "w", encoding="utf-8", newline="") as f:
+                    writer = csv.writer(f)
+                    writer.writerow(
+                        [
+                            "timestamp",
+                            "stock_code",
+                            "action",
+                            "quantity",
+                            "price",
+                            "confidence",
+                            "rationale",
+                            "pnl",
+                        ]
+                    )
+            return output_file
+
+        # Get column names from first trade
+        fieldnames = list(trades[0].keys())
+
+        if compress:
+            with gzip.open(output_file, "wt", encoding="utf-8", newline="") as f:
+                writer = csv.DictWriter(f, fieldnames=fieldnames)
+                writer.writeheader()
+                writer.writerows(trades)
+        else:
+            with open(output_file, "w", encoding="utf-8", newline="") as f:
+                writer = csv.DictWriter(f, fieldnames=fieldnames)
+                writer.writeheader()
+                writer.writerows(trades)
+
+        return output_file
+
+    def _export_parquet(
+        self,
+        output_dir: Path,
+        timestamp: str,
+        compress: bool,
+        incremental_since: datetime | None,
+    ) -> Path:
+        """Export to Parquet format.
+
+        Args:
+            output_dir: Output directory
+            timestamp: Timestamp for filename
+            compress: Whether to compress (Parquet has built-in compression)
+            incremental_since: Incremental cutoff
+
+        Returns:
+            Path to output file
+        """
+        trades = self._get_trades(incremental_since)
+
+        filename = f"trades_{timestamp}.parquet"
+        output_file = output_dir / filename
+
+        try:
+            import pyarrow as pa
+            import pyarrow.parquet as pq
+        except ImportError:
+            raise ImportError(
+                "pyarrow is required for Parquet export. "
+                "Install with: pip install pyarrow"
+            )
+
+        # Convert to pyarrow table
+        table = pa.Table.from_pylist(trades)
+
+        # Write with compression
+        compression = "gzip" if compress else "none"
+        pq.write_table(table, output_file, compression=compression)
+
+        return output_file
+
+    def get_export_stats(self) -> dict[str, Any]:
+        """Get statistics about exportable data.
+
+        Returns:
+            Dictionary with data statistics
+        """
+        conn = sqlite3.connect(self.db_path)
+        cursor = conn.cursor()
+
+        stats = {}
+
+        # Total trades
+        cursor.execute("SELECT COUNT(*) FROM trades")
+        stats["total_trades"] = cursor.fetchone()[0]
+
+        # Date range
+        cursor.execute("SELECT MIN(timestamp), MAX(timestamp) FROM trades")
+        min_date, max_date = cursor.fetchone()
+        stats["date_range"] = {"earliest": min_date, "latest": max_date}
+
+        # Database size
+        cursor.execute("SELECT page_count * page_size FROM pragma_page_count(), pragma_page_size()")
+        stats["db_size_bytes"] = cursor.fetchone()[0]
+
+        conn.close()
+
+        return stats

src/backup/health_monitor.py

@@ -0,0 +1,282 @@
+"""Health monitoring for backup system.
+
+Checks:
+- Database accessibility and integrity
+- Disk space availability
+- Backup success/failure tracking
+- Self-healing capabilities
+"""
+
+from __future__ import annotations
+
+import logging
+import shutil
+import sqlite3
+from dataclasses import dataclass
+from datetime import UTC, datetime, timedelta
+from enum import Enum
+from pathlib import Path
+from typing import Any
+
+logger = logging.getLogger(__name__)
+
+
+class HealthStatus(str, Enum):
+    """Health check status."""
+
+    HEALTHY = "healthy"
+    DEGRADED = "degraded"
+    UNHEALTHY = "unhealthy"
+
+
+@dataclass
+class HealthCheckResult:
+    """Result of a health check."""
+
+    status: HealthStatus
+    message: str
+    details: dict[str, Any] | None = None
+    timestamp: datetime | None = None
+
+    def __post_init__(self) -> None:
+        if self.timestamp is None:
+            self.timestamp = datetime.now(UTC)
+
+
+class HealthMonitor:
+    """Monitor system health and backup status."""
+
+    def __init__(
+        self,
+        db_path: str,
+        backup_dir: Path,
+        min_disk_space_gb: float = 10.0,
+        max_backup_age_hours: int = 25,  # Daily backups should be < 25 hours old
+    ) -> None:
+        """Initialize health monitor.
+
+        Args:
+            db_path: Path to SQLite database
+            backup_dir: Backup directory
+            min_disk_space_gb: Minimum required disk space in GB
+            max_backup_age_hours: Maximum acceptable backup age in hours
+        """
+        self.db_path = Path(db_path)
+        self.backup_dir = backup_dir
+        self.min_disk_space_bytes = int(min_disk_space_gb * 1024 * 1024 * 1024)
+        self.max_backup_age = timedelta(hours=max_backup_age_hours)
+
+    def check_database_health(self) -> HealthCheckResult:
+        """Check database accessibility and integrity.
+
+        Returns:
+            HealthCheckResult
+        """
+        # Check if database exists
+        if not self.db_path.exists():
+            return HealthCheckResult(
+                status=HealthStatus.UNHEALTHY,
+                message=f"Database not found: {self.db_path}",
+            )
+
+        # Check if database is accessible
+        try:
+            conn = sqlite3.connect(str(self.db_path))
+            cursor = conn.cursor()
+
+            # Run integrity check
+            cursor.execute("PRAGMA integrity_check")
+            result = cursor.fetchone()[0]
+
+            if result != "ok":
+                conn.close()
+                return HealthCheckResult(
+                    status=HealthStatus.UNHEALTHY,
+                    message=f"Database integrity check failed: {result}",
+                )
+
+            # Get database size
+            cursor.execute(
+                "SELECT page_count * page_size FROM pragma_page_count(), pragma_page_size()"
+            )
+            db_size = cursor.fetchone()[0]
+
+            # Get row counts
+            cursor.execute("SELECT COUNT(*) FROM trades")
+            trade_count = cursor.fetchone()[0]
+
+            conn.close()
+
+            return HealthCheckResult(
+                status=HealthStatus.HEALTHY,
+                message="Database is healthy",
+                details={
+                    "size_bytes": db_size,
+                    "size_mb": db_size / 1024 / 1024,
+                    "trade_count": trade_count,
+                },
+            )
+
+        except sqlite3.Error as exc:
+            return HealthCheckResult(
+                status=HealthStatus.UNHEALTHY,
+                message=f"Database access error: {exc}",
+            )
+
+    def check_disk_space(self) -> HealthCheckResult:
+        """Check available disk space.
+
+        Returns:
+            HealthCheckResult
+        """
+        try:
+            stat = shutil.disk_usage(self.backup_dir)
+
+            free_gb = stat.free / 1024 / 1024 / 1024
+            total_gb = stat.total / 1024 / 1024 / 1024
+            used_percent = (stat.used / stat.total) * 100
+
+            if stat.free < self.min_disk_space_bytes:
+                return HealthCheckResult(
+                    status=HealthStatus.UNHEALTHY,
+                    message=f"Low disk space: {free_gb:.2f} GB free (minimum: {self.min_disk_space_bytes / 1024 / 1024 / 1024:.2f} GB)",
+                    details={
+                        "free_gb": free_gb,
+                        "total_gb": total_gb,
+                        "used_percent": used_percent,
+                    },
+                )
+            elif stat.free < self.min_disk_space_bytes * 2:
+                return HealthCheckResult(
+                    status=HealthStatus.DEGRADED,
+                    message=f"Disk space low: {free_gb:.2f} GB free",
+                    details={
+                        "free_gb": free_gb,
+                        "total_gb": total_gb,
+                        "used_percent": used_percent,
+                    },
+                )
+            else:
+                return HealthCheckResult(
+                    status=HealthStatus.HEALTHY,
+                    message=f"Disk space healthy: {free_gb:.2f} GB free",
+                    details={
+                        "free_gb": free_gb,
+                        "total_gb": total_gb,
+                        "used_percent": used_percent,
+                    },
+                )
+
+        except Exception as exc:
+            return HealthCheckResult(
+                status=HealthStatus.UNHEALTHY,
+                message=f"Failed to check disk space: {exc}",
+            )
+
+    def check_backup_recency(self) -> HealthCheckResult:
+        """Check if backups are recent enough.
+
+        Returns:
+            HealthCheckResult
+        """
+        daily_dir = self.backup_dir / "daily"
+
+        if not daily_dir.exists():
+            return HealthCheckResult(
+                status=HealthStatus.DEGRADED,
+                message="Daily backup directory not found",
+            )
+
+        # Find most recent backup
+        backups = sorted(daily_dir.glob("*.db"), key=lambda p: p.stat().st_mtime, reverse=True)
+
+        if not backups:
+            return HealthCheckResult(
+                status=HealthStatus.UNHEALTHY,
+                message="No daily backups found",
+            )
+
+        most_recent = backups[0]
+        mtime = datetime.fromtimestamp(most_recent.stat().st_mtime, tz=UTC)
+        age = datetime.now(UTC) - mtime
+
+        if age > self.max_backup_age:
+            return HealthCheckResult(
+                status=HealthStatus.DEGRADED,
+                message=f"Most recent backup is {age.total_seconds() / 3600:.1f} hours old",
+                details={
+                    "backup_file": most_recent.name,
+                    "age_hours": age.total_seconds() / 3600,
+                    "threshold_hours": self.max_backup_age.total_seconds() / 3600,
+                },
+            )
+        else:
+            return HealthCheckResult(
+                status=HealthStatus.HEALTHY,
+                message=f"Recent backup found ({age.total_seconds() / 3600:.1f} hours old)",
+                details={
+                    "backup_file": most_recent.name,
+                    "age_hours": age.total_seconds() / 3600,
+                },
+            )
+
+    def run_all_checks(self) -> dict[str, HealthCheckResult]:
+        """Run all health checks.
+
+        Returns:
+            Dictionary mapping check name to result
+        """
+        checks = {
+            "database": self.check_database_health(),
+            "disk_space": self.check_disk_space(),
+            "backup_recency": self.check_backup_recency(),
+        }
+
+        # Log results
+        for check_name, result in checks.items():
+            if result.status == HealthStatus.UNHEALTHY:
+                logger.error("[%s] %s: %s", check_name, result.status.value, result.message)
+            elif result.status == HealthStatus.DEGRADED:
+                logger.warning("[%s] %s: %s", check_name, result.status.value, result.message)
+            else:
+                logger.info("[%s] %s: %s", check_name, result.status.value, result.message)
+
+        return checks
+
+    def get_overall_status(self) -> HealthStatus:
+        """Get overall system health status.
+
+        Returns:
+            HealthStatus (worst status from all checks)
+        """
+        checks = self.run_all_checks()
+
+        # Return worst status
+        if any(c.status == HealthStatus.UNHEALTHY for c in checks.values()):
+            return HealthStatus.UNHEALTHY
+        elif any(c.status == HealthStatus.DEGRADED for c in checks.values()):
+            return HealthStatus.DEGRADED
+        else:
+            return HealthStatus.HEALTHY
+
+    def get_health_report(self) -> dict[str, Any]:
+        """Get comprehensive health report.
+
+        Returns:
+            Dictionary with health report
+        """
+        checks = self.run_all_checks()
+        overall = self.get_overall_status()
+
+        return {
+            "overall_status": overall.value,
+            "timestamp": datetime.now(UTC).isoformat(),
+            "checks": {
+                name: {
+                    "status": result.status.value,
+                    "message": result.message,
+                    "details": result.details,
+                }
+                for name, result in checks.items()
+            },
+        }

src/backup/scheduler.py

@@ -0,0 +1,336 @@
+"""Backup scheduler for automated database backups.
+
+Implements backup policies:
+- Daily: Keep for 30 days (hot storage)
+- Weekly: Keep for 1 year (warm storage)
+- Monthly: Keep forever (cold storage)
+"""
+
+from __future__ import annotations
+
+import logging
+import shutil
+from dataclasses import dataclass
+from datetime import UTC, datetime, timedelta
+from enum import Enum
+from pathlib import Path
+from typing import Any
+
+logger = logging.getLogger(__name__)
+
+
+class BackupPolicy(str, Enum):
+    """Backup retention policies."""
+
+    DAILY = "daily"
+    WEEKLY = "weekly"
+    MONTHLY = "monthly"
+
+
+@dataclass
+class BackupMetadata:
+    """Metadata for a backup."""
+
+    timestamp: datetime
+    policy: BackupPolicy
+    file_path: Path
+    size_bytes: int
+    checksum: str | None = None
+
+
+class BackupScheduler:
+    """Manage automated database backups with retention policies."""
+
+    def __init__(
+        self,
+        db_path: str,
+        backup_dir: Path,
+        daily_retention_days: int = 30,
+        weekly_retention_days: int = 365,
+    ) -> None:
+        """Initialize the backup scheduler.
+
+        Args:
+            db_path: Path to SQLite database
+            backup_dir: Root directory for backups
+            daily_retention_days: Days to keep daily backups
+            weekly_retention_days: Days to keep weekly backups
+        """
+        self.db_path = Path(db_path)
+        self.backup_dir = backup_dir
+        self.daily_retention = timedelta(days=daily_retention_days)
+        self.weekly_retention = timedelta(days=weekly_retention_days)
+
+        # Create policy-specific directories
+        self.daily_dir = backup_dir / "daily"
+        self.weekly_dir = backup_dir / "weekly"
+        self.monthly_dir = backup_dir / "monthly"
+
+        for d in [self.daily_dir, self.weekly_dir, self.monthly_dir]:
+            d.mkdir(parents=True, exist_ok=True)
+
+    def create_backup(
+        self, policy: BackupPolicy, verify: bool = True
+    ) -> BackupMetadata:
+        """Create a database backup.
+
+        Args:
+            policy: Backup policy (daily/weekly/monthly)
+            verify: Whether to verify backup integrity
+
+        Returns:
+            BackupMetadata object
+
+        Raises:
+            FileNotFoundError: If database doesn't exist
+            OSError: If backup fails
+        """
+        if not self.db_path.exists():
+            raise FileNotFoundError(f"Database not found: {self.db_path}")
+
+        timestamp = datetime.now(UTC)
+        backup_filename = self._get_backup_filename(timestamp, policy)
+
+        # Determine output directory
+        if policy == BackupPolicy.DAILY:
+            output_dir = self.daily_dir
+        elif policy == BackupPolicy.WEEKLY:
+            output_dir = self.weekly_dir
+        else:  # MONTHLY
+            output_dir = self.monthly_dir
+
+        backup_path = output_dir / backup_filename
+
+        # Create backup (copy database file)
+        logger.info("Creating %s backup: %s", policy.value, backup_path)
+        shutil.copy2(self.db_path, backup_path)
+
+        # Get file size
+        size_bytes = backup_path.stat().st_size
+
+        # Verify backup if requested
+        checksum = None
+        if verify:
+            checksum = self._verify_backup(backup_path)
+
+        metadata = BackupMetadata(
+            timestamp=timestamp,
+            policy=policy,
+            file_path=backup_path,
+            size_bytes=size_bytes,
+            checksum=checksum,
+        )
+
+        logger.info(
+            "Backup created: %s (%.2f MB)",
+            backup_path.name,
+            size_bytes / 1024 / 1024,
+        )
+
+        return metadata
+
+    def _get_backup_filename(self, timestamp: datetime, policy: BackupPolicy) -> str:
+        """Generate backup filename.
+
+        Args:
+            timestamp: Backup timestamp
+            policy: Backup policy
+
+        Returns:
+            Filename string
+        """
+        ts_str = timestamp.strftime("%Y%m%d_%H%M%S")
+        return f"trade_logs_{policy.value}_{ts_str}.db"
+
+    def _verify_backup(self, backup_path: Path) -> str:
+        """Verify backup integrity using SQLite integrity check.
+
+        Args:
+            backup_path: Path to backup file
+
+        Returns:
+            Checksum string (MD5 hash)
+
+        Raises:
+            RuntimeError: If integrity check fails
+        """
+        import hashlib
+        import sqlite3
+
+        # Integrity check
+        try:
+            conn = sqlite3.connect(str(backup_path))
+            cursor = conn.cursor()
+            cursor.execute("PRAGMA integrity_check")
+            result = cursor.fetchone()[0]
+            conn.close()
+
+            if result != "ok":
+                raise RuntimeError(f"Integrity check failed: {result}")
+        except sqlite3.Error as exc:
+            raise RuntimeError(f"Failed to verify backup: {exc}")
+
+        # Calculate MD5 checksum
+        md5 = hashlib.md5()
+        with open(backup_path, "rb") as f:
+            for chunk in iter(lambda: f.read(8192), b""):
+                md5.update(chunk)
+
+        return md5.hexdigest()
+
+    def cleanup_old_backups(self) -> dict[BackupPolicy, int]:
+        """Remove backups older than retention policies.
+
+        Returns:
+            Dictionary mapping policy to number of backups removed
+        """
+        now = datetime.now(UTC)
+        removed_counts: dict[BackupPolicy, int] = {}
+
+        # Daily backups: remove older than retention
+        removed_counts[BackupPolicy.DAILY] = self._cleanup_directory(
+            self.daily_dir, now - self.daily_retention
+        )
+
+        # Weekly backups: remove older than retention
+        removed_counts[BackupPolicy.WEEKLY] = self._cleanup_directory(
+            self.weekly_dir, now - self.weekly_retention
+        )
+
+        # Monthly backups: never remove (kept forever)
+        removed_counts[BackupPolicy.MONTHLY] = 0
+
+        total = sum(removed_counts.values())
+        if total > 0:
+            logger.info("Cleaned up %d old backup(s)", total)
+
+        return removed_counts
+
+    def _cleanup_directory(self, directory: Path, cutoff: datetime) -> int:
+        """Remove backups older than cutoff date.
+
+        Args:
+            directory: Directory to clean
+            cutoff: Remove files older than this
+
+        Returns:
+            Number of files removed
+        """
+        removed = 0
+
+        for backup_file in directory.glob("*.db"):
+            # Get file modification time
+            mtime = datetime.fromtimestamp(backup_file.stat().st_mtime, tz=UTC)
+
+            if mtime < cutoff:
+                logger.debug("Removing old backup: %s", backup_file.name)
+                backup_file.unlink()
+                removed += 1
+
+        return removed
+
+    def list_backups(
+        self, policy: BackupPolicy | None = None
+    ) -> list[BackupMetadata]:
+        """List available backups.
+
+        Args:
+            policy: Filter by policy (None for all)
+
+        Returns:
+            List of BackupMetadata objects
+        """
+        backups: list[BackupMetadata] = []
+
+        policies_to_check = (
+            [policy] if policy else [BackupPolicy.DAILY, BackupPolicy.WEEKLY, BackupPolicy.MONTHLY]
+        )
+
+        for pol in policies_to_check:
+            if pol == BackupPolicy.DAILY:
+                directory = self.daily_dir
+            elif pol == BackupPolicy.WEEKLY:
+                directory = self.weekly_dir
+            else:
+                directory = self.monthly_dir
+
+            for backup_file in sorted(directory.glob("*.db")):
+                mtime = datetime.fromtimestamp(backup_file.stat().st_mtime, tz=UTC)
+                size = backup_file.stat().st_size
+
+                backups.append(
+                    BackupMetadata(
+                        timestamp=mtime,
+                        policy=pol,
+                        file_path=backup_file,
+                        size_bytes=size,
+                    )
+                )
+
+        # Sort by timestamp (newest first)
+        backups.sort(key=lambda b: b.timestamp, reverse=True)
+
+        return backups
+
+    def get_backup_stats(self) -> dict[str, Any]:
+        """Get backup statistics.
+
+        Returns:
+            Dictionary with backup stats
+        """
+        stats: dict[str, Any] = {}
+
+        for policy in BackupPolicy:
+            if policy == BackupPolicy.DAILY:
+                directory = self.daily_dir
+            elif policy == BackupPolicy.WEEKLY:
+                directory = self.weekly_dir
+            else:
+                directory = self.monthly_dir
+
+            backups = list(directory.glob("*.db"))
+            total_size = sum(b.stat().st_size for b in backups)
+
+            stats[policy.value] = {
+                "count": len(backups),
+                "total_size_bytes": total_size,
+                "total_size_mb": total_size / 1024 / 1024,
+            }
+
+        return stats
+
+    def restore_backup(self, backup_metadata: BackupMetadata, verify: bool = True) -> None:
+        """Restore database from backup.
+
+        Args:
+            backup_metadata: Backup to restore
+            verify: Whether to verify restored database
+
+        Raises:
+            FileNotFoundError: If backup file doesn't exist
+            RuntimeError: If verification fails
+        """
+        if not backup_metadata.file_path.exists():
+            raise FileNotFoundError(f"Backup not found: {backup_metadata.file_path}")
+
+        # Create backup of current database
+        if self.db_path.exists():
+            backup_current = self.db_path.with_suffix(".db.before_restore")
+            logger.info("Backing up current database to: %s", backup_current)
+            shutil.copy2(self.db_path, backup_current)
+
+        # Restore backup
+        logger.info("Restoring backup: %s", backup_metadata.file_path.name)
+        shutil.copy2(backup_metadata.file_path, self.db_path)
+
+        # Verify restored database
+        if verify:
+            try:
+                self._verify_backup(self.db_path)
+                logger.info("Backup restored and verified successfully")
+            except RuntimeError as exc:
+                # Restore failed, revert to backup
+                if backup_current.exists():
+                    logger.error("Restore verification failed, reverting: %s", exc)
+                    shutil.copy2(backup_current, self.db_path)
+                raise

src/brain/cache.py

@@ -0,0 +1,293 @@
+"""Response caching system for reducing redundant LLM calls.
+
+This module provides caching for common trading scenarios:
+- TTL-based cache invalidation
+- Cache key based on market conditions
+- Cache hit rate monitoring
+- Special handling for HOLD decisions in quiet markets
+"""
+
+from __future__ import annotations
+
+import hashlib
+import json
+import logging
+import time
+from dataclasses import dataclass, field
+from typing import Any, TYPE_CHECKING
+
+if TYPE_CHECKING:
+    from src.brain.gemini_client import TradeDecision
+
+logger = logging.getLogger(__name__)
+
+
+@dataclass
+class CacheEntry:
+    """Cached decision with metadata."""
+
+    decision: "TradeDecision"
+    cached_at: float  # Unix timestamp
+    hit_count: int = 0
+    market_data_hash: str = ""
+
+
+@dataclass
+class CacheMetrics:
+    """Metrics for cache performance monitoring."""
+
+    total_requests: int = 0
+    cache_hits: int = 0
+    cache_misses: int = 0
+    evictions: int = 0
+    total_entries: int = 0
+
+    @property
+    def hit_rate(self) -> float:
+        """Calculate cache hit rate."""
+        if self.total_requests == 0:
+            return 0.0
+        return self.cache_hits / self.total_requests
+
+    def to_dict(self) -> dict[str, Any]:
+        """Convert metrics to dictionary."""
+        return {
+            "total_requests": self.total_requests,
+            "cache_hits": self.cache_hits,
+            "cache_misses": self.cache_misses,
+            "hit_rate": self.hit_rate,
+            "evictions": self.evictions,
+            "total_entries": self.total_entries,
+        }
+
+
+class DecisionCache:
+    """TTL-based cache for trade decisions."""
+
+    def __init__(self, ttl_seconds: int = 300, max_size: int = 1000) -> None:
+        """Initialize the decision cache.
+
+        Args:
+            ttl_seconds: Time-to-live for cache entries in seconds (default: 5 minutes)
+            max_size: Maximum number of cache entries
+        """
+        self.ttl_seconds = ttl_seconds
+        self.max_size = max_size
+        self._cache: dict[str, CacheEntry] = {}
+        self._metrics = CacheMetrics()
+
+    def _generate_cache_key(self, market_data: dict[str, Any]) -> str:
+        """Generate cache key from market data.
+
+        Key is based on:
+        - Stock code
+        - Current price (rounded to reduce sensitivity)
+        - Market conditions (orderbook snapshot)
+
+        Args:
+            market_data: Market data dictionary
+
+        Returns:
+            Cache key string
+        """
+        # Extract key components
+        stock_code = market_data.get("stock_code", "UNKNOWN")
+        current_price = market_data.get("current_price", 0)
+
+        # Round price to reduce sensitivity (cache hits for similar prices)
+        # For prices > 1000, round to nearest 10
+        # For prices < 1000, round to nearest 1
+        if current_price > 1000:
+            price_rounded = round(current_price / 10) * 10
+        else:
+            price_rounded = round(current_price)
+
+        # Include orderbook snapshot (if available)
+        orderbook_key = ""
+        if "orderbook" in market_data and market_data["orderbook"]:
+            ob = market_data["orderbook"]
+            # Just use bid/ask spread as indicator
+            if "bid" in ob and "ask" in ob and ob["bid"] and ob["ask"]:
+                bid_price = ob["bid"][0].get("price", 0) if ob["bid"] else 0
+                ask_price = ob["ask"][0].get("price", 0) if ob["ask"] else 0
+                spread = ask_price - bid_price
+                orderbook_key = f"_spread{spread}"
+
+        # Generate cache key
+        key_str = f"{stock_code}_{price_rounded}{orderbook_key}"
+
+        return key_str
+
+    def _generate_market_hash(self, market_data: dict[str, Any]) -> str:
+        """Generate hash of full market data for invalidation checks.
+
+        Args:
+            market_data: Market data dictionary
+
+        Returns:
+            Hash string
+        """
+        # Create stable JSON representation
+        stable_json = json.dumps(market_data, sort_keys=True, ensure_ascii=False)
+        return hashlib.md5(stable_json.encode()).hexdigest()
+
+    def get(self, market_data: dict[str, Any]) -> TradeDecision | None:
+        """Retrieve cached decision if valid.
+
+        Args:
+            market_data: Market data dictionary
+
+        Returns:
+            Cached TradeDecision if valid, None otherwise
+        """
+        self._metrics.total_requests += 1
+
+        cache_key = self._generate_cache_key(market_data)
+
+        if cache_key not in self._cache:
+            self._metrics.cache_misses += 1
+            return None
+
+        entry = self._cache[cache_key]
+        current_time = time.time()
+
+        # Check TTL
+        if current_time - entry.cached_at > self.ttl_seconds:
+            # Expired
+            del self._cache[cache_key]
+            self._metrics.cache_misses += 1
+            self._metrics.evictions += 1
+            logger.debug("Cache expired for key: %s", cache_key)
+            return None
+
+        # Cache hit
+        entry.hit_count += 1
+        self._metrics.cache_hits += 1
+        logger.debug("Cache hit for key: %s (hits: %d)", cache_key, entry.hit_count)
+
+        return entry.decision
+
+    def set(
+        self,
+        market_data: dict[str, Any],
+        decision: TradeDecision,
+    ) -> None:
+        """Store decision in cache.
+
+        Args:
+            market_data: Market data dictionary
+            decision: TradeDecision to cache
+        """
+        cache_key = self._generate_cache_key(market_data)
+        market_hash = self._generate_market_hash(market_data)
+
+        # Enforce max size (evict oldest if full)
+        if len(self._cache) >= self.max_size:
+            # Find oldest entry
+            oldest_key = min(self._cache.keys(), key=lambda k: self._cache[k].cached_at)
+            del self._cache[oldest_key]
+            self._metrics.evictions += 1
+            logger.debug("Cache full, evicted key: %s", oldest_key)
+
+        # Store entry
+        entry = CacheEntry(
+            decision=decision,
+            cached_at=time.time(),
+            market_data_hash=market_hash,
+        )
+        self._cache[cache_key] = entry
+        self._metrics.total_entries = len(self._cache)
+
+        logger.debug("Cached decision for key: %s", cache_key)
+
+    def invalidate(self, stock_code: str | None = None) -> int:
+        """Invalidate cache entries.
+
+        Args:
+            stock_code: Specific stock code to invalidate, or None for all
+
+        Returns:
+            Number of entries invalidated
+        """
+        if stock_code is None:
+            # Clear all
+            count = len(self._cache)
+            self._cache.clear()
+            self._metrics.evictions += count
+            self._metrics.total_entries = 0
+            logger.info("Invalidated all cache entries (%d)", count)
+            return count
+
+        # Invalidate specific stock
+        keys_to_remove = [k for k in self._cache.keys() if k.startswith(f"{stock_code}_")]
+        count = len(keys_to_remove)
+
+        for key in keys_to_remove:
+            del self._cache[key]
+
+        self._metrics.evictions += count
+        self._metrics.total_entries = len(self._cache)
+        logger.info("Invalidated %d cache entries for stock: %s", count, stock_code)
+
+        return count
+
+    def cleanup_expired(self) -> int:
+        """Remove expired entries from cache.
+
+        Returns:
+            Number of entries removed
+        """
+        current_time = time.time()
+        expired_keys = [
+            k
+            for k, v in self._cache.items()
+            if current_time - v.cached_at > self.ttl_seconds
+        ]
+
+        count = len(expired_keys)
+        for key in expired_keys:
+            del self._cache[key]
+
+        self._metrics.evictions += count
+        self._metrics.total_entries = len(self._cache)
+
+        if count > 0:
+            logger.debug("Cleaned up %d expired cache entries", count)
+
+        return count
+
+    def get_metrics(self) -> CacheMetrics:
+        """Get current cache metrics.
+
+        Returns:
+            CacheMetrics object with current statistics
+        """
+        return self._metrics
+
+    def reset_metrics(self) -> None:
+        """Reset cache metrics."""
+        self._metrics = CacheMetrics(total_entries=len(self._cache))
+        logger.info("Cache metrics reset")
+
+    def should_cache_decision(self, decision: TradeDecision) -> bool:
+        """Determine if a decision should be cached.
+
+        HOLD decisions with low confidence are good candidates for caching,
+        as they're likely to recur in quiet markets.
+
+        Args:
+            decision: TradeDecision to evaluate
+
+        Returns:
+            True if decision should be cached
+        """
+        # Cache HOLD decisions (common in quiet markets)
+        if decision.action == "HOLD":
+            return True
+
+        # Cache high-confidence decisions (stable signals)
+        if decision.confidence >= 90:
+            return True
+
+        # Don't cache low-confidence BUY/SELL (volatile signals)
+        return False

src/brain/context_selector.py

@@ -0,0 +1,296 @@
+"""Smart context selection for optimizing token usage.
+
+This module implements intelligent selection of context layers (L1-L7) based on
+decision type and market conditions:
+- L7 (real-time) for normal trading decisions
+- L6-L5 (daily/weekly) for strategic decisions
+- L4-L1 (monthly/legacy) only for major events or policy changes
+"""
+
+from __future__ import annotations
+
+from dataclasses import dataclass
+from datetime import UTC, datetime
+from enum import Enum
+from typing import Any
+
+from src.context.layer import ContextLayer
+from src.context.store import ContextStore
+
+
+class DecisionType(str, Enum):
+    """Type of trading decision being made."""
+
+    NORMAL = "normal"  # Regular trade decision
+    STRATEGIC = "strategic"  # Strategy adjustment
+    MAJOR_EVENT = "major_event"  # Portfolio rebalancing, policy change
+
+
+@dataclass(frozen=True)
+class ContextSelection:
+    """Selected context layers and their relevance scores."""
+
+    layers: list[ContextLayer]
+    relevance_scores: dict[ContextLayer, float]
+    total_score: float
+
+
+class ContextSelector:
+    """Selects optimal context layers to minimize token usage."""
+
+    def __init__(self, store: ContextStore) -> None:
+        """Initialize the context selector.
+
+        Args:
+            store: ContextStore instance for retrieving context data
+        """
+        self.store = store
+
+    def select_layers(
+        self,
+        decision_type: DecisionType = DecisionType.NORMAL,
+        include_realtime: bool = True,
+    ) -> list[ContextLayer]:
+        """Select context layers based on decision type.
+
+        Strategy:
+        - NORMAL: L7 (real-time) only
+        - STRATEGIC: L7 + L6 + L5 (real-time + daily + weekly)
+        - MAJOR_EVENT: All layers L1-L7
+
+        Args:
+            decision_type: Type of decision being made
+            include_realtime: Whether to include L7 real-time data
+
+        Returns:
+            List of context layers to use (ordered by priority)
+        """
+        if decision_type == DecisionType.NORMAL:
+            # Normal trading: only real-time data
+            return [ContextLayer.L7_REALTIME] if include_realtime else []
+
+        elif decision_type == DecisionType.STRATEGIC:
+            # Strategic decisions: real-time + recent history
+            layers = []
+            if include_realtime:
+                layers.append(ContextLayer.L7_REALTIME)
+            layers.extend([ContextLayer.L6_DAILY, ContextLayer.L5_WEEKLY])
+            return layers
+
+        else:  # MAJOR_EVENT
+            # Major events: all layers for comprehensive context
+            layers = []
+            if include_realtime:
+                layers.append(ContextLayer.L7_REALTIME)
+            layers.extend(
+                [
+                    ContextLayer.L6_DAILY,
+                    ContextLayer.L5_WEEKLY,
+                    ContextLayer.L4_MONTHLY,
+                    ContextLayer.L3_QUARTERLY,
+                    ContextLayer.L2_ANNUAL,
+                    ContextLayer.L1_LEGACY,
+                ]
+            )
+            return layers
+
+    def score_layer_relevance(
+        self,
+        layer: ContextLayer,
+        decision_type: DecisionType,
+        current_time: datetime | None = None,
+    ) -> float:
+        """Calculate relevance score for a context layer.
+
+        Relevance is based on:
+        1. Decision type (normal, strategic, major event)
+        2. Layer recency (L7 > L6 > ... > L1)
+        3. Data availability
+
+        Args:
+            layer: Context layer to score
+            decision_type: Type of decision being made
+            current_time: Current time (defaults to now)
+
+        Returns:
+            Relevance score (0.0 to 1.0)
+        """
+        if current_time is None:
+            current_time = datetime.now(UTC)
+
+        # Base scores by decision type
+        base_scores = {
+            DecisionType.NORMAL: {
+                ContextLayer.L7_REALTIME: 1.0,
+                ContextLayer.L6_DAILY: 0.1,
+                ContextLayer.L5_WEEKLY: 0.05,
+                ContextLayer.L4_MONTHLY: 0.01,
+                ContextLayer.L3_QUARTERLY: 0.0,
+                ContextLayer.L2_ANNUAL: 0.0,
+                ContextLayer.L1_LEGACY: 0.0,
+            },
+            DecisionType.STRATEGIC: {
+                ContextLayer.L7_REALTIME: 0.9,
+                ContextLayer.L6_DAILY: 0.8,
+                ContextLayer.L5_WEEKLY: 0.7,
+                ContextLayer.L4_MONTHLY: 0.3,
+                ContextLayer.L3_QUARTERLY: 0.2,
+                ContextLayer.L2_ANNUAL: 0.1,
+                ContextLayer.L1_LEGACY: 0.05,
+            },
+            DecisionType.MAJOR_EVENT: {
+                ContextLayer.L7_REALTIME: 0.7,
+                ContextLayer.L6_DAILY: 0.7,
+                ContextLayer.L5_WEEKLY: 0.7,
+                ContextLayer.L4_MONTHLY: 0.8,
+                ContextLayer.L3_QUARTERLY: 0.8,
+                ContextLayer.L2_ANNUAL: 0.9,
+                ContextLayer.L1_LEGACY: 1.0,
+            },
+        }
+
+        score = base_scores[decision_type].get(layer, 0.0)
+
+        # Check data availability
+        latest_timeframe = self.store.get_latest_timeframe(layer)
+        if latest_timeframe is None:
+            # No data available - reduce score significantly
+            score *= 0.1
+
+        return score
+
+    def select_with_scoring(
+        self,
+        decision_type: DecisionType = DecisionType.NORMAL,
+        min_score: float = 0.5,
+    ) -> ContextSelection:
+        """Select context layers with relevance scoring.
+
+        Args:
+            decision_type: Type of decision being made
+            min_score: Minimum relevance score to include a layer
+
+        Returns:
+            ContextSelection with selected layers and scores
+        """
+        all_layers = [
+            ContextLayer.L7_REALTIME,
+            ContextLayer.L6_DAILY,
+            ContextLayer.L5_WEEKLY,
+            ContextLayer.L4_MONTHLY,
+            ContextLayer.L3_QUARTERLY,
+            ContextLayer.L2_ANNUAL,
+            ContextLayer.L1_LEGACY,
+        ]
+
+        scores = {
+            layer: self.score_layer_relevance(layer, decision_type) for layer in all_layers
+        }
+
+        # Filter by minimum score
+        selected_layers = [layer for layer, score in scores.items() if score >= min_score]
+
+        # Sort by score (descending)
+        selected_layers.sort(key=lambda layer: scores[layer], reverse=True)
+
+        total_score = sum(scores[layer] for layer in selected_layers)
+
+        return ContextSelection(
+            layers=selected_layers,
+            relevance_scores=scores,
+            total_score=total_score,
+        )
+
+    def get_context_data(
+        self,
+        layers: list[ContextLayer],
+        max_items_per_layer: int = 10,
+    ) -> dict[str, Any]:
+        """Retrieve context data for selected layers.
+
+        Args:
+            layers: List of context layers to retrieve
+            max_items_per_layer: Maximum number of items per layer
+
+        Returns:
+            Dictionary with context data organized by layer
+        """
+        result: dict[str, Any] = {}
+
+        for layer in layers:
+            # Get latest timeframe for this layer
+            latest_timeframe = self.store.get_latest_timeframe(layer)
+            if latest_timeframe:
+                # Get all contexts for latest timeframe
+                contexts = self.store.get_all_contexts(layer, latest_timeframe)
+
+                # Limit number of items
+                if len(contexts) > max_items_per_layer:
+                    # Keep only first N items
+                    contexts = dict(list(contexts.items())[:max_items_per_layer])
+
+                result[layer.value] = contexts
+
+        return result
+
+    def estimate_context_tokens(self, context_data: dict[str, Any]) -> int:
+        """Estimate total tokens for context data.
+
+        Args:
+            context_data: Context data dictionary
+
+        Returns:
+            Estimated token count
+        """
+        import json
+
+        from src.brain.prompt_optimizer import PromptOptimizer
+
+        # Serialize to JSON and estimate tokens
+        json_str = json.dumps(context_data, ensure_ascii=False)
+        return PromptOptimizer.estimate_tokens(json_str)
+
+    def optimize_context_for_budget(
+        self,
+        decision_type: DecisionType,
+        max_tokens: int,
+    ) -> dict[str, Any]:
+        """Select and retrieve context data within a token budget.
+
+        Args:
+            decision_type: Type of decision being made
+            max_tokens: Maximum token budget for context
+
+        Returns:
+            Optimized context data within budget
+        """
+        # Start with minimal selection
+        selection = self.select_with_scoring(decision_type, min_score=0.5)
+
+        # Retrieve data
+        context_data = self.get_context_data(selection.layers)
+
+        # Check if within budget
+        estimated_tokens = self.estimate_context_tokens(context_data)
+
+        if estimated_tokens <= max_tokens:
+            return context_data
+
+        # If over budget, progressively reduce
+        # 1. Reduce items per layer
+        for max_items in [5, 3, 1]:
+            context_data = self.get_context_data(selection.layers, max_items)
+            estimated_tokens = self.estimate_context_tokens(context_data)
+            if estimated_tokens <= max_tokens:
+                return context_data
+
+        # 2. Remove lower-priority layers
+        for min_score in [0.6, 0.7, 0.8, 0.9]:
+            selection = self.select_with_scoring(decision_type, min_score=min_score)
+            context_data = self.get_context_data(selection.layers, max_items_per_layer=1)
+            estimated_tokens = self.estimate_context_tokens(context_data)
+            if estimated_tokens <= max_tokens:
+                return context_data
+
+        # Last resort: return only L7 with minimal data
+        return self.get_context_data([ContextLayer.L7_REALTIME], max_items_per_layer=1)

src/brain/gemini_client.py

@@ -2,6 +2,17 @@
 
 Constructs prompts from market data, calls Gemini, and parses structured
 JSON responses into validated TradeDecision objects.
+
+Includes token efficiency optimizations:
+- Prompt compression and abbreviation
+- Response caching for common scenarios
+- Smart context selection
+- Token usage tracking and metrics
+
+Includes external data integration:
+- News sentiment analysis
+- Economic calendar events
+- Market indicators
 """
 
 from __future__ import annotations
@@ -15,6 +26,11 @@ from typing import Any
 from google import genai
 
 from src.config import Settings
+from src.data.news_api import NewsAPI, NewsSentiment
+from src.data.economic_calendar import EconomicCalendar
+from src.data.market_data import MarketData
+from src.brain.cache import DecisionCache
+from src.brain.prompt_optimizer import PromptOptimizer
 
 logger = logging.getLogger(__name__)
 
@@ -28,23 +44,176 @@ class TradeDecision:
     action: str  # "BUY" | "SELL" | "HOLD"
     confidence: int  # 0-100
     rationale: str
+    token_count: int = 0  # Estimated tokens used
+    cached: bool = False  # Whether decision came from cache
 
 
 class GeminiClient:
     """Wraps the Gemini API for trade decision-making."""
 
-    def __init__(self, settings: Settings) -> None:
+    def __init__(
+        self,
+        settings: Settings,
+        news_api: NewsAPI | None = None,
+        economic_calendar: EconomicCalendar | None = None,
+        market_data: MarketData | None = None,
+        enable_cache: bool = True,
+        enable_optimization: bool = True,
+    ) -> None:
         self._settings = settings
         self._confidence_threshold = settings.CONFIDENCE_THRESHOLD
         self._client = genai.Client(api_key=settings.GEMINI_API_KEY)
         self._model_name = settings.GEMINI_MODEL
 
+        # External data sources (optional)
+        self._news_api = news_api
+        self._economic_calendar = economic_calendar
+        self._market_data = market_data
+
+        # Token efficiency features
+        self._enable_cache = enable_cache
+        self._enable_optimization = enable_optimization
+        self._cache = DecisionCache(ttl_seconds=300) if enable_cache else None
+        self._optimizer = PromptOptimizer()
+
+        # Token usage metrics
+        self._total_tokens_used = 0
+        self._total_decisions = 0
+        self._total_cached_decisions = 0
+
+    # ------------------------------------------------------------------
+    # External Data Integration
+    # ------------------------------------------------------------------
+
+    async def _build_external_context(
+        self, stock_code: str, news_sentiment: NewsSentiment | None = None
+    ) -> str:
+        """Build external data context for the prompt.
+
+        Args:
+            stock_code: Stock ticker symbol
+            news_sentiment: Optional pre-fetched news sentiment
+
+        Returns:
+            Formatted string with external data context
+        """
+        context_parts: list[str] = []
+
+        # News sentiment
+        if news_sentiment is not None:
+            sentiment_str = self._format_news_sentiment(news_sentiment)
+            if sentiment_str:
+                context_parts.append(sentiment_str)
+        elif self._news_api is not None:
+            # Fetch news sentiment if not provided
+            try:
+                sentiment = await self._news_api.get_news_sentiment(stock_code)
+                if sentiment is not None:
+                    sentiment_str = self._format_news_sentiment(sentiment)
+                    if sentiment_str:
+                        context_parts.append(sentiment_str)
+            except Exception as exc:
+                logger.warning("Failed to fetch news sentiment: %s", exc)
+
+        # Economic events
+        if self._economic_calendar is not None:
+            events_str = self._format_economic_events(stock_code)
+            if events_str:
+                context_parts.append(events_str)
+
+        # Market indicators
+        if self._market_data is not None:
+            indicators_str = self._format_market_indicators()
+            if indicators_str:
+                context_parts.append(indicators_str)
+
+        if not context_parts:
+            return ""
+
+        return "EXTERNAL DATA:\n" + "\n\n".join(context_parts)
+
+    def _format_news_sentiment(self, sentiment: NewsSentiment) -> str:
+        """Format news sentiment for prompt."""
+        if sentiment.article_count == 0:
+            return ""
+
+        # Select top 3 most relevant articles
+        top_articles = sentiment.articles[:3]
+
+        lines = [
+            f"News Sentiment: {sentiment.avg_sentiment:.2f} "
+            f"(from {sentiment.article_count} articles)",
+        ]
+
+        for i, article in enumerate(top_articles, 1):
+            lines.append(
+                f"  {i}. [{article.source}] {article.title} "
+                f"(sentiment: {article.sentiment_score:.2f})"
+            )
+
+        return "\n".join(lines)
+
+    def _format_economic_events(self, stock_code: str) -> str:
+        """Format upcoming economic events for prompt."""
+        if self._economic_calendar is None:
+            return ""
+
+        # Check for upcoming high-impact events
+        upcoming = self._economic_calendar.get_upcoming_events(
+            days_ahead=7, min_impact="HIGH"
+        )
+
+        if upcoming.high_impact_count == 0:
+            return ""
+
+        lines = [
+            f"Upcoming High-Impact Events: {upcoming.high_impact_count} in next 7 days"
+        ]
+
+        if upcoming.next_major_event is not None:
+            event = upcoming.next_major_event
+            lines.append(
+                f"  Next: {event.name} ({event.event_type}) "
+                f"on {event.datetime.strftime('%Y-%m-%d')}"
+            )
+
+        # Check for earnings
+        earnings_date = self._economic_calendar.get_earnings_date(stock_code)
+        if earnings_date is not None:
+            lines.append(
+                f"  Earnings: {stock_code} on {earnings_date.strftime('%Y-%m-%d')}"
+            )
+
+        return "\n".join(lines)
+
+    def _format_market_indicators(self) -> str:
+        """Format market indicators for prompt."""
+        if self._market_data is None:
+            return ""
+
+        try:
+            indicators = self._market_data.get_market_indicators()
+            lines = [f"Market Sentiment: {indicators.sentiment.name}"]
+
+            # Add breadth if meaningful
+            if indicators.breadth.advance_decline_ratio != 1.0:
+                lines.append(
+                    f"Advance/Decline Ratio: {indicators.breadth.advance_decline_ratio:.2f}"
+                )
+
+            return "\n".join(lines)
+        except Exception as exc:
+            logger.warning("Failed to get market indicators: %s", exc)
+            return ""
+
     # ------------------------------------------------------------------
     # Prompt Construction
     # ------------------------------------------------------------------
 
-    def build_prompt(self, market_data: dict[str, Any]) -> str:
-        """Build a structured prompt from market data.
+    async def build_prompt(
+        self, market_data: dict[str, Any], news_sentiment: NewsSentiment | None = None
+    ) -> str:
+        """Build a structured prompt from market data and external sources.
 
         The prompt instructs Gemini to return valid JSON with action,
         confidence, and rationale fields.
@@ -72,6 +241,60 @@ class GeminiClient:
 
         market_info = "\n".join(market_info_lines)
 
+        # Add external data context if available
+        external_context = await self._build_external_context(
+            market_data["stock_code"], news_sentiment
+        )
+        if external_context:
+            market_info += f"\n\n{external_context}"
+
+        json_format = (
+            '{"action": "BUY"|"SELL"|"HOLD", '
+            '"confidence": <int 0-100>, "rationale": "<string>"}'
+        )
+        return (
+            f"You are a professional {market_name} trading analyst.\n"
+            "Analyze the following market data and decide whether to "
+            "BUY, SELL, or HOLD.\n\n"
+            f"{market_info}\n\n"
+            "You MUST respond with ONLY valid JSON in the following format:\n"
+            f"{json_format}\n\n"
+            "Rules:\n"
+            "- action must be exactly one of: BUY, SELL, HOLD\n"
+            "- confidence must be an integer from 0 to 100\n"
+            "- rationale must explain your reasoning concisely\n"
+            "- Do NOT wrap the JSON in markdown code blocks\n"
+        )
+
+    def build_prompt_sync(self, market_data: dict[str, Any]) -> str:
+        """Synchronous version of build_prompt (for backward compatibility).
+
+        This version does NOT include external data integration.
+        Use async build_prompt() for full functionality.
+        """
+        market_name = market_data.get("market_name", "Korean stock market")
+
+        # Build market data section dynamically based on available fields
+        market_info_lines = [
+            f"Market: {market_name}",
+            f"Stock Code: {market_data['stock_code']}",
+            f"Current Price: {market_data['current_price']}",
+        ]
+
+        # Add orderbook if available (domestic markets)
+        if "orderbook" in market_data:
+            market_info_lines.append(
+                f"Orderbook: {json.dumps(market_data['orderbook'], ensure_ascii=False)}"
+            )
+
+        # Add foreigner net if non-zero
+        if market_data.get("foreigner_net", 0) != 0:
+            market_info_lines.append(
+                f"Foreigner Net Buy/Sell: {market_data['foreigner_net']}"
+            )
+
+        market_info = "\n".join(market_info_lines)
+
         json_format = (
             '{"action": "BUY"|"SELL"|"HOLD", '
             '"confidence": <int 0-100>, "rationale": "<string>"}'
@@ -123,8 +346,10 @@ class GeminiClient:
         # Validate required fields
         if not all(k in data for k in ("action", "confidence", "rationale")):
             logger.warning("Missing fields in Gemini response — defaulting to HOLD")
+            # Preserve raw text in rationale so prompt_override callers (e.g. pre_market_planner)
+            # can extract their own JSON format from decision.rationale (#245)
             return TradeDecision(
-                action="HOLD", confidence=0, rationale="Missing required fields"
+                action="HOLD", confidence=0, rationale=raw
             )
 
         action = str(data["action"]).upper()
@@ -152,28 +377,397 @@ class GeminiClient:
     # API Call
     # ------------------------------------------------------------------
 
-    async def decide(self, market_data: dict[str, Any]) -> TradeDecision:
-        """Build prompt, call Gemini, and return a parsed decision."""
-        prompt = self.build_prompt(market_data)
-        logger.info("Requesting trade decision from Gemini")
+    async def decide(
+        self, market_data: dict[str, Any], news_sentiment: NewsSentiment | None = None
+    ) -> TradeDecision:
+        """Build prompt, call Gemini, and return a parsed decision.
+
+        Args:
+            market_data: Market data dictionary with price, orderbook, etc.
+            news_sentiment: Optional pre-fetched news sentiment
+
+        Returns:
+            Parsed TradeDecision
+        """
+        # Check cache first
+        if self._cache:
+            cached_decision = self._cache.get(market_data)
+            if cached_decision:
+                self._total_cached_decisions += 1
+                self._total_decisions += 1
+                logger.info(
+                    "Cache hit for decision",
+                    extra={
+                        "action": cached_decision.action,
+                        "confidence": cached_decision.confidence,
+                        "cache_hit_rate": self.get_cache_hit_rate(),
+                    },
+                )
+                # Return cached decision with cached flag
+                return TradeDecision(
+                    action=cached_decision.action,
+                    confidence=cached_decision.confidence,
+                    rationale=cached_decision.rationale,
+                    token_count=0,
+                    cached=True,
+                )
+
+        # Build prompt (prompt_override takes priority for callers like pre_market_planner)
+        if "prompt_override" in market_data:
+            prompt = market_data["prompt_override"]
+        elif self._enable_optimization:
+            prompt = self._optimizer.build_compressed_prompt(market_data)
+        else:
+            prompt = await self.build_prompt(market_data, news_sentiment)
+
+        # Estimate tokens
+        token_count = self._optimizer.estimate_tokens(prompt)
+        self._total_tokens_used += token_count
+
+        logger.info(
+            "Requesting trade decision from Gemini",
+            extra={"estimated_tokens": token_count, "optimized": self._enable_optimization},
+        )
 
         try:
             response = await self._client.aio.models.generate_content(
-                model=self._model_name, contents=prompt,
+                model=self._model_name,
+                contents=prompt,
             )
             raw = response.text
         except Exception as exc:
             logger.error("Gemini API error: %s", exc)
             return TradeDecision(
-                action="HOLD", confidence=0, rationale=f"API error: {exc}"
+                action="HOLD", confidence=0, rationale=f"API error: {exc}", token_count=token_count
+            )
+
+        # prompt_override callers (e.g. pre_market_planner) expect raw text back,
+        # not a parsed TradeDecision. Skip parse_response to avoid spurious
+        # "Missing fields" warnings and return the raw response directly. (#247)
+        if "prompt_override" in market_data:
+            logger.info(
+                "Gemini raw response received (prompt_override, tokens=%d)", token_count
+            )
+            # Not a trade decision — don't inflate _total_decisions metrics
+            return TradeDecision(
+                action="HOLD", confidence=0, rationale=raw, token_count=token_count
             )
 
         decision = self.parse_response(raw)
+        self._total_decisions += 1
+
+        # Add token count to decision
+        decision_with_tokens = TradeDecision(
+            action=decision.action,
+            confidence=decision.confidence,
+            rationale=decision.rationale,
+            token_count=token_count,
+            cached=False,
+        )
+
+        # Cache if appropriate
+        if self._cache and self._cache.should_cache_decision(decision):
+            self._cache.set(market_data, decision)
+
         logger.info(
             "Gemini decision",
             extra={
                 "action": decision.action,
                 "confidence": decision.confidence,
+                "tokens": token_count,
+                "avg_tokens": self.get_avg_tokens_per_decision(),
             },
         )
-        return decision
+
+        return decision_with_tokens
+
+    # ------------------------------------------------------------------
+    # Token Efficiency Metrics
+    # ------------------------------------------------------------------
+
+    def get_token_metrics(self) -> dict[str, Any]:
+        """Get token usage metrics.
+
+        Returns:
+            Dictionary with token usage statistics
+        """
+        metrics = {
+            "total_tokens_used": self._total_tokens_used,
+            "total_decisions": self._total_decisions,
+            "total_cached_decisions": self._total_cached_decisions,
+            "avg_tokens_per_decision": self.get_avg_tokens_per_decision(),
+            "cache_hit_rate": self.get_cache_hit_rate(),
+        }
+
+        if self._cache:
+            cache_metrics = self._cache.get_metrics()
+            metrics["cache_metrics"] = cache_metrics.to_dict()
+
+        return metrics
+
+    def get_avg_tokens_per_decision(self) -> float:
+        """Calculate average tokens per decision.
+
+        Returns:
+            Average tokens per decision
+        """
+        if self._total_decisions == 0:
+            return 0.0
+        return self._total_tokens_used / self._total_decisions
+
+    def get_cache_hit_rate(self) -> float:
+        """Calculate cache hit rate.
+
+        Returns:
+            Cache hit rate (0.0 to 1.0)
+        """
+        if self._total_decisions == 0:
+            return 0.0
+        return self._total_cached_decisions / self._total_decisions
+
+    def reset_metrics(self) -> None:
+        """Reset token usage metrics."""
+        self._total_tokens_used = 0
+        self._total_decisions = 0
+        self._total_cached_decisions = 0
+        if self._cache:
+            self._cache.reset_metrics()
+        logger.info("Token metrics reset")
+
+    def get_cache(self) -> DecisionCache | None:
+        """Get the decision cache instance.
+
+        Returns:
+            DecisionCache instance or None if caching disabled
+        """
+        return self._cache
+
+    # ------------------------------------------------------------------
+    # Batch Decision Making (for daily trading mode)
+    # ------------------------------------------------------------------
+
+    async def decide_batch(
+        self, stocks_data: list[dict[str, Any]]
+    ) -> dict[str, TradeDecision]:
+        """Make decisions for multiple stocks in a single API call.
+
+        This is designed for daily trading mode to minimize API usage
+        when working with Gemini Free tier (20 calls/day limit).
+
+        Args:
+            stocks_data: List of market data dictionaries, each with:
+                - stock_code: Stock ticker
+                - current_price: Current price
+                - market_name: Market name (optional)
+                - foreigner_net: Foreigner net buy/sell (optional)
+
+        Returns:
+            Dictionary mapping stock_code to TradeDecision
+
+        Example:
+            >>> stocks_data = [
+            ...     {"stock_code": "AAPL", "current_price": 185.5},
+            ...     {"stock_code": "MSFT", "current_price": 420.0},
+            ... ]
+            >>> decisions = await client.decide_batch(stocks_data)
+            >>> decisions["AAPL"].action
+            'BUY'
+        """
+        if not stocks_data:
+            return {}
+
+        # Build compressed batch prompt
+        market_name = stocks_data[0].get("market_name", "stock market")
+
+        # Format stock data as compact JSON array
+        compact_stocks = []
+        for stock in stocks_data:
+            compact = {
+                "code": stock["stock_code"],
+                "price": stock["current_price"],
+            }
+            if stock.get("foreigner_net", 0) != 0:
+                compact["frgn"] = stock["foreigner_net"]
+            compact_stocks.append(compact)
+
+        data_str = json.dumps(compact_stocks, ensure_ascii=False)
+
+        prompt = (
+            f"You are a professional {market_name} trading analyst.\n"
+            "Analyze the following stocks and decide whether to BUY, SELL, or HOLD each one.\n\n"
+            f"Stock Data: {data_str}\n\n"
+            "You MUST respond with ONLY a valid JSON array in this format:\n"
+            '[{"code": "AAPL", "action": "BUY", "confidence": 85, "rationale": "..."},\n'
+            ' {"code": "MSFT", "action": "HOLD", "confidence": 50, "rationale": "..."}, ...]\n\n'
+            "Rules:\n"
+            "- Return one decision object per stock\n"
+            "- action must be exactly: BUY, SELL, or HOLD\n"
+            "- confidence must be 0-100\n"
+            "- rationale should be concise (1-2 sentences)\n"
+            "- Do NOT wrap JSON in markdown code blocks\n"
+        )
+
+        # Estimate tokens
+        token_count = self._optimizer.estimate_tokens(prompt)
+        self._total_tokens_used += token_count
+
+        logger.info(
+            "Requesting batch decision for %d stocks from Gemini",
+            len(stocks_data),
+            extra={"estimated_tokens": token_count},
+        )
+
+        try:
+            response = await self._client.aio.models.generate_content(
+                model=self._model_name,
+                contents=prompt,
+            )
+            raw = response.text
+        except Exception as exc:
+            logger.error("Gemini API error in batch decision: %s", exc)
+            # Return HOLD for all stocks on API error
+            return {
+                stock["stock_code"]: TradeDecision(
+                    action="HOLD",
+                    confidence=0,
+                    rationale=f"API error: {exc}",
+                    token_count=token_count,
+                    cached=False,
+                )
+                for stock in stocks_data
+            }
+
+        # Parse batch response
+        return self._parse_batch_response(raw, stocks_data, token_count)
+
+    def _parse_batch_response(
+        self, raw: str, stocks_data: list[dict[str, Any]], token_count: int
+    ) -> dict[str, TradeDecision]:
+        """Parse batch response into a dictionary of decisions.
+
+        Args:
+            raw: Raw response from Gemini
+            stocks_data: Original stock data list
+            token_count: Token count for the request
+
+        Returns:
+            Dictionary mapping stock_code to TradeDecision
+        """
+        if not raw or not raw.strip():
+            logger.warning("Empty batch response from Gemini — defaulting all to HOLD")
+            return {
+                stock["stock_code"]: TradeDecision(
+                    action="HOLD",
+                    confidence=0,
+                    rationale="Empty response",
+                    token_count=0,
+                    cached=False,
+                )
+                for stock in stocks_data
+            }
+
+        # Strip markdown code fences if present
+        cleaned = raw.strip()
+        match = re.search(r"```(?:json)?\s*\n?(.*?)\n?```", cleaned, re.DOTALL)
+        if match:
+            cleaned = match.group(1).strip()
+
+        try:
+            data = json.loads(cleaned)
+        except json.JSONDecodeError:
+            logger.warning("Malformed JSON in batch response — defaulting all to HOLD")
+            return {
+                stock["stock_code"]: TradeDecision(
+                    action="HOLD",
+                    confidence=0,
+                    rationale="Malformed JSON response",
+                    token_count=0,
+                    cached=False,
+                )
+                for stock in stocks_data
+            }
+
+        if not isinstance(data, list):
+            logger.warning("Batch response is not a JSON array — defaulting all to HOLD")
+            return {
+                stock["stock_code"]: TradeDecision(
+                    action="HOLD",
+                    confidence=0,
+                    rationale="Invalid response format",
+                    token_count=0,
+                    cached=False,
+                )
+                for stock in stocks_data
+            }
+
+        # Build decision map
+        decisions: dict[str, TradeDecision] = {}
+        stock_codes = {stock["stock_code"] for stock in stocks_data}
+
+        for item in data:
+            if not isinstance(item, dict):
+                continue
+
+            code = item.get("code")
+            if not code or code not in stock_codes:
+                continue
+
+            # Validate required fields
+            if not all(k in item for k in ("action", "confidence", "rationale")):
+                logger.warning("Missing fields for %s — using HOLD", code)
+                decisions[code] = TradeDecision(
+                    action="HOLD",
+                    confidence=0,
+                    rationale="Missing required fields",
+                    token_count=0,
+                    cached=False,
+                )
+                continue
+
+            action = str(item["action"]).upper()
+            if action not in VALID_ACTIONS:
+                logger.warning("Invalid action '%s' for %s — forcing HOLD", action, code)
+                action = "HOLD"
+
+            confidence = int(item["confidence"])
+            rationale = str(item["rationale"])
+
+            # Enforce confidence threshold
+            if confidence < self._confidence_threshold:
+                logger.info(
+                    "Confidence %d < threshold %d for %s — forcing HOLD",
+                    confidence,
+                    self._confidence_threshold,
+                    code,
+                )
+                action = "HOLD"
+
+            decisions[code] = TradeDecision(
+                action=action,
+                confidence=confidence,
+                rationale=rationale,
+                token_count=token_count // len(stocks_data),  # Split token cost
+                cached=False,
+            )
+            self._total_decisions += 1
+
+        # Fill in missing stocks with HOLD
+        for stock in stocks_data:
+            code = stock["stock_code"]
+            if code not in decisions:
+                logger.warning("No decision for %s in batch response — using HOLD", code)
+                decisions[code] = TradeDecision(
+                    action="HOLD",
+                    confidence=0,
+                    rationale="Not found in batch response",
+                    token_count=0,
+                    cached=False,
+                )
+
+        logger.info(
+            "Batch decision completed for %d stocks",
+            len(decisions),
+            extra={"tokens": token_count},
+        )
+
+        return decisions

src/brain/prompt_optimizer.py

@@ -0,0 +1,267 @@
+"""Prompt optimization utilities for reducing token usage.
+
+This module provides tools to compress prompts while maintaining decision quality:
+- Token counting
+- Text compression and abbreviation
+- Template-based prompts with variable slots
+- Priority-based context truncation
+"""
+
+from __future__ import annotations
+
+import json
+import re
+from dataclasses import dataclass
+from typing import Any
+
+# Abbreviation mapping for common terms
+ABBREVIATIONS = {
+    "price": "P",
+    "volume": "V",
+    "current": "cur",
+    "previous": "prev",
+    "change": "chg",
+    "percentage": "pct",
+    "market": "mkt",
+    "orderbook": "ob",
+    "foreigner": "fgn",
+    "buy": "B",
+    "sell": "S",
+    "hold": "H",
+    "confidence": "conf",
+    "rationale": "reason",
+    "action": "act",
+    "net": "net",
+}
+
+# Reverse mapping for decompression
+REVERSE_ABBREVIATIONS = {v: k for k, v in ABBREVIATIONS.items()}
+
+
+@dataclass(frozen=True)
+class TokenMetrics:
+    """Metrics about token usage in a prompt."""
+
+    char_count: int
+    word_count: int
+    estimated_tokens: int  # Rough estimate: ~4 chars per token
+    compression_ratio: float = 1.0  # Original / Compressed
+
+
+class PromptOptimizer:
+    """Optimizes prompts to reduce token usage while maintaining quality."""
+
+    @staticmethod
+    def estimate_tokens(text: str) -> int:
+        """Estimate token count for text.
+
+        Uses a simple heuristic: ~4 characters per token for English.
+        This is approximate but sufficient for optimization purposes.
+
+        Args:
+            text: Input text to estimate tokens for
+
+        Returns:
+            Estimated token count
+        """
+        if not text:
+            return 0
+        # Simple estimate: 1 token ≈ 4 characters
+        return max(1, len(text) // 4)
+
+    @staticmethod
+    def count_tokens(text: str) -> TokenMetrics:
+        """Count various metrics for a text.
+
+        Args:
+            text: Input text to analyze
+
+        Returns:
+            TokenMetrics with character, word, and estimated token counts
+        """
+        char_count = len(text)
+        word_count = len(text.split())
+        estimated_tokens = PromptOptimizer.estimate_tokens(text)
+
+        return TokenMetrics(
+            char_count=char_count,
+            word_count=word_count,
+            estimated_tokens=estimated_tokens,
+        )
+
+    @staticmethod
+    def compress_json(data: dict[str, Any]) -> str:
+        """Compress JSON by removing whitespace.
+
+        Args:
+            data: Dictionary to serialize
+
+        Returns:
+            Compact JSON string without whitespace
+        """
+        return json.dumps(data, separators=(",", ":"), ensure_ascii=False)
+
+    @staticmethod
+    def abbreviate_text(text: str, aggressive: bool = False) -> str:
+        """Apply abbreviations to reduce text length.
+
+        Args:
+            text: Input text to abbreviate
+            aggressive: If True, apply more aggressive compression
+
+        Returns:
+            Abbreviated text
+        """
+        result = text
+
+        # Apply word-level abbreviations (case-insensitive)
+        for full, abbr in ABBREVIATIONS.items():
+            # Word boundaries to avoid partial replacements
+            pattern = r"\b" + re.escape(full) + r"\b"
+            result = re.sub(pattern, abbr, result, flags=re.IGNORECASE)
+
+        if aggressive:
+            # Remove articles and filler words
+            result = re.sub(r"\b(a|an|the)\b", "", result, flags=re.IGNORECASE)
+            result = re.sub(r"\b(is|are|was|were)\b", "", result, flags=re.IGNORECASE)
+            # Collapse multiple spaces
+            result = re.sub(r"\s+", " ", result)
+
+        return result.strip()
+
+    @staticmethod
+    def build_compressed_prompt(
+        market_data: dict[str, Any],
+        include_instructions: bool = True,
+        max_length: int | None = None,
+    ) -> str:
+        """Build a compressed prompt from market data.
+
+        Args:
+            market_data: Market data dictionary with stock info
+            include_instructions: Whether to include full instructions
+            max_length: Maximum character length (truncates if needed)
+
+        Returns:
+            Compressed prompt string
+        """
+        # Abbreviated market name
+        market_name = market_data.get("market_name", "KR")
+        if "Korea" in market_name:
+            market_name = "KR"
+        elif "United States" in market_name or "US" in market_name:
+            market_name = "US"
+
+        # Core data - always included
+        core_info = {
+            "mkt": market_name,
+            "code": market_data["stock_code"],
+            "P": market_data["current_price"],
+        }
+
+        # Optional fields
+        if "orderbook" in market_data and market_data["orderbook"]:
+            ob = market_data["orderbook"]
+            # Compress orderbook: keep only top 3 levels
+            compressed_ob = {
+                "bid": ob.get("bid", [])[:3],
+                "ask": ob.get("ask", [])[:3],
+            }
+            core_info["ob"] = compressed_ob
+
+        if market_data.get("foreigner_net", 0) != 0:
+            core_info["fgn_net"] = market_data["foreigner_net"]
+
+        # Compress to JSON
+        data_str = PromptOptimizer.compress_json(core_info)
+
+        if include_instructions:
+            # Minimal instructions
+            prompt = (
+                f"{market_name} trader. Analyze:\n{data_str}\n\n"
+                'Return JSON: {"action":"BUY"|"SELL"|"HOLD","confidence":<0-100>,"rationale":"<text>"}\n'
+                "Rules: action=BUY/SELL/HOLD, confidence=0-100, rationale=concise. No markdown."
+            )
+        else:
+            # Data only (for cached contexts where instructions are known)
+            prompt = data_str
+
+        # Truncate if needed
+        if max_length and len(prompt) > max_length:
+            prompt = prompt[:max_length] + "..."
+
+        return prompt
+
+    @staticmethod
+    def truncate_context(
+        context: dict[str, Any],
+        max_tokens: int,
+        priority_keys: list[str] | None = None,
+    ) -> dict[str, Any]:
+        """Truncate context data to fit within token budget.
+
+        Keeps high-priority keys first, then truncates less important data.
+
+        Args:
+            context: Context dictionary to truncate
+            max_tokens: Maximum token budget
+            priority_keys: List of keys to keep (in order of priority)
+
+        Returns:
+            Truncated context dictionary
+        """
+        if not context:
+            return {}
+
+        if priority_keys is None:
+            priority_keys = []
+
+        result: dict[str, Any] = {}
+        current_tokens = 0
+
+        # Add priority keys first
+        for key in priority_keys:
+            if key in context:
+                value_str = json.dumps(context[key])
+                tokens = PromptOptimizer.estimate_tokens(value_str)
+
+                if current_tokens + tokens <= max_tokens:
+                    result[key] = context[key]
+                    current_tokens += tokens
+                else:
+                    break
+
+        # Add remaining keys if space available
+        for key, value in context.items():
+            if key in result:
+                continue
+
+            value_str = json.dumps(value)
+            tokens = PromptOptimizer.estimate_tokens(value_str)
+
+            if current_tokens + tokens <= max_tokens:
+                result[key] = value
+                current_tokens += tokens
+            else:
+                break
+
+        return result
+
+    @staticmethod
+    def calculate_compression_ratio(original: str, compressed: str) -> float:
+        """Calculate compression ratio between original and compressed text.
+
+        Args:
+            original: Original text
+            compressed: Compressed text
+
+        Returns:
+            Compression ratio (original_tokens / compressed_tokens)
+        """
+        original_tokens = PromptOptimizer.estimate_tokens(original)
+        compressed_tokens = PromptOptimizer.estimate_tokens(compressed)
+
+        if compressed_tokens == 0:
+            return 1.0
+
+        return original_tokens / compressed_tokens

src/broker/kis_api.py

@@ -8,7 +8,7 @@ from __future__ import annotations
 import asyncio
 import logging
 import ssl
-from typing import Any
+from typing import Any, cast
 
 import aiohttp
 
@@ -20,6 +20,39 @@ _KIS_VTS_HOST = "openapivts.koreainvestment.com"
 logger = logging.getLogger(__name__)
 
 
+def kr_tick_unit(price: float) -> int:
+    """Return KRX tick size for the given price level.
+
+    KRX price tick rules (domestic stocks):
+        price < 2,000        →   1원
+        2,000  ≤ price < 5,000     →   5원
+        5,000  ≤ price < 20,000    →  10원
+        20,000 ≤ price < 50,000    →  50원
+        50,000 ≤ price < 200,000   → 100원
+        200,000 ≤ price < 500,000  → 500원
+        500,000 ≤ price            → 1,000원
+    """
+    if price < 2_000:
+        return 1
+    if price < 5_000:
+        return 5
+    if price < 20_000:
+        return 10
+    if price < 50_000:
+        return 50
+    if price < 200_000:
+        return 100
+    if price < 500_000:
+        return 500
+    return 1_000
+
+
+def kr_round_down(price: float) -> int:
+    """Round *down* price to the nearest KRX tick unit."""
+    tick = kr_tick_unit(price)
+    return int(price // tick * tick)
+
+
 class LeakyBucket:
     """Simple leaky-bucket rate limiter for async code."""
 
@@ -55,6 +88,9 @@ class KISBroker:
         self._session: aiohttp.ClientSession | None = None
         self._access_token: str | None = None
         self._token_expires_at: float = 0.0
+        self._token_lock = asyncio.Lock()
+        self._last_refresh_attempt: float = 0.0
+        self._refresh_cooldown: float = 60.0  # Seconds (matches KIS 1/minute limit)
         self._rate_limiter = LeakyBucket(settings.RATE_LIMIT_RPS)
 
     def _get_session(self) -> aiohttp.ClientSession:
@@ -80,30 +116,56 @@ class KISBroker:
     # ------------------------------------------------------------------
 
     async def _ensure_token(self) -> str:
-        """Return a valid access token, refreshing if expired."""
+        """Return a valid access token, refreshing if expired.
+
+        Uses a lock to prevent concurrent token refresh attempts that would
+        hit the API's 1-per-minute rate limit (EGW00133).
+        """
+        # Fast path: check without lock
         now = asyncio.get_event_loop().time()
         if self._access_token and now < self._token_expires_at:
             return self._access_token
 
-        logger.info("Refreshing KIS access token")
-        session = self._get_session()
-        url = f"{self._base_url}/oauth2/tokenP"
-        body = {
-            "grant_type": "client_credentials",
-            "appkey": self._app_key,
-            "appsecret": self._app_secret,
-        }
+        # Slow path: acquire lock and refresh
+        async with self._token_lock:
+            # Re-check after acquiring lock (another coroutine may have refreshed)
+            now = asyncio.get_event_loop().time()
+            if self._access_token and now < self._token_expires_at:
+                return self._access_token
 
-        async with session.post(url, json=body) as resp:
-            if resp.status != 200:
-                text = await resp.text()
-                raise ConnectionError(f"Token refresh failed ({resp.status}): {text}")
-            data = await resp.json()
+            # Check cooldown period (prevents hitting EGW00133: 1/minute limit)
+            time_since_last_attempt = now - self._last_refresh_attempt
+            if time_since_last_attempt < self._refresh_cooldown:
+                remaining = self._refresh_cooldown - time_since_last_attempt
+                # Do not fail fast here. If token is unavailable, upstream calls
+                # will all fail for up to a minute and scanning returns no trades.
+                logger.warning(
+                    "Token refresh on cooldown. Waiting %.1fs before retry (KIS allows 1/minute)",
+                    remaining,
+                )
+                await asyncio.sleep(remaining)
+                now = asyncio.get_event_loop().time()
 
-        self._access_token = data["access_token"]
-        self._token_expires_at = now + data.get("expires_in", 86400) - 60  # 1-min buffer
-        logger.info("Token refreshed successfully")
-        return self._access_token
+            logger.info("Refreshing KIS access token")
+            self._last_refresh_attempt = now
+            session = self._get_session()
+            url = f"{self._base_url}/oauth2/tokenP"
+            body = {
+                "grant_type": "client_credentials",
+                "appkey": self._app_key,
+                "appsecret": self._app_secret,
+            }
+
+            async with session.post(url, json=body) as resp:
+                if resp.status != 200:
+                    text = await resp.text()
+                    raise ConnectionError(f"Token refresh failed ({resp.status}): {text}")
+                data = await resp.json()
+
+            self._access_token = data["access_token"]
+            self._token_expires_at = now + data.get("expires_in", 86400) - 60  # 1-min buffer
+            logger.info("Token refreshed successfully")
+            return self._access_token
 
     # ------------------------------------------------------------------
     # Hash Key (required for POST bodies)
@@ -111,6 +173,7 @@ class KISBroker:
 
     async def _get_hash_key(self, body: dict[str, Any]) -> str:
         """Request a hash key from KIS for POST request body signing."""
+        await self._rate_limiter.acquire()
         session = self._get_session()
         url = f"{self._base_url}/uapi/hashkey"
         headers = {
@@ -168,12 +231,64 @@ class KISBroker:
         except (TimeoutError, aiohttp.ClientError) as exc:
             raise ConnectionError(f"Network error fetching orderbook: {exc}") from exc
 
+    async def get_current_price(
+        self, stock_code: str
+    ) -> tuple[float, float, float]:
+        """Fetch current price data for a domestic stock.
+
+        Uses the ``inquire-price`` API (FHKST01010100), which works in both
+        real and VTS environments and returns the actual last-traded price.
+
+        Returns:
+            (current_price, prdy_ctrt, frgn_ntby_qty)
+            - current_price: Last traded price in KRW.
+            - prdy_ctrt: Day change rate (%).
+            - frgn_ntby_qty: Foreigner net buy quantity.
+        """
+        await self._rate_limiter.acquire()
+        session = self._get_session()
+
+        headers = await self._auth_headers("FHKST01010100")
+        params = {
+            "FID_COND_MRKT_DIV_CODE": "J",
+            "FID_INPUT_ISCD": stock_code,
+        }
+        url = f"{self._base_url}/uapi/domestic-stock/v1/quotations/inquire-price"
+
+        def _f(val: str | None) -> float:
+            try:
+                return float(val or "0")
+            except ValueError:
+                return 0.0
+
+        try:
+            async with session.get(url, headers=headers, params=params) as resp:
+                if resp.status != 200:
+                    text = await resp.text()
+                    raise ConnectionError(
+                        f"get_current_price failed ({resp.status}): {text}"
+                    )
+                data = await resp.json()
+                out = data.get("output", {})
+                return (
+                    _f(out.get("stck_prpr")),
+                    _f(out.get("prdy_ctrt")),
+                    _f(out.get("frgn_ntby_qty")),
+                )
+        except (TimeoutError, aiohttp.ClientError) as exc:
+            raise ConnectionError(
+                f"Network error fetching current price: {exc}"
+            ) from exc
+
     async def get_balance(self) -> dict[str, Any]:
         """Fetch current account balance and holdings."""
         await self._rate_limiter.acquire()
         session = self._get_session()
 
-        headers = await self._auth_headers("VTTC8434R")  # 모의투자 잔고조회
+        # TR_ID: 실전 TTTC8434R, 모의 VTTC8434R
+        # Source: 한국투자증권 오픈API 전체문서 (20260221) — '국내주식 잔고조회' 시트
+        tr_id = "TTTC8434R" if self._settings.MODE == "live" else "VTTC8434R"
+        headers = await self._auth_headers(tr_id)
         params = {
             "CANO": self._account_no,
             "ACNT_PRDT_CD": self._product_cd,
@@ -218,14 +333,30 @@ class KISBroker:
         await self._rate_limiter.acquire()
         session = self._get_session()
 
-        tr_id = "VTTC0802U" if order_type == "BUY" else "VTTC0801U"
+        # TR_ID: 실전 BUY=TTTC0012U SELL=TTTC0011U, 모의 BUY=VTTC0012U SELL=VTTC0011U
+        # Source: 한국투자증권 오픈API 전체문서 (20260221) — '주식주문(현금)' 시트
+        # ※ TTTC0802U/VTTC0802U는 미수매수(증거금40% 계좌 전용) — 현금주문에 사용 금지
+        if self._settings.MODE == "live":
+            tr_id = "TTTC0012U" if order_type == "BUY" else "TTTC0011U"
+        else:
+            tr_id = "VTTC0012U" if order_type == "BUY" else "VTTC0011U"
+
+        # KRX requires limit orders to be rounded down to the tick unit.
+        # ORD_DVSN: "00"=지정가, "01"=시장가
+        if price > 0:
+            ord_dvsn = "00"  # 지정가
+            ord_price = kr_round_down(price)
+        else:
+            ord_dvsn = "01"  # 시장가
+            ord_price = 0
+
         body = {
             "CANO": self._account_no,
             "ACNT_PRDT_CD": self._product_cd,
             "PDNO": stock_code,
-            "ORD_DVSN": "01" if price > 0 else "06",  # 01=지정가, 06=시장가
+            "ORD_DVSN": ord_dvsn,
             "ORD_QTY": str(quantity),
-            "ORD_UNPR": str(price),
+            "ORD_UNPR": str(ord_price),
         }
 
         hash_key = await self._get_hash_key(body)
@@ -252,3 +383,279 @@ class KISBroker:
                 return data
         except (TimeoutError, aiohttp.ClientError) as exc:
             raise ConnectionError(f"Network error sending order: {exc}") from exc
+
+    async def fetch_market_rankings(
+        self,
+        ranking_type: str = "volume",
+        limit: int = 30,
+    ) -> list[dict[str, Any]]:
+        """Fetch market rankings from KIS API.
+
+        Args:
+            ranking_type: Type of ranking ("volume" or "fluctuation")
+            limit: Maximum number of results to return
+
+        Returns:
+            List of stock data dicts with keys: stock_code, name, price, volume,
+            change_rate, volume_increase_rate
+
+        Raises:
+            ConnectionError: If API request fails
+        """
+        await self._rate_limiter.acquire()
+        session = self._get_session()
+
+        if ranking_type == "volume":
+            # 거래량순위: FHPST01710000 / /quotations/volume-rank
+            tr_id = "FHPST01710000"
+            url = f"{self._base_url}/uapi/domestic-stock/v1/quotations/volume-rank"
+            params: dict[str, str] = {
+                "FID_COND_MRKT_DIV_CODE": "J",
+                "FID_COND_SCR_DIV_CODE": "20171",
+                "FID_INPUT_ISCD": "0000",
+                "FID_DIV_CLS_CODE": "0",
+                "FID_BLNG_CLS_CODE": "0",
+                "FID_TRGT_CLS_CODE": "111111111",
+                "FID_TRGT_EXLS_CLS_CODE": "0000000000",
+                "FID_INPUT_PRICE_1": "0",
+                "FID_INPUT_PRICE_2": "0",
+                "FID_VOL_CNT": "0",
+                "FID_INPUT_DATE_1": "",
+            }
+        else:
+            # 등락률순위: FHPST01700000 / /ranking/fluctuation (소문자 파라미터)
+            tr_id = "FHPST01700000"
+            url = f"{self._base_url}/uapi/domestic-stock/v1/ranking/fluctuation"
+            params = {
+                "fid_cond_mrkt_div_code": "J",
+                "fid_cond_scr_div_code": "20170",
+                "fid_input_iscd": "0000",
+                "fid_rank_sort_cls_code": "0",
+                "fid_input_cnt_1": str(limit),
+                "fid_prc_cls_code": "0",
+                "fid_input_price_1": "0",
+                "fid_input_price_2": "0",
+                "fid_vol_cnt": "0",
+                "fid_trgt_cls_code": "0",
+                "fid_trgt_exls_cls_code": "0",
+                "fid_div_cls_code": "0",
+                "fid_rsfl_rate1": "0",
+                "fid_rsfl_rate2": "0",
+            }
+
+        headers = await self._auth_headers(tr_id)
+
+        try:
+            async with session.get(url, headers=headers, params=params) as resp:
+                if resp.status != 200:
+                    text = await resp.text()
+                    raise ConnectionError(
+                        f"fetch_market_rankings failed ({resp.status}): {text}"
+                    )
+                data = await resp.json()
+
+            # Parse response - output is a list of ranked stocks
+            def _safe_float(value: str | float | None, default: float = 0.0) -> float:
+                if value is None or value == "":
+                    return default
+                try:
+                    return float(value)
+                except (ValueError, TypeError):
+                    return default
+
+            rankings = []
+            for item in data.get("output", [])[:limit]:
+                rankings.append({
+                    "stock_code": item.get("stck_shrn_iscd") or item.get("mksc_shrn_iscd", ""),
+                    "name": item.get("hts_kor_isnm", ""),
+                    "price": _safe_float(item.get("stck_prpr", "0")),
+                    "volume": _safe_float(item.get("acml_vol", "0")),
+                    "change_rate": _safe_float(item.get("prdy_ctrt", "0")),
+                    "volume_increase_rate": _safe_float(item.get("vol_inrt", "0")),
+                })
+            return rankings
+
+        except (TimeoutError, aiohttp.ClientError) as exc:
+            raise ConnectionError(f"Network error fetching rankings: {exc}") from exc
+
+    async def get_domestic_pending_orders(self) -> list[dict[str, Any]]:
+        """Fetch unfilled (pending) domestic limit orders.
+
+        The KIS pending-orders API (TTTC0084R) is unsupported in paper (VTS)
+        mode, so this method returns an empty list immediately when MODE is
+        not "live".
+
+        Returns:
+            List of pending order dicts from the KIS ``output`` field.
+            Each dict includes keys such as ``odno``, ``orgn_odno``,
+            ``ord_gno_brno``, ``psbl_qty``, ``sll_buy_dvsn_cd``, ``pdno``.
+        """
+        if self._settings.MODE != "live":
+            logger.debug(
+                "get_domestic_pending_orders: paper mode — TTTC0084R unsupported, returning []"
+            )
+            return []
+
+        await self._rate_limiter.acquire()
+        session = self._get_session()
+
+        # TR_ID: 실전 TTTC0084R (모의 미지원)
+        # Source: 한국투자증권 오픈API 전체문서 (20260221) — '주식 미체결조회' 시트
+        headers = await self._auth_headers("TTTC0084R")
+        params = {
+            "CANO": self._account_no,
+            "ACNT_PRDT_CD": self._product_cd,
+            "INQR_DVSN_1": "0",
+            "INQR_DVSN_2": "0",
+            "CTX_AREA_FK100": "",
+            "CTX_AREA_NK100": "",
+        }
+        url = f"{self._base_url}/uapi/domestic-stock/v1/trading/inquire-psbl-rvsecncl"
+
+        try:
+            async with session.get(url, headers=headers, params=params) as resp:
+                if resp.status != 200:
+                    text = await resp.text()
+                    raise ConnectionError(
+                        f"get_domestic_pending_orders failed ({resp.status}): {text}"
+                    )
+                data = await resp.json()
+                return data.get("output", []) or []
+        except (TimeoutError, aiohttp.ClientError) as exc:
+            raise ConnectionError(
+                f"Network error fetching domestic pending orders: {exc}"
+            ) from exc
+
+    async def cancel_domestic_order(
+        self,
+        stock_code: str,
+        orgn_odno: str,
+        krx_fwdg_ord_orgno: str,
+        qty: int,
+    ) -> dict[str, Any]:
+        """Cancel an unfilled domestic limit order.
+
+        Args:
+            stock_code: 6-digit domestic stock code (``pdno``).
+            orgn_odno: Original order number from pending-orders response
+                (``orgn_odno`` field).
+            krx_fwdg_ord_orgno: KRX forwarding order branch number from
+                pending-orders response (``ord_gno_brno`` field).
+            qty: Quantity to cancel (use ``psbl_qty`` from pending order).
+
+        Returns:
+            Raw KIS API response dict (check ``rt_cd == "0"`` for success).
+        """
+        await self._rate_limiter.acquire()
+        session = self._get_session()
+
+        # TR_ID: 실전 TTTC0013U, 모의 VTTC0013U
+        # Source: 한국투자증권 오픈API 전체문서 (20260221) — '주식주문(정정취소)' 시트
+        tr_id = "TTTC0013U" if self._settings.MODE == "live" else "VTTC0013U"
+
+        body = {
+            "CANO": self._account_no,
+            "ACNT_PRDT_CD": self._product_cd,
+            "KRX_FWDG_ORD_ORGNO": krx_fwdg_ord_orgno,
+            "ORGN_ODNO": orgn_odno,
+            "ORD_DVSN": "00",
+            "ORD_QTY": str(qty),
+            "ORD_UNPR": "0",
+            "RVSE_CNCL_DVSN_CD": "02",
+            "QTY_ALL_ORD_YN": "Y",
+        }
+
+        hash_key = await self._get_hash_key(body)
+        headers = await self._auth_headers(tr_id)
+        headers["hashkey"] = hash_key
+
+        url = f"{self._base_url}/uapi/domestic-stock/v1/trading/order-rvsecncl"
+
+        try:
+            async with session.post(url, headers=headers, json=body) as resp:
+                if resp.status != 200:
+                    text = await resp.text()
+                    raise ConnectionError(
+                        f"cancel_domestic_order failed ({resp.status}): {text}"
+                    )
+                return cast(dict[str, Any], await resp.json())
+        except (TimeoutError, aiohttp.ClientError) as exc:
+            raise ConnectionError(
+                f"Network error cancelling domestic order: {exc}"
+            ) from exc
+
+    async def get_daily_prices(
+        self,
+        stock_code: str,
+        days: int = 20,
+    ) -> list[dict[str, Any]]:
+        """Fetch daily OHLCV price history for a stock.
+
+        Args:
+            stock_code: 6-digit stock code
+            days: Number of trading days to fetch (default 20 for RSI calculation)
+
+        Returns:
+            List of daily price dicts with keys: date, open, high, low, close, volume
+            Sorted oldest to newest
+
+        Raises:
+            ConnectionError: If API request fails
+        """
+        await self._rate_limiter.acquire()
+        session = self._get_session()
+
+        headers = await self._auth_headers("FHKST03010100")
+
+        # Calculate date range (today and N days ago)
+        from datetime import datetime, timedelta
+        end_date = datetime.now().strftime("%Y%m%d")
+        start_date = (datetime.now() - timedelta(days=days + 10)).strftime("%Y%m%d")
+
+        params = {
+            "FID_COND_MRKT_DIV_CODE": "J",
+            "FID_INPUT_ISCD": stock_code,
+            "FID_INPUT_DATE_1": start_date,
+            "FID_INPUT_DATE_2": end_date,
+            "FID_PERIOD_DIV_CODE": "D",  # Daily
+            "FID_ORG_ADJ_PRC": "0",  # Adjusted price
+        }
+
+        url = f"{self._base_url}/uapi/domestic-stock/v1/quotations/inquire-daily-itemchartprice"
+
+        try:
+            async with session.get(url, headers=headers, params=params) as resp:
+                if resp.status != 200:
+                    text = await resp.text()
+                    raise ConnectionError(
+                        f"get_daily_prices failed ({resp.status}): {text}"
+                    )
+                data = await resp.json()
+
+            # Parse response
+            def _safe_float(value: str | float | None, default: float = 0.0) -> float:
+                if value is None or value == "":
+                    return default
+                try:
+                    return float(value)
+                except (ValueError, TypeError):
+                    return default
+
+            prices = []
+            for item in data.get("output2", []):
+                prices.append({
+                    "date": item.get("stck_bsop_date", ""),
+                    "open": _safe_float(item.get("stck_oprc", "0")),
+                    "high": _safe_float(item.get("stck_hgpr", "0")),
+                    "low": _safe_float(item.get("stck_lwpr", "0")),
+                    "close": _safe_float(item.get("stck_clpr", "0")),
+                    "volume": _safe_float(item.get("acml_vol", "0")),
+                })
+
+            # Sort oldest to newest (KIS returns newest first)
+            prices.reverse()
+
+            return prices[:days]  # Return only requested number of days
+
+        except (TimeoutError, aiohttp.ClientError) as exc:
+            raise ConnectionError(f"Network error fetching daily prices: {exc}") from exc

src/broker/overseas.py

@@ -12,6 +12,38 @@ from src.broker.kis_api import KISBroker
 logger = logging.getLogger(__name__)
 
 
+# Ranking API uses different exchange codes than order/quote APIs.
+_RANKING_EXCHANGE_MAP: dict[str, str] = {
+    "NASD": "NAS",
+    "NYSE": "NYS",
+    "AMEX": "AMS",
+    "SEHK": "HKS",
+    "SHAA": "SHS",
+    "SZAA": "SZS",
+    "HSX": "HSX",
+    "HNX": "HNX",
+    "TSE": "TSE",
+}
+
+# Price inquiry API (HHDFS00000300) uses the same short exchange codes as rankings.
+# NASD → NAS, NYSE → NYS, AMEX → AMS (confirmed: AMEX returns empty, AMS returns price).
+_PRICE_EXCHANGE_MAP: dict[str, str] = _RANKING_EXCHANGE_MAP
+
+# Cancel order TR_IDs per exchange code — (live_tr_id, paper_tr_id).
+# Source: 한국투자증권 오픈API 전체문서 (20260221) — '해외주식 주문취소' 시트
+_CANCEL_TR_ID_MAP: dict[str, tuple[str, str]] = {
+    "NASD": ("TTTT1004U", "VTTT1004U"),
+    "NYSE": ("TTTT1004U", "VTTT1004U"),
+    "AMEX": ("TTTT1004U", "VTTT1004U"),
+    "SEHK": ("TTTS1003U", "VTTS1003U"),
+    "TSE":  ("TTTS0309U", "VTTS0309U"),
+    "SHAA": ("TTTS0302U", "VTTS0302U"),
+    "SZAA": ("TTTS0306U", "VTTS0306U"),
+    "HNX":  ("TTTS0312U", "VTTS0312U"),
+    "HSX":  ("TTTS0312U", "VTTS0312U"),
+}
+
+
 class OverseasBroker:
     """KIS Overseas Stock API wrapper that reuses KISBroker infrastructure."""
 
@@ -44,9 +76,11 @@ class OverseasBroker:
         session = self._broker._get_session()
 
         headers = await self._broker._auth_headers("HHDFS00000300")
+        # Map internal exchange codes to the short form expected by the price API.
+        price_excd = _PRICE_EXCHANGE_MAP.get(exchange_code, exchange_code)
         params = {
             "AUTH": "",
-            "EXCD": exchange_code,
+            "EXCD": price_excd,
             "SYMB": stock_code,
         }
         url = f"{self._broker._base_url}/uapi/overseas-price/v1/quotations/price"
@@ -64,6 +98,83 @@ class OverseasBroker:
                 f"Network error fetching overseas price: {exc}"
             ) from exc
 
+    async def fetch_overseas_rankings(
+        self,
+        exchange_code: str,
+        ranking_type: str = "fluctuation",
+        limit: int = 30,
+    ) -> list[dict[str, Any]]:
+        """Fetch overseas rankings (price change or volume surge).
+
+        Ranking API specs may differ by account/product. Endpoint paths and
+        TR_IDs are configurable via settings and can be overridden in .env.
+        """
+        if not self._broker._settings.OVERSEAS_RANKING_ENABLED:
+            return []
+
+        await self._broker._rate_limiter.acquire()
+        session = self._broker._get_session()
+
+        ranking_excd = _RANKING_EXCHANGE_MAP.get(exchange_code, exchange_code)
+
+        if ranking_type == "volume":
+            tr_id = self._broker._settings.OVERSEAS_RANKING_VOLUME_TR_ID
+            path = self._broker._settings.OVERSEAS_RANKING_VOLUME_PATH
+            params: dict[str, str] = {
+                "KEYB": "",  # NEXT KEY BUFF — Required, 공백
+                "AUTH": "",
+                "EXCD": ranking_excd,
+                "MIXN": "0",
+                "VOL_RANG": "0",
+            }
+        else:
+            tr_id = self._broker._settings.OVERSEAS_RANKING_FLUCT_TR_ID
+            path = self._broker._settings.OVERSEAS_RANKING_FLUCT_PATH
+            params = {
+                "KEYB": "",  # NEXT KEY BUFF — Required, 공백
+                "AUTH": "",
+                "EXCD": ranking_excd,
+                "NDAY": "0",
+                "GUBN": "1",  # 0=하락율, 1=상승율 — 변동성 스캐너는 급등 종목 우선
+                "VOL_RANG": "0",
+            }
+
+        headers = await self._broker._auth_headers(tr_id)
+        url = f"{self._broker._base_url}{path}"
+
+        try:
+            async with session.get(url, headers=headers, params=params) as resp:
+                if resp.status != 200:
+                    text = await resp.text()
+                    if resp.status == 404:
+                        logger.warning(
+                            "Overseas ranking endpoint unavailable (404) for %s/%s; "
+                            "using symbol fallback scan",
+                            exchange_code,
+                            ranking_type,
+                        )
+                        return []
+                    raise ConnectionError(
+                        f"fetch_overseas_rankings failed ({resp.status}): {text}"
+                    )
+
+                data = await resp.json()
+                rows = self._extract_ranking_rows(data)
+                if rows:
+                    return rows[:limit]
+
+                logger.debug(
+                    "Overseas ranking returned empty for %s/%s (keys=%s)",
+                    exchange_code,
+                    ranking_type,
+                    list(data.keys()),
+                )
+                return []
+        except (TimeoutError, aiohttp.ClientError) as exc:
+            raise ConnectionError(
+                f"Network error fetching overseas rankings: {exc}"
+            ) from exc
+
     async def get_overseas_balance(self, exchange_code: str) -> dict[str, Any]:
         """
         Fetch overseas account balance.
@@ -80,8 +191,12 @@ class OverseasBroker:
         await self._broker._rate_limiter.acquire()
         session = self._broker._get_session()
 
-        # Virtual trading TR_ID for overseas balance inquiry
-        headers = await self._broker._auth_headers("VTTS3012R")
+        # TR_ID: 실전 TTTS3012R, 모의 VTTS3012R
+        # Source: 한국투자증권 오픈API 전체문서 (20260221) — '해외주식 잔고조회' 시트
+        balance_tr_id = (
+            "TTTS3012R" if self._broker._settings.MODE == "live" else "VTTS3012R"
+        )
+        headers = await self._broker._auth_headers(balance_tr_id)
         params = {
             "CANO": self._broker._account_no,
             "ACNT_PRDT_CD": self._broker._product_cd,
@@ -107,6 +222,59 @@ class OverseasBroker:
                 f"Network error fetching overseas balance: {exc}"
             ) from exc
 
+    async def get_overseas_buying_power(
+        self,
+        exchange_code: str,
+        stock_code: str,
+        price: float,
+    ) -> dict[str, Any]:
+        """
+        Fetch overseas buying power for a specific stock and price.
+
+        Args:
+            exchange_code: Exchange code (e.g., "NASD", "NYSE")
+            stock_code: Stock ticker symbol
+            price: Current stock price (used for quantity calculation)
+
+        Returns:
+            API response; key field: output.ord_psbl_frcr_amt (주문가능외화금액)
+
+        Raises:
+            ConnectionError: On network or API errors
+        """
+        await self._broker._rate_limiter.acquire()
+        session = self._broker._get_session()
+
+        # TR_ID: 실전 TTTS3007R, 모의 VTTS3007R
+        # Source: 한국투자증권 오픈API 전체문서 (20260221) — '해외주식 매수가능금액조회' 시트
+        ps_tr_id = (
+            "TTTS3007R" if self._broker._settings.MODE == "live" else "VTTS3007R"
+        )
+        headers = await self._broker._auth_headers(ps_tr_id)
+        params = {
+            "CANO": self._broker._account_no,
+            "ACNT_PRDT_CD": self._broker._product_cd,
+            "OVRS_EXCG_CD": exchange_code,
+            "OVRS_ORD_UNPR": f"{price:.2f}",
+            "ITEM_CD": stock_code,
+        }
+        url = (
+            f"{self._broker._base_url}/uapi/overseas-stock/v1/trading/inquire-psamount"
+        )
+
+        try:
+            async with session.get(url, headers=headers, params=params) as resp:
+                if resp.status != 200:
+                    text = await resp.text()
+                    raise ConnectionError(
+                        f"get_overseas_buying_power failed ({resp.status}): {text}"
+                    )
+                return await resp.json()
+        except (TimeoutError, aiohttp.ClientError) as exc:
+            raise ConnectionError(
+                f"Network error fetching overseas buying power: {exc}"
+            ) from exc
+
     async def send_overseas_order(
         self,
         exchange_code: str,
@@ -134,8 +302,12 @@ class OverseasBroker:
         await self._broker._rate_limiter.acquire()
         session = self._broker._get_session()
 
-        # Virtual trading TR_IDs for overseas orders
-        tr_id = "VTTT1002U" if order_type == "BUY" else "VTTT1006U"
+        # TR_ID: 실전 BUY=TTTT1002U SELL=TTTT1006U, 모의 BUY=VTTT1002U SELL=VTTT1001U
+        # Source: 한국투자증권 오픈API 전체문서 (20260221) — '해외주식 주문' 시트
+        if self._broker._settings.MODE == "live":
+            tr_id = "TTTT1002U" if order_type == "BUY" else "TTTT1006U"
+        else:
+            tr_id = "VTTT1002U" if order_type == "BUY" else "VTTT1001U"
 
         body = {
             "CANO": self._broker._account_no,
@@ -162,20 +334,158 @@ class OverseasBroker:
                         f"send_overseas_order failed ({resp.status}): {text}"
                     )
                 data = await resp.json()
-                logger.info(
-                    "Overseas order submitted",
-                    extra={
-                        "exchange": exchange_code,
-                        "stock_code": stock_code,
-                        "action": order_type,
-                    },
-                )
+                rt_cd = data.get("rt_cd", "")
+                msg1 = data.get("msg1", "")
+                if rt_cd == "0":
+                    logger.info(
+                        "Overseas order submitted",
+                        extra={
+                            "exchange": exchange_code,
+                            "stock_code": stock_code,
+                            "action": order_type,
+                        },
+                    )
+                else:
+                    logger.warning(
+                        "Overseas order rejected (rt_cd=%s): %s [%s %s %s qty=%d]",
+                        rt_cd,
+                        msg1,
+                        order_type,
+                        stock_code,
+                        exchange_code,
+                        quantity,
+                    )
                 return data
         except (TimeoutError, aiohttp.ClientError) as exc:
             raise ConnectionError(
                 f"Network error sending overseas order: {exc}"
             ) from exc
 
+    async def get_overseas_pending_orders(
+        self, exchange_code: str
+    ) -> list[dict[str, Any]]:
+        """Fetch unfilled (pending) overseas orders for a given exchange.
+
+        Args:
+            exchange_code: Exchange code (e.g., "NASD", "SEHK").
+                For US markets, NASD returns all US pending orders (NASD/NYSE/AMEX).
+
+        Returns:
+            List of pending order dicts with fields: odno, pdno, sll_buy_dvsn_cd,
+            ft_ord_qty, nccs_qty, ft_ord_unpr3, ovrs_excg_cd.
+            Always returns [] in paper mode (TTTS3018R is live-only).
+
+        Raises:
+            ConnectionError: On network or API errors (live mode only).
+        """
+        if self._broker._settings.MODE != "live":
+            logger.debug(
+                "Pending orders API (TTTS3018R) not supported in paper mode; returning []"
+            )
+            return []
+
+        await self._broker._rate_limiter.acquire()
+        session = self._broker._get_session()
+
+        # TTTS3018R: 해외주식 미체결내역조회 (실전 전용)
+        # Source: 한국투자증권 오픈API 전체문서 (20260221) — '해외주식 미체결조회' 시트
+        headers = await self._broker._auth_headers("TTTS3018R")
+        params = {
+            "CANO": self._broker._account_no,
+            "ACNT_PRDT_CD": self._broker._product_cd,
+            "OVRS_EXCG_CD": exchange_code,
+            "SORT_SQN": "DS",
+            "CTX_AREA_FK200": "",
+            "CTX_AREA_NK200": "",
+        }
+        url = (
+            f"{self._broker._base_url}/uapi/overseas-stock/v1/trading/inquire-nccs"
+        )
+
+        try:
+            async with session.get(url, headers=headers, params=params) as resp:
+                if resp.status != 200:
+                    text = await resp.text()
+                    raise ConnectionError(
+                        f"get_overseas_pending_orders failed ({resp.status}): {text}"
+                    )
+                data = await resp.json()
+                output = data.get("output", [])
+                if isinstance(output, list):
+                    return output
+                return []
+        except (TimeoutError, aiohttp.ClientError) as exc:
+            raise ConnectionError(
+                f"Network error fetching pending orders: {exc}"
+            ) from exc
+
+    async def cancel_overseas_order(
+        self,
+        exchange_code: str,
+        stock_code: str,
+        odno: str,
+        qty: int,
+    ) -> dict[str, Any]:
+        """Cancel an overseas limit order.
+
+        Args:
+            exchange_code: Exchange code (e.g., "NASD", "SEHK").
+            stock_code: Stock ticker symbol.
+            odno: Original order number to cancel.
+            qty: Unfilled quantity to cancel.
+
+        Returns:
+            API response dict containing rt_cd and msg1.
+
+        Raises:
+            ValueError: If exchange_code has no cancel TR_ID mapping.
+            ConnectionError: On network or API errors.
+        """
+        tr_ids = _CANCEL_TR_ID_MAP.get(exchange_code)
+        if tr_ids is None:
+            raise ValueError(f"No cancel TR_ID mapping for exchange: {exchange_code}")
+        live_tr_id, paper_tr_id = tr_ids
+        tr_id = live_tr_id if self._broker._settings.MODE == "live" else paper_tr_id
+
+        await self._broker._rate_limiter.acquire()
+        session = self._broker._get_session()
+
+        # RVSE_CNCL_DVSN_CD="02" means cancel (not revision).
+        # OVRS_ORD_UNPR must be "0" for cancellations.
+        # Source: 한국투자증권 오픈API 전체문서 (20260221) — '해외주식 정정취소주문' 시트
+        body = {
+            "CANO": self._broker._account_no,
+            "ACNT_PRDT_CD": self._broker._product_cd,
+            "OVRS_EXCG_CD": exchange_code,
+            "PDNO": stock_code,
+            "ORGN_ODNO": odno,
+            "RVSE_CNCL_DVSN_CD": "02",
+            "ORD_QTY": str(qty),
+            "OVRS_ORD_UNPR": "0",
+            "ORD_SVR_DVSN_CD": "0",
+        }
+
+        hash_key = await self._broker._get_hash_key(body)
+        headers = await self._broker._auth_headers(tr_id)
+        headers["hashkey"] = hash_key
+
+        url = (
+            f"{self._broker._base_url}/uapi/overseas-stock/v1/trading/order-rvsecncl"
+        )
+
+        try:
+            async with session.post(url, headers=headers, json=body) as resp:
+                if resp.status != 200:
+                    text = await resp.text()
+                    raise ConnectionError(
+                        f"cancel_overseas_order failed ({resp.status}): {text}"
+                    )
+                return await resp.json()
+        except (TimeoutError, aiohttp.ClientError) as exc:
+            raise ConnectionError(
+                f"Network error cancelling overseas order: {exc}"
+            ) from exc
+
     def _get_currency_code(self, exchange_code: str) -> str:
         """
         Map exchange code to currency code.
@@ -198,3 +508,11 @@ class OverseasBroker:
             "HSX": "VND",
         }
         return currency_map.get(exchange_code, "USD")
+
+    def _extract_ranking_rows(self, data: dict[str, Any]) -> list[dict[str, Any]]:
+        """Extract list rows from ranking response across schema variants."""
+        candidates = [data.get("output"), data.get("output1"), data.get("output2")]
+        for value in candidates:
+            if isinstance(value, list):
+                return [row for row in value if isinstance(row, dict)]
+        return []

src/config.py

@@ -13,28 +13,116 @@ class Settings(BaseSettings):
     KIS_APP_KEY: str
     KIS_APP_SECRET: str
     KIS_ACCOUNT_NO: str  # format: "XXXXXXXX-XX"
-    KIS_BASE_URL: str = "https://openapivts.koreainvestment.com:9443"
+    KIS_BASE_URL: str = "https://openapivts.koreainvestment.com:29443"
 
     # Google Gemini
     GEMINI_API_KEY: str
-    GEMINI_MODEL: str = "gemini-pro"
+    GEMINI_MODEL: str = "gemini-2.0-flash"
+
+    # External Data APIs (optional — for data-driven decisions)
+    NEWS_API_KEY: str | None = None
+    NEWS_API_PROVIDER: str = "alphavantage"  # "alphavantage" or "newsapi"
+    MARKET_DATA_API_KEY: str | None = None
+
+    # Legacy field names (for backward compatibility)
+    ALPHA_VANTAGE_API_KEY: str | None = None
+    NEWSAPI_KEY: str | None = None
 
     # Risk Management
     CIRCUIT_BREAKER_PCT: float = Field(default=-3.0, le=0.0)
     FAT_FINGER_PCT: float = Field(default=30.0, gt=0.0, le=100.0)
     CONFIDENCE_THRESHOLD: int = Field(default=80, ge=0, le=100)
 
+    # Smart Scanner Configuration
+    RSI_OVERSOLD_THRESHOLD: int = Field(default=30, ge=0, le=50)
+    RSI_MOMENTUM_THRESHOLD: int = Field(default=70, ge=50, le=100)
+    VOL_MULTIPLIER: float = Field(default=2.0, gt=1.0, le=10.0)
+    SCANNER_TOP_N: int = Field(default=3, ge=1, le=10)
+    POSITION_SIZING_ENABLED: bool = True
+    POSITION_BASE_ALLOCATION_PCT: float = Field(default=5.0, gt=0.0, le=30.0)
+    POSITION_MIN_ALLOCATION_PCT: float = Field(default=1.0, gt=0.0, le=20.0)
+    POSITION_MAX_ALLOCATION_PCT: float = Field(default=10.0, gt=0.0, le=50.0)
+    POSITION_VOLATILITY_TARGET_SCORE: float = Field(default=50.0, gt=0.0, le=100.0)
+
     # Database
     DB_PATH: str = "data/trade_logs.db"
 
     # Rate Limiting (requests per second for KIS API)
-    RATE_LIMIT_RPS: float = 10.0
+    # Conservative limit to avoid EGW00201 "초당 거래건수 초과" errors.
+    # KIS API real limit is ~2 RPS; 2.0 provides maximum safety.
+    RATE_LIMIT_RPS: float = 2.0
 
     # Trading mode
     MODE: str = Field(default="paper", pattern="^(paper|live)$")
 
+    # Simulated USD cash for VTS (paper) overseas trading.
+    # KIS VTS overseas balance API returns errors for most accounts.
+    # This value is used as a fallback when the balance API returns 0 in paper mode.
+    PAPER_OVERSEAS_CASH: float = Field(default=50000.0, ge=0.0)
+    USD_BUFFER_MIN: float = Field(default=1000.0, ge=0.0)
+
+    # Trading frequency mode (daily = batch API calls, realtime = per-stock calls)
+    TRADE_MODE: str = Field(default="daily", pattern="^(daily|realtime)$")
+    DAILY_SESSIONS: int = Field(default=4, ge=1, le=10)
+    SESSION_INTERVAL_HOURS: int = Field(default=6, ge=1, le=24)
+    ORDER_BLACKOUT_ENABLED: bool = True
+    ORDER_BLACKOUT_WINDOWS_KST: str = "23:30-00:10"
+    ORDER_BLACKOUT_QUEUE_MAX: int = Field(default=500, ge=10, le=5000)
+
+    # Pre-Market Planner
+    PRE_MARKET_MINUTES: int = Field(default=30, ge=10, le=120)
+    MAX_SCENARIOS_PER_STOCK: int = Field(default=5, ge=1, le=10)
+    PLANNER_TIMEOUT_SECONDS: int = Field(default=60, ge=10, le=300)
+    DEFENSIVE_PLAYBOOK_ON_FAILURE: bool = True
+    RESCAN_INTERVAL_SECONDS: int = Field(default=300, ge=60, le=900)
+
     # Market selection (comma-separated market codes)
-    ENABLED_MARKETS: str = "KR"
+    ENABLED_MARKETS: str = "KR,US"
+
+    # Backup and Disaster Recovery (optional)
+    BACKUP_ENABLED: bool = True
+    BACKUP_DIR: str = "data/backups"
+    S3_ENDPOINT_URL: str | None = None  # For MinIO, Backblaze B2, etc.
+    S3_ACCESS_KEY: str | None = None
+    S3_SECRET_KEY: str | None = None
+    S3_BUCKET_NAME: str | None = None
+    S3_REGION: str = "us-east-1"
+
+    # Telegram Notifications (optional)
+    TELEGRAM_BOT_TOKEN: str | None = None
+    TELEGRAM_CHAT_ID: str | None = None
+    TELEGRAM_ENABLED: bool = True
+
+    # Telegram Commands (optional)
+    TELEGRAM_COMMANDS_ENABLED: bool = True
+    TELEGRAM_POLLING_INTERVAL: float = 1.0  # seconds
+
+    # Telegram notification type filters (granular control)
+    # circuit_breaker is always sent regardless — safety-critical
+    TELEGRAM_NOTIFY_TRADES: bool = True           # BUY/SELL execution alerts
+    TELEGRAM_NOTIFY_MARKET_OPEN_CLOSE: bool = True  # Market open/close alerts
+    TELEGRAM_NOTIFY_FAT_FINGER: bool = True       # Fat-finger rejection alerts
+    TELEGRAM_NOTIFY_SYSTEM_EVENTS: bool = True    # System start/shutdown alerts
+    TELEGRAM_NOTIFY_PLAYBOOK: bool = True         # Playbook generated/failed alerts
+    TELEGRAM_NOTIFY_SCENARIO_MATCH: bool = True   # Scenario matched alerts (most frequent)
+    TELEGRAM_NOTIFY_ERRORS: bool = True           # Error alerts
+
+    # Overseas ranking API (KIS endpoint/TR_ID may vary by account/product)
+    # Override these from .env if your account uses different specs.
+    OVERSEAS_RANKING_ENABLED: bool = True
+    OVERSEAS_RANKING_FLUCT_TR_ID: str = "HHDFS76290000"
+    OVERSEAS_RANKING_VOLUME_TR_ID: str = "HHDFS76270000"
+    OVERSEAS_RANKING_FLUCT_PATH: str = (
+        "/uapi/overseas-stock/v1/ranking/updown-rate"
+    )
+    OVERSEAS_RANKING_VOLUME_PATH: str = (
+        "/uapi/overseas-stock/v1/ranking/volume-surge"
+    )
+
+    # Dashboard (optional)
+    DASHBOARD_ENABLED: bool = False
+    DASHBOARD_HOST: str = "127.0.0.1"
+    DASHBOARD_PORT: int = Field(default=8080, ge=1, le=65535)
 
     model_config = {"env_file": ".env", "env_file_encoding": "utf-8"}
 
@@ -49,4 +137,7 @@ class Settings(BaseSettings):
     @property
     def enabled_market_list(self) -> list[str]:
         """Parse ENABLED_MARKETS into list of market codes."""
-        return [m.strip() for m in self.ENABLED_MARKETS.split(",") if m.strip()]
+        from src.markets.schedule import expand_market_codes
+
+        raw = [m.strip() for m in self.ENABLED_MARKETS.split(",") if m.strip()]
+        return expand_market_codes(raw)

src/context/__init__.py

@@ -0,0 +1,11 @@
+"""Multi-layered context management system for trading decisions.
+
+The context tree implements Pillar 2: hierarchical memory management across
+7 time horizons, from real-time quotes to generational wisdom.
+"""
+
+from src.context.layer import ContextLayer
+from src.context.scheduler import ContextScheduler
+from src.context.store import ContextStore
+
+__all__ = ["ContextLayer", "ContextScheduler", "ContextStore"]

src/context/aggregator.py

@@ -0,0 +1,334 @@
+"""Context aggregation logic for rolling up data from lower to higher layers."""
+
+from __future__ import annotations
+
+import sqlite3
+from datetime import UTC, datetime
+from typing import Any
+
+from src.context.layer import ContextLayer
+from src.context.store import ContextStore
+
+
+class ContextAggregator:
+    """Aggregates context data from lower (finer) to higher (coarser) layers."""
+
+    def __init__(self, conn: sqlite3.Connection) -> None:
+        """Initialize the aggregator with a database connection."""
+        self.conn = conn
+        self.store = ContextStore(conn)
+
+    def aggregate_daily_from_trades(
+        self, date: str | None = None, market: str | None = None
+    ) -> None:
+        """Aggregate L6 (daily) context from trades table.
+
+        Args:
+            date: Date in YYYY-MM-DD format. If None, uses today.
+            market: Market code filter (e.g., "KR", "US"). If None, aggregates all markets.
+        """
+        if date is None:
+            date = datetime.now(UTC).date().isoformat()
+
+        if market is None:
+            cursor = self.conn.execute(
+                """
+                SELECT DISTINCT market
+                FROM trades
+                WHERE DATE(timestamp) = ?
+                """,
+                (date,),
+            )
+            markets = [row[0] for row in cursor.fetchall() if row[0]]
+        else:
+            markets = [market]
+
+        for market_code in markets:
+            # Calculate daily metrics from trades for the market
+            cursor = self.conn.execute(
+                """
+                SELECT
+                    COUNT(*) as trade_count,
+                    SUM(CASE WHEN action = 'BUY' THEN 1 ELSE 0 END) as buys,
+                    SUM(CASE WHEN action = 'SELL' THEN 1 ELSE 0 END) as sells,
+                    SUM(CASE WHEN action = 'HOLD' THEN 1 ELSE 0 END) as holds,
+                    AVG(confidence) as avg_confidence,
+                    SUM(pnl) as total_pnl,
+                    COUNT(DISTINCT stock_code) as unique_stocks,
+                    SUM(CASE WHEN pnl > 0 THEN 1 ELSE 0 END) as wins,
+                    SUM(CASE WHEN pnl < 0 THEN 1 ELSE 0 END) as losses
+                FROM trades
+                WHERE DATE(timestamp) = ? AND market = ?
+                """,
+                (date, market_code),
+            )
+            row = cursor.fetchone()
+
+            if row and row[0] > 0:  # At least one trade
+                trade_count, buys, sells, holds, avg_conf, total_pnl, stocks, wins, losses = row
+
+                key_suffix = f"_{market_code}"
+
+                # Store daily metrics in L6 with market suffix
+                self.store.set_context(
+                    ContextLayer.L6_DAILY, date, f"trade_count{key_suffix}", trade_count
+                )
+                self.store.set_context(ContextLayer.L6_DAILY, date, f"buys{key_suffix}", buys)
+                self.store.set_context(ContextLayer.L6_DAILY, date, f"sells{key_suffix}", sells)
+                self.store.set_context(ContextLayer.L6_DAILY, date, f"holds{key_suffix}", holds)
+                self.store.set_context(
+                    ContextLayer.L6_DAILY,
+                    date,
+                    f"avg_confidence{key_suffix}",
+                    round(avg_conf, 2),
+                )
+                self.store.set_context(
+                    ContextLayer.L6_DAILY,
+                    date,
+                    f"total_pnl{key_suffix}",
+                    round(total_pnl, 2),
+                )
+                self.store.set_context(
+                    ContextLayer.L6_DAILY, date, f"unique_stocks{key_suffix}", stocks
+                )
+                win_rate = round(wins / max(wins + losses, 1) * 100, 2)
+                self.store.set_context(
+                    ContextLayer.L6_DAILY, date, f"win_rate{key_suffix}", win_rate
+                )
+
+    def aggregate_weekly_from_daily(self, week: str | None = None) -> None:
+        """Aggregate L5 (weekly) context from L6 (daily).
+
+        Args:
+            week: Week in YYYY-Www format (ISO week). If None, uses current week.
+        """
+        if week is None:
+            week = datetime.now(UTC).strftime("%Y-W%V")
+
+        # Get all daily contexts for this week
+        cursor = self.conn.execute(
+            """
+            SELECT key, value FROM contexts
+            WHERE layer = ? AND timeframe LIKE ?
+            """,
+            (ContextLayer.L6_DAILY.value, f"{week[:4]}-%"),  # All days in the year
+        )
+
+        # Group by key and collect all values
+        import json
+        from collections import defaultdict
+
+        daily_data: dict[str, list[Any]] = defaultdict(list)
+        for row in cursor.fetchall():
+            daily_data[row[0]].append(json.loads(row[1]))
+
+        if daily_data:
+            # Sum all PnL values (market-specific if suffixed)
+            if "total_pnl" in daily_data:
+                total_pnl = sum(daily_data["total_pnl"])
+                self.store.set_context(
+                    ContextLayer.L5_WEEKLY, week, "weekly_pnl", round(total_pnl, 2)
+                )
+
+            for key, values in daily_data.items():
+                if key.startswith("total_pnl_"):
+                    market_code = key.split("total_pnl_", 1)[1]
+                    total_pnl = sum(values)
+                    self.store.set_context(
+                        ContextLayer.L5_WEEKLY,
+                        week,
+                        f"weekly_pnl_{market_code}",
+                        round(total_pnl, 2),
+                    )
+
+            # Average all confidence values (market-specific if suffixed)
+            if "avg_confidence" in daily_data:
+                conf_values = daily_data["avg_confidence"]
+                avg_conf = sum(conf_values) / len(conf_values)
+                self.store.set_context(
+                    ContextLayer.L5_WEEKLY, week, "avg_confidence", round(avg_conf, 2)
+                )
+
+            for key, values in daily_data.items():
+                if key.startswith("avg_confidence_"):
+                    market_code = key.split("avg_confidence_", 1)[1]
+                    avg_conf = sum(values) / len(values)
+                    self.store.set_context(
+                        ContextLayer.L5_WEEKLY,
+                        week,
+                        f"avg_confidence_{market_code}",
+                        round(avg_conf, 2),
+                    )
+
+    def aggregate_monthly_from_weekly(self, month: str | None = None) -> None:
+        """Aggregate L4 (monthly) context from L5 (weekly).
+
+        Args:
+            month: Month in YYYY-MM format. If None, uses current month.
+        """
+        if month is None:
+            month = datetime.now(UTC).strftime("%Y-%m")
+
+        # Get all weekly contexts for this month
+        cursor = self.conn.execute(
+            """
+            SELECT key, value FROM contexts
+            WHERE layer = ? AND timeframe LIKE ?
+            """,
+            (ContextLayer.L5_WEEKLY.value, f"{month[:4]}-W%"),
+        )
+
+        # Group by key and collect all values
+        import json
+        from collections import defaultdict
+
+        weekly_data: dict[str, list[Any]] = defaultdict(list)
+        for row in cursor.fetchall():
+            weekly_data[row[0]].append(json.loads(row[1]))
+
+        if weekly_data:
+            # Sum all weekly PnL values
+            total_pnl_values: list[float] = []
+            if "weekly_pnl" in weekly_data:
+                total_pnl_values.extend(weekly_data["weekly_pnl"])
+
+            for key, values in weekly_data.items():
+                if key.startswith("weekly_pnl_"):
+                    total_pnl_values.extend(values)
+
+            if total_pnl_values:
+                total_pnl = sum(total_pnl_values)
+                self.store.set_context(
+                    ContextLayer.L4_MONTHLY, month, "monthly_pnl", round(total_pnl, 2)
+                )
+
+    def aggregate_quarterly_from_monthly(self, quarter: str | None = None) -> None:
+        """Aggregate L3 (quarterly) context from L4 (monthly).
+
+        Args:
+            quarter: Quarter in YYYY-Qn format. If None, uses current quarter.
+        """
+        if quarter is None:
+            from datetime import datetime
+
+            now = datetime.now(UTC)
+            q = (now.month - 1) // 3 + 1
+            quarter = f"{now.year}-Q{q}"
+
+        # Get all monthly contexts for this quarter
+        # Q1: 01-03, Q2: 04-06, Q3: 07-09, Q4: 10-12
+        q_num = int(quarter.split("-Q")[1])
+        months = [f"{quarter[:4]}-{m:02d}" for m in range((q_num - 1) * 3 + 1, q_num * 3 + 1)]
+
+        total_pnl = 0.0
+        for month in months:
+            monthly_pnl = self.store.get_context(
+                ContextLayer.L4_MONTHLY, month, "monthly_pnl"
+            )
+            if monthly_pnl is not None:
+                total_pnl += monthly_pnl
+
+        self.store.set_context(
+            ContextLayer.L3_QUARTERLY, quarter, "quarterly_pnl", round(total_pnl, 2)
+        )
+
+    def aggregate_annual_from_quarterly(self, year: str | None = None) -> None:
+        """Aggregate L2 (annual) context from L3 (quarterly).
+
+        Args:
+            year: Year in YYYY format. If None, uses current year.
+        """
+        if year is None:
+            year = str(datetime.now(UTC).year)
+
+        # Get all quarterly contexts for this year
+        total_pnl = 0.0
+        for q in range(1, 5):
+            quarter = f"{year}-Q{q}"
+            quarterly_pnl = self.store.get_context(
+                ContextLayer.L3_QUARTERLY, quarter, "quarterly_pnl"
+            )
+            if quarterly_pnl is not None:
+                total_pnl += quarterly_pnl
+
+        self.store.set_context(
+            ContextLayer.L2_ANNUAL, year, "annual_pnl", round(total_pnl, 2)
+        )
+
+    def aggregate_legacy_from_annual(self) -> None:
+        """Aggregate L1 (legacy) context from all L2 (annual) data."""
+        # Get all annual PnL
+        cursor = self.conn.execute(
+            """
+            SELECT timeframe, value FROM contexts
+            WHERE layer = ? AND key = ?
+            ORDER BY timeframe
+            """,
+            (ContextLayer.L2_ANNUAL.value, "annual_pnl"),
+        )
+
+        import json
+
+        annual_data = [(row[0], json.loads(row[1])) for row in cursor.fetchall()]
+
+        if annual_data:
+            total_pnl = sum(pnl for _, pnl in annual_data)
+            years_traded = len(annual_data)
+            avg_annual_pnl = total_pnl / years_traded
+
+            # Store in L1 (single "LEGACY" timeframe)
+            self.store.set_context(
+                ContextLayer.L1_LEGACY, "LEGACY", "total_pnl", round(total_pnl, 2)
+            )
+            self.store.set_context(
+                ContextLayer.L1_LEGACY, "LEGACY", "years_traded", years_traded
+            )
+            self.store.set_context(
+                ContextLayer.L1_LEGACY,
+                "LEGACY",
+                "avg_annual_pnl",
+                round(avg_annual_pnl, 2),
+            )
+
+    def run_all_aggregations(self) -> None:
+        """Run all aggregations from L7 to L1 (bottom-up).
+
+        All timeframes are derived from the latest trade timestamp so that
+        past data re-aggregation produces consistent results across layers.
+        """
+        cursor = self.conn.execute("SELECT MAX(timestamp) FROM trades")
+        row = cursor.fetchone()
+        if not row or row[0] is None:
+            return
+
+        ts_raw = row[0]
+        if ts_raw.endswith("Z"):
+            ts_raw = ts_raw.replace("Z", "+00:00")
+        latest_ts = datetime.fromisoformat(ts_raw)
+        trade_date = latest_ts.date()
+        date_str = trade_date.isoformat()
+
+        iso_year, iso_week, _ = trade_date.isocalendar()
+        week_str = f"{iso_year}-W{iso_week:02d}"
+        month_str = f"{trade_date.year}-{trade_date.month:02d}"
+        quarter = (trade_date.month - 1) // 3 + 1
+        quarter_str = f"{trade_date.year}-Q{quarter}"
+        year_str = str(trade_date.year)
+
+        # L7 (trades) → L6 (daily)
+        self.aggregate_daily_from_trades(date_str)
+
+        # L6 (daily) → L5 (weekly)
+        self.aggregate_weekly_from_daily(week_str)
+
+        # L5 (weekly) → L4 (monthly)
+        self.aggregate_monthly_from_weekly(month_str)
+
+        # L4 (monthly) → L3 (quarterly)
+        self.aggregate_quarterly_from_monthly(quarter_str)
+
+        # L3 (quarterly) → L2 (annual)
+        self.aggregate_annual_from_quarterly(year_str)
+
+        # L2 (annual) → L1 (legacy)
+        self.aggregate_legacy_from_annual()

src/context/layer.py

@@ -0,0 +1,75 @@
+"""Context layer definitions for multi-tier memory management."""
+
+from __future__ import annotations
+
+from dataclasses import dataclass
+from enum import Enum
+
+
+class ContextLayer(str, Enum):
+    """7-tier context hierarchy from real-time to generational."""
+
+    L1_LEGACY = "L1_LEGACY"  # Cumulative/generational wisdom
+    L2_ANNUAL = "L2_ANNUAL"  # Yearly performance
+    L3_QUARTERLY = "L3_QUARTERLY"  # Quarterly strategy adjustments
+    L4_MONTHLY = "L4_MONTHLY"  # Monthly rebalancing
+    L5_WEEKLY = "L5_WEEKLY"  # Weekly stock selection
+    L6_DAILY = "L6_DAILY"  # Daily trade logs
+    L7_REALTIME = "L7_REALTIME"  # Real-time market data
+
+
+@dataclass(frozen=True)
+class LayerMetadata:
+    """Metadata for each context layer."""
+
+    layer: ContextLayer
+    description: str
+    retention_days: int | None  # None = keep forever
+    aggregation_source: ContextLayer | None  # Parent layer for aggregation
+
+
+# Layer configuration
+LAYER_CONFIG: dict[ContextLayer, LayerMetadata] = {
+    ContextLayer.L1_LEGACY: LayerMetadata(
+        layer=ContextLayer.L1_LEGACY,
+        description="Cumulative trading history and core lessons learned across generations",
+        retention_days=None,  # Keep forever
+        aggregation_source=ContextLayer.L2_ANNUAL,
+    ),
+    ContextLayer.L2_ANNUAL: LayerMetadata(
+        layer=ContextLayer.L2_ANNUAL,
+        description="Yearly returns, Sharpe ratio, max drawdown, win rate",
+        retention_days=365 * 10,  # 10 years
+        aggregation_source=ContextLayer.L3_QUARTERLY,
+    ),
+    ContextLayer.L3_QUARTERLY: LayerMetadata(
+        layer=ContextLayer.L3_QUARTERLY,
+        description="Quarterly strategy adjustments, market phase detection, sector rotation",
+        retention_days=365 * 3,  # 3 years
+        aggregation_source=ContextLayer.L4_MONTHLY,
+    ),
+    ContextLayer.L4_MONTHLY: LayerMetadata(
+        layer=ContextLayer.L4_MONTHLY,
+        description="Monthly portfolio rebalancing, risk exposure, drawdown recovery",
+        retention_days=365 * 2,  # 2 years
+        aggregation_source=ContextLayer.L5_WEEKLY,
+    ),
+    ContextLayer.L5_WEEKLY: LayerMetadata(
+        layer=ContextLayer.L5_WEEKLY,
+        description="Weekly stock selection, sector focus, volatility regime",
+        retention_days=365,  # 1 year
+        aggregation_source=ContextLayer.L6_DAILY,
+    ),
+    ContextLayer.L6_DAILY: LayerMetadata(
+        layer=ContextLayer.L6_DAILY,
+        description="Daily trade logs, P&L, market summaries, decision accuracy",
+        retention_days=90,  # 90 days
+        aggregation_source=ContextLayer.L7_REALTIME,
+    ),
+    ContextLayer.L7_REALTIME: LayerMetadata(
+        layer=ContextLayer.L7_REALTIME,
+        description="Real-time positions, quotes, orderbook, volatility, live P&L",
+        retention_days=7,  # 7 days (real-time data is ephemeral)
+        aggregation_source=None,  # No aggregation source (leaf layer)
+    ),
+}

src/context/scheduler.py

@@ -0,0 +1,135 @@
+"""Context aggregation scheduler for periodic rollups and cleanup."""
+
+from __future__ import annotations
+
+import sqlite3
+from calendar import monthrange
+from dataclasses import dataclass
+from datetime import UTC, datetime
+
+from src.context.aggregator import ContextAggregator
+from src.context.store import ContextStore
+
+
+@dataclass(frozen=True)
+class ScheduleResult:
+    """Represents which scheduled tasks ran."""
+
+    weekly: bool = False
+    monthly: bool = False
+    quarterly: bool = False
+    annual: bool = False
+    legacy: bool = False
+    cleanup: bool = False
+
+
+class ContextScheduler:
+    """Run periodic context aggregations and cleanup when due."""
+
+    def __init__(
+        self,
+        conn: sqlite3.Connection | None = None,
+        aggregator: ContextAggregator | None = None,
+        store: ContextStore | None = None,
+    ) -> None:
+        if aggregator is None:
+            if conn is None:
+                raise ValueError("conn is required when aggregator is not provided")
+            aggregator = ContextAggregator(conn)
+        self.aggregator = aggregator
+
+        if store is None:
+            store = getattr(aggregator, "store", None)
+        if store is None:
+            if conn is None:
+                raise ValueError("conn is required when store is not provided")
+            store = ContextStore(conn)
+        self.store = store
+
+        self._last_run: dict[str, str] = {}
+
+    def run_if_due(self, now: datetime | None = None) -> ScheduleResult:
+        """Run scheduled aggregations if their schedule is due.
+
+        Args:
+            now: Current datetime (UTC). If None, uses current time.
+
+        Returns:
+            ScheduleResult indicating which tasks ran.
+        """
+        if now is None:
+            now = datetime.now(UTC)
+
+        today = now.date().isoformat()
+        result = ScheduleResult()
+
+        if self._should_run("cleanup", today):
+            self.store.cleanup_expired_contexts()
+            result = self._with(result, cleanup=True)
+
+        if self._is_sunday(now) and self._should_run("weekly", today):
+            week = now.strftime("%Y-W%V")
+            self.aggregator.aggregate_weekly_from_daily(week)
+            result = self._with(result, weekly=True)
+
+        if self._is_last_day_of_month(now) and self._should_run("monthly", today):
+            month = now.strftime("%Y-%m")
+            self.aggregator.aggregate_monthly_from_weekly(month)
+            result = self._with(result, monthly=True)
+
+        if self._is_last_day_of_quarter(now) and self._should_run("quarterly", today):
+            quarter = self._current_quarter(now)
+            self.aggregator.aggregate_quarterly_from_monthly(quarter)
+            result = self._with(result, quarterly=True)
+
+        if self._is_last_day_of_year(now) and self._should_run("annual", today):
+            year = str(now.year)
+            self.aggregator.aggregate_annual_from_quarterly(year)
+            result = self._with(result, annual=True)
+
+            # Legacy rollup runs after annual aggregation.
+            self.aggregator.aggregate_legacy_from_annual()
+            result = self._with(result, legacy=True)
+
+        return result
+
+    def _should_run(self, key: str, date_str: str) -> bool:
+        if self._last_run.get(key) == date_str:
+            return False
+        self._last_run[key] = date_str
+        return True
+
+    @staticmethod
+    def _is_sunday(now: datetime) -> bool:
+        return now.weekday() == 6
+
+    @staticmethod
+    def _is_last_day_of_month(now: datetime) -> bool:
+        last_day = monthrange(now.year, now.month)[1]
+        return now.day == last_day
+
+    @classmethod
+    def _is_last_day_of_quarter(cls, now: datetime) -> bool:
+        if now.month not in (3, 6, 9, 12):
+            return False
+        return cls._is_last_day_of_month(now)
+
+    @staticmethod
+    def _is_last_day_of_year(now: datetime) -> bool:
+        return now.month == 12 and now.day == 31
+
+    @staticmethod
+    def _current_quarter(now: datetime) -> str:
+        quarter = (now.month - 1) // 3 + 1
+        return f"{now.year}-Q{quarter}"
+
+    @staticmethod
+    def _with(result: ScheduleResult, **kwargs: bool) -> ScheduleResult:
+        return ScheduleResult(
+            weekly=kwargs.get("weekly", result.weekly),
+            monthly=kwargs.get("monthly", result.monthly),
+            quarterly=kwargs.get("quarterly", result.quarterly),
+            annual=kwargs.get("annual", result.annual),
+            legacy=kwargs.get("legacy", result.legacy),
+            cleanup=kwargs.get("cleanup", result.cleanup),
+        )

src/context/store.py

@@ -0,0 +1,193 @@
+"""Context storage and retrieval for the 7-tier memory system."""
+
+from __future__ import annotations
+
+import json
+import sqlite3
+from datetime import UTC, datetime
+from typing import Any
+
+from src.context.layer import LAYER_CONFIG, ContextLayer
+
+
+class ContextStore:
+    """Manages context data across the 7-tier hierarchy."""
+
+    def __init__(self, conn: sqlite3.Connection) -> None:
+        """Initialize the context store with a database connection."""
+        self.conn = conn
+        self._init_metadata()
+
+    def _init_metadata(self) -> None:
+        """Initialize context_metadata table with layer configurations."""
+        for config in LAYER_CONFIG.values():
+            self.conn.execute(
+                """
+                INSERT OR REPLACE INTO context_metadata
+                (layer, description, retention_days, aggregation_source)
+                VALUES (?, ?, ?, ?)
+                """,
+                (
+                    config.layer.value,
+                    config.description,
+                    config.retention_days,
+                    config.aggregation_source.value if config.aggregation_source else None,
+                ),
+            )
+        self.conn.commit()
+
+    def set_context(
+        self,
+        layer: ContextLayer,
+        timeframe: str,
+        key: str,
+        value: Any,
+    ) -> None:
+        """Set a context value for a given layer and timeframe.
+
+        Args:
+            layer: The context layer (L1-L7)
+            timeframe: Time identifier (e.g., "2026", "2026-Q1", "2026-01",
+                "2026-W05", "2026-02-04")
+            key: Context key (e.g., "sharpe_ratio", "win_rate", "lesson_learned")
+            value: Context value (will be JSON-serialized)
+        """
+        now = datetime.now(UTC).isoformat()
+        value_json = json.dumps(value)
+
+        self.conn.execute(
+            """
+            INSERT INTO contexts (layer, timeframe, key, value, created_at, updated_at)
+            VALUES (?, ?, ?, ?, ?, ?)
+            ON CONFLICT(layer, timeframe, key)
+            DO UPDATE SET value = excluded.value, updated_at = excluded.updated_at
+            """,
+            (layer.value, timeframe, key, value_json, now, now),
+        )
+        self.conn.commit()
+
+    def get_context(
+        self,
+        layer: ContextLayer,
+        timeframe: str,
+        key: str,
+    ) -> Any | None:
+        """Get a context value for a given layer and timeframe.
+
+        Args:
+            layer: The context layer (L1-L7)
+            timeframe: Time identifier
+            key: Context key
+
+        Returns:
+            The context value (deserialized from JSON), or None if not found
+        """
+        cursor = self.conn.execute(
+            """
+            SELECT value FROM contexts
+            WHERE layer = ? AND timeframe = ? AND key = ?
+            """,
+            (layer.value, timeframe, key),
+        )
+        row = cursor.fetchone()
+        if row:
+            return json.loads(row[0])
+        return None
+
+    def get_all_contexts(
+        self,
+        layer: ContextLayer,
+        timeframe: str | None = None,
+    ) -> dict[str, Any]:
+        """Get all context values for a given layer and optional timeframe.
+
+        Args:
+            layer: The context layer (L1-L7)
+            timeframe: Optional time identifier filter
+
+        Returns:
+            Dictionary of key-value pairs for the specified layer/timeframe
+        """
+        if timeframe:
+            cursor = self.conn.execute(
+                """
+                SELECT key, value FROM contexts
+                WHERE layer = ? AND timeframe = ?
+                ORDER BY key
+                """,
+                (layer.value, timeframe),
+            )
+        else:
+            cursor = self.conn.execute(
+                """
+                SELECT key, value FROM contexts
+                WHERE layer = ?
+                ORDER BY timeframe DESC, key
+                """,
+                (layer.value,),
+            )
+
+        return {row[0]: json.loads(row[1]) for row in cursor.fetchall()}
+
+    def get_latest_timeframe(self, layer: ContextLayer) -> str | None:
+        """Get the most recent timeframe for a given layer.
+
+        Args:
+            layer: The context layer (L1-L7)
+
+        Returns:
+            The latest timeframe string, or None if no data exists
+        """
+        cursor = self.conn.execute(
+            """
+            SELECT timeframe FROM contexts
+            WHERE layer = ?
+            ORDER BY updated_at DESC
+            LIMIT 1
+            """,
+            (layer.value,),
+        )
+        row = cursor.fetchone()
+        return row[0] if row else None
+
+    def delete_old_contexts(self, layer: ContextLayer, cutoff_date: str) -> int:
+        """Delete contexts older than the cutoff date for a given layer.
+
+        Args:
+            layer: The context layer (L1-L7)
+            cutoff_date: ISO format date string (contexts before this will be deleted)
+
+        Returns:
+            Number of rows deleted
+        """
+        cursor = self.conn.execute(
+            """
+            DELETE FROM contexts
+            WHERE layer = ? AND updated_at < ?
+            """,
+            (layer.value, cutoff_date),
+        )
+        self.conn.commit()
+        return cursor.rowcount
+
+    def cleanup_expired_contexts(self) -> dict[ContextLayer, int]:
+        """Delete expired contexts based on retention policies.
+
+        Returns:
+            Dictionary mapping layer to number of deleted rows
+        """
+        deleted_counts: dict[ContextLayer, int] = {}
+
+        for layer, config in LAYER_CONFIG.items():
+            if config.retention_days is None:
+                # Keep forever (e.g., L1_LEGACY)
+                deleted_counts[layer] = 0
+                continue
+
+            # Calculate cutoff date
+            from datetime import timedelta
+
+            cutoff = datetime.now(UTC) - timedelta(days=config.retention_days)
+            deleted_counts[layer] = self.delete_old_contexts(layer, cutoff.isoformat())
+
+        return deleted_counts

src/context/summarizer.py

@@ -0,0 +1,328 @@
+"""Context summarization for efficient historical data representation.
+
+This module summarizes old context data instead of including raw details:
+- Key metrics only (averages, trends, not details)
+- Rolling window (keep last N days detailed, summarize older)
+- Aggregate historical data efficiently
+"""
+
+from __future__ import annotations
+
+from dataclasses import dataclass
+from datetime import UTC, datetime, timedelta
+from typing import Any
+
+from src.context.layer import ContextLayer
+from src.context.store import ContextStore
+
+
+@dataclass(frozen=True)
+class SummaryStats:
+    """Statistical summary of historical data."""
+
+    count: int
+    mean: float | None = None
+    min: float | None = None
+    max: float | None = None
+    std: float | None = None
+    trend: str | None = None  # "up", "down", "flat"
+
+
+class ContextSummarizer:
+    """Summarizes historical context data to reduce token usage."""
+
+    def __init__(self, store: ContextStore) -> None:
+        """Initialize the context summarizer.
+
+        Args:
+            store: ContextStore instance for retrieving context data
+        """
+        self.store = store
+
+    def summarize_numeric_values(self, values: list[float]) -> SummaryStats:
+        """Summarize a list of numeric values.
+
+        Args:
+            values: List of numeric values to summarize
+
+        Returns:
+            SummaryStats with mean, min, max, std, and trend
+        """
+        if not values:
+            return SummaryStats(count=0)
+
+        count = len(values)
+        mean = sum(values) / count
+        min_val = min(values)
+        max_val = max(values)
+
+        # Calculate standard deviation
+        if count > 1:
+            variance = sum((x - mean) ** 2 for x in values) / (count - 1)
+            std = variance**0.5
+        else:
+            std = 0.0
+
+        # Determine trend
+        trend = "flat"
+        if count >= 3:
+            # Simple trend: compare first third vs last third
+            first_third = values[: count // 3]
+            last_third = values[-(count // 3) :]
+            first_avg = sum(first_third) / len(first_third)
+            last_avg = sum(last_third) / len(last_third)
+
+            # Trend threshold: 5% change
+            threshold = 0.05 * abs(first_avg) if first_avg != 0 else 0.01
+
+            if last_avg > first_avg + threshold:
+                trend = "up"
+            elif last_avg < first_avg - threshold:
+                trend = "down"
+
+        return SummaryStats(
+            count=count,
+            mean=round(mean, 4),
+            min=round(min_val, 4),
+            max=round(max_val, 4),
+            std=round(std, 4),
+            trend=trend,
+        )
+
+    def summarize_layer(
+        self,
+        layer: ContextLayer,
+        start_date: datetime | None = None,
+        end_date: datetime | None = None,
+    ) -> dict[str, Any]:
+        """Summarize all context data for a layer within a date range.
+
+        Args:
+            layer: Context layer to summarize
+            start_date: Start date (inclusive), None for all
+            end_date: End date (inclusive), None for now
+
+        Returns:
+            Dictionary with summarized metrics
+        """
+        if end_date is None:
+            end_date = datetime.now(UTC)
+
+        # Get all contexts for this layer
+        all_contexts = self.store.get_all_contexts(layer)
+
+        if not all_contexts:
+            return {"summary": "No data available", "count": 0}
+
+        # Group numeric values by key
+        numeric_data: dict[str, list[float]] = {}
+        text_data: dict[str, list[str]] = {}
+
+        for key, value in all_contexts.items():
+            # Try to extract numeric values
+            if isinstance(value, (int, float)):
+                if key not in numeric_data:
+                    numeric_data[key] = []
+                numeric_data[key].append(float(value))
+            elif isinstance(value, dict):
+                # Extract numeric fields from dict
+                for subkey, subvalue in value.items():
+                    if isinstance(subvalue, (int, float)):
+                        full_key = f"{key}.{subkey}"
+                        if full_key not in numeric_data:
+                            numeric_data[full_key] = []
+                        numeric_data[full_key].append(float(subvalue))
+            elif isinstance(value, str):
+                if key not in text_data:
+                    text_data[key] = []
+                text_data[key].append(value)
+
+        # Summarize numeric data
+        summary: dict[str, Any] = {}
+
+        for key, values in numeric_data.items():
+            stats = self.summarize_numeric_values(values)
+            summary[key] = {
+                "count": stats.count,
+                "avg": stats.mean,
+                "range": [stats.min, stats.max],
+                "trend": stats.trend,
+            }
+
+        # Summarize text data (just counts)
+        for key, values in text_data.items():
+            summary[f"{key}_count"] = len(values)
+
+        summary["total_entries"] = len(all_contexts)
+
+        return summary
+
+    def rolling_window_summary(
+        self,
+        layer: ContextLayer,
+        window_days: int = 30,
+        summarize_older: bool = True,
+    ) -> dict[str, Any]:
+        """Create a rolling window summary.
+
+        Recent data (within window) is kept detailed.
+        Older data is summarized to key metrics.
+
+        Args:
+            layer: Context layer to summarize
+            window_days: Number of days to keep detailed
+            summarize_older: Whether to summarize data older than window
+
+        Returns:
+            Dictionary with recent (detailed) and historical (summary) data
+        """
+        result: dict[str, Any] = {
+            "window_days": window_days,
+            "recent_data": {},
+            "historical_summary": {},
+        }
+
+        # Get all contexts
+        all_contexts = self.store.get_all_contexts(layer)
+
+        recent_values: dict[str, list[float]] = {}
+        historical_values: dict[str, list[float]] = {}
+
+        for key, value in all_contexts.items():
+            # For simplicity, treat all numeric values
+            if isinstance(value, (int, float)):
+                # Note: We don't have timestamps in context keys
+                # This is a simplified implementation
+                # In practice, would need to check timeframe field
+
+                # For now, put recent data in window
+                if key not in recent_values:
+                    recent_values[key] = []
+                recent_values[key].append(float(value))
+
+        # Detailed recent data
+        result["recent_data"] = {key: values[-10:] for key, values in recent_values.items()}
+
+        # Summarized historical data
+        if summarize_older:
+            for key, values in historical_values.items():
+                stats = self.summarize_numeric_values(values)
+                result["historical_summary"][key] = {
+                    "count": stats.count,
+                    "avg": stats.mean,
+                    "trend": stats.trend,
+                }
+
+        return result
+
+    def aggregate_to_higher_layer(
+        self,
+        source_layer: ContextLayer,
+        target_layer: ContextLayer,
+        metric_key: str,
+        aggregation_func: str = "mean",
+    ) -> float | None:
+        """Aggregate data from source layer to target layer.
+
+        Args:
+            source_layer: Source context layer (more granular)
+            target_layer: Target context layer (less granular)
+            metric_key: Key of metric to aggregate
+            aggregation_func: Aggregation function ("mean", "sum", "max", "min")
+
+        Returns:
+            Aggregated value, or None if no data available
+        """
+        # Get all contexts from source layer
+        source_contexts = self.store.get_all_contexts(source_layer)
+
+        # Extract values for metric_key
+        values = []
+        for key, value in source_contexts.items():
+            if key == metric_key and isinstance(value, (int, float)):
+                values.append(float(value))
+            elif isinstance(value, dict) and metric_key in value:
+                subvalue = value[metric_key]
+                if isinstance(subvalue, (int, float)):
+                    values.append(float(subvalue))
+
+        if not values:
+            return None
+
+        # Apply aggregation function
+        if aggregation_func == "mean":
+            return sum(values) / len(values)
+        elif aggregation_func == "sum":
+            return sum(values)
+        elif aggregation_func == "max":
+            return max(values)
+        elif aggregation_func == "min":
+            return min(values)
+        else:
+            return sum(values) / len(values)  # Default to mean
+
+    def create_compact_summary(
+        self,
+        layers: list[ContextLayer],
+        top_n_metrics: int = 5,
+    ) -> dict[str, Any]:
+        """Create a compact summary across multiple layers.
+
+        Args:
+            layers: List of context layers to summarize
+            top_n_metrics: Number of top metrics to include per layer
+
+        Returns:
+            Compact summary dictionary
+        """
+        summary: dict[str, Any] = {}
+
+        for layer in layers:
+            layer_summary = self.summarize_layer(layer)
+
+            # Keep only top N metrics (by count/relevance)
+            metrics = []
+            for key, value in layer_summary.items():
+                if isinstance(value, dict) and "count" in value:
+                    metrics.append((key, value, value["count"]))
+
+            # Sort by count (descending)
+            metrics.sort(key=lambda x: x[2], reverse=True)
+
+            # Keep top N
+            top_metrics = {m[0]: m[1] for m in metrics[:top_n_metrics]}
+
+            summary[layer.value] = top_metrics
+
+        return summary
+
+    def format_summary_for_prompt(self, summary: dict[str, Any]) -> str:
+        """Format summary for inclusion in a prompt.
+
+        Args:
+            summary: Summary dictionary
+
+        Returns:
+            Formatted string for prompt
+        """
+        lines = []
+
+        for layer, metrics in summary.items():
+            if not metrics:
+                continue
+
+            lines.append(f"{layer}:")
+            for key, value in metrics.items():
+                if isinstance(value, dict):
+                    # Format as: key: avg=X, trend=Y
+                    parts = []
+                    if "avg" in value and value["avg"] is not None:
+                        parts.append(f"avg={value['avg']:.2f}")
+                    if "trend" in value and value["trend"]:
+                        parts.append(f"trend={value['trend']}")
+                    if parts:
+                        lines.append(f"  {key}: {', '.join(parts)}")
+                else:
+                    lines.append(f"  {key}: {value}")
+
+        return "\n".join(lines)

src/core/blackout_manager.py

@@ -0,0 +1,105 @@
+"""Blackout policy and queued order-intent manager."""
+
+from __future__ import annotations
+
+from collections import deque
+from dataclasses import dataclass
+from datetime import UTC, datetime, time
+from zoneinfo import ZoneInfo
+
+
+@dataclass(frozen=True)
+class BlackoutWindow:
+    start: time
+    end: time
+
+    def contains(self, kst_time: time) -> bool:
+        if self.start <= self.end:
+            return self.start <= kst_time < self.end
+        return kst_time >= self.start or kst_time < self.end
+
+
+@dataclass
+class QueuedOrderIntent:
+    market_code: str
+    exchange_code: str
+    stock_code: str
+    order_type: str
+    quantity: int
+    price: float
+    source: str
+    queued_at: datetime
+    attempts: int = 0
+
+
+def parse_blackout_windows_kst(raw: str) -> list[BlackoutWindow]:
+    """Parse comma-separated KST windows like '23:30-00:10,11:20-11:30'."""
+    windows: list[BlackoutWindow] = []
+    for token in raw.split(","):
+        span = token.strip()
+        if not span or "-" not in span:
+            continue
+        start_raw, end_raw = [part.strip() for part in span.split("-", 1)]
+        try:
+            start_h, start_m = [int(v) for v in start_raw.split(":", 1)]
+            end_h, end_m = [int(v) for v in end_raw.split(":", 1)]
+        except (ValueError, TypeError):
+            continue
+        if not (0 <= start_h <= 23 and 0 <= end_h <= 23):
+            continue
+        if not (0 <= start_m <= 59 and 0 <= end_m <= 59):
+            continue
+        windows.append(BlackoutWindow(start=time(start_h, start_m), end=time(end_h, end_m)))
+    return windows
+
+
+class BlackoutOrderManager:
+    """Tracks blackout mode and queues order intents until recovery."""
+
+    def __init__(
+        self,
+        *,
+        enabled: bool,
+        windows: list[BlackoutWindow],
+        max_queue_size: int = 500,
+    ) -> None:
+        self.enabled = enabled
+        self._windows = windows
+        self._queue: deque[QueuedOrderIntent] = deque()
+        self._was_blackout = False
+        self._max_queue_size = max_queue_size
+
+    @property
+    def pending_count(self) -> int:
+        return len(self._queue)
+
+    def in_blackout(self, now: datetime | None = None) -> bool:
+        if not self.enabled or not self._windows:
+            return False
+        now = now or datetime.now(UTC)
+        kst_now = now.astimezone(ZoneInfo("Asia/Seoul")).timetz().replace(tzinfo=None)
+        return any(window.contains(kst_now) for window in self._windows)
+
+    def enqueue(self, intent: QueuedOrderIntent) -> bool:
+        if len(self._queue) >= self._max_queue_size:
+            return False
+        self._queue.append(intent)
+        return True
+
+    def pop_recovery_batch(self, now: datetime | None = None) -> list[QueuedOrderIntent]:
+        in_blackout_now = self.in_blackout(now)
+        batch: list[QueuedOrderIntent] = []
+        if not in_blackout_now and self._queue:
+            while self._queue:
+                batch.append(self._queue.popleft())
+        self._was_blackout = in_blackout_now
+        return batch
+
+    def requeue(self, intent: QueuedOrderIntent) -> None:
+        if len(self._queue) < self._max_queue_size:
+            self._queue.append(intent)
+
+    def clear(self) -> int:
+        count = len(self._queue)
+        self._queue.clear()
+        return count

src/core/criticality.py

@@ -0,0 +1,110 @@
+"""Criticality assessment for urgency-based response system.
+
+Evaluates market conditions to determine response urgency and enable
+faster reactions in critical situations.
+"""
+
+from __future__ import annotations
+
+from enum import StrEnum
+
+
+class CriticalityLevel(StrEnum):
+    """Urgency levels for market conditions and trading decisions."""
+
+    CRITICAL = "CRITICAL"  # <5s timeout - Emergency response required
+    HIGH = "HIGH"  # <30s timeout - Elevated priority
+    NORMAL = "NORMAL"  # <60s timeout - Standard processing
+    LOW = "LOW"  # No timeout - Batch processing
+
+
+class CriticalityAssessor:
+    """Assesses market conditions to determine response criticality level."""
+
+    def __init__(
+        self,
+        critical_pnl_threshold: float = -2.5,
+        critical_price_change_threshold: float = 5.0,
+        critical_volume_surge_threshold: float = 10.0,
+        high_volatility_threshold: float = 70.0,
+        low_volatility_threshold: float = 30.0,
+    ) -> None:
+        """Initialize the criticality assessor.
+
+        Args:
+            critical_pnl_threshold: P&L % that triggers CRITICAL (default -2.5%)
+            critical_price_change_threshold: Price change % that triggers CRITICAL
+                (default 5.0% in 1 minute)
+            critical_volume_surge_threshold: Volume surge ratio that triggers CRITICAL
+                (default 10x average)
+            high_volatility_threshold: Volatility score that triggers HIGH
+                (default 70.0)
+            low_volatility_threshold: Volatility score below which is LOW
+                (default 30.0)
+        """
+        self.critical_pnl_threshold = critical_pnl_threshold
+        self.critical_price_change_threshold = critical_price_change_threshold
+        self.critical_volume_surge_threshold = critical_volume_surge_threshold
+        self.high_volatility_threshold = high_volatility_threshold
+        self.low_volatility_threshold = low_volatility_threshold
+
+    def assess_market_conditions(
+        self,
+        pnl_pct: float,
+        volatility_score: float,
+        volume_surge: float,
+        price_change_1m: float = 0.0,
+        is_market_open: bool = True,
+    ) -> CriticalityLevel:
+        """Assess criticality level based on market conditions.
+
+        Args:
+            pnl_pct: Current P&L percentage
+            volatility_score: Momentum score from VolatilityAnalyzer (0-100)
+            volume_surge: Volume surge ratio (current / average)
+            price_change_1m: 1-minute price change percentage
+            is_market_open: Whether the market is currently open
+
+        Returns:
+            CriticalityLevel indicating required response urgency
+        """
+        # Market closed or very quiet → LOW priority (batch processing)
+        if not is_market_open or volatility_score < self.low_volatility_threshold:
+            return CriticalityLevel.LOW
+
+        # CRITICAL conditions: immediate action required
+        # 1. P&L near circuit breaker (-2.5% is close to -3.0% breaker)
+        if pnl_pct <= self.critical_pnl_threshold:
+            return CriticalityLevel.CRITICAL
+
+        # 2. Large sudden price movement (>5% in 1 minute)
+        if abs(price_change_1m) >= self.critical_price_change_threshold:
+            return CriticalityLevel.CRITICAL
+
+        # 3. Extreme volume surge (>10x average) indicates major event
+        if volume_surge >= self.critical_volume_surge_threshold:
+            return CriticalityLevel.CRITICAL
+
+        # HIGH priority: elevated volatility requires faster response
+        if volatility_score >= self.high_volatility_threshold:
+            return CriticalityLevel.HIGH
+
+        # NORMAL: standard trading conditions
+        return CriticalityLevel.NORMAL
+
+    def get_timeout(self, level: CriticalityLevel) -> float | None:
+        """Get timeout in seconds for a given criticality level.
+
+        Args:
+            level: Criticality level
+
+        Returns:
+            Timeout in seconds, or None for no timeout (LOW priority)
+        """
+        timeout_map = {
+            CriticalityLevel.CRITICAL: 5.0,
+            CriticalityLevel.HIGH: 30.0,
+            CriticalityLevel.NORMAL: 60.0,
+            CriticalityLevel.LOW: None,
+        }
+        return timeout_map[level]

src/core/kill_switch.py

@@ -0,0 +1,71 @@
+"""Kill switch orchestration for emergency risk actions.
+
+Order is fixed:
+1) block new orders
+2) cancel pending orders
+3) refresh order state
+4) reduce risk
+5) snapshot and notify
+"""
+
+from __future__ import annotations
+
+import inspect
+from dataclasses import dataclass, field
+from typing import Any, Awaitable, Callable
+
+StepCallable = Callable[[], Any | Awaitable[Any]]
+
+
+@dataclass
+class KillSwitchReport:
+    reason: str
+    steps: list[str] = field(default_factory=list)
+    errors: list[str] = field(default_factory=list)
+
+
+class KillSwitchOrchestrator:
+    def __init__(self) -> None:
+        self.new_orders_blocked = False
+
+    async def _run_step(
+        self,
+        report: KillSwitchReport,
+        name: str,
+        fn: StepCallable | None,
+    ) -> None:
+        report.steps.append(name)
+        if fn is None:
+            return
+        try:
+            result = fn()
+            if inspect.isawaitable(result):
+                await result
+        except Exception as exc:  # pragma: no cover - intentionally resilient
+            report.errors.append(f"{name}: {exc}")
+
+    async def trigger(
+        self,
+        *,
+        reason: str,
+        cancel_pending_orders: StepCallable | None = None,
+        refresh_order_state: StepCallable | None = None,
+        reduce_risk: StepCallable | None = None,
+        snapshot_state: StepCallable | None = None,
+        notify: StepCallable | None = None,
+    ) -> KillSwitchReport:
+        report = KillSwitchReport(reason=reason)
+
+        self.new_orders_blocked = True
+        report.steps.append("block_new_orders")
+
+        await self._run_step(report, "cancel_pending_orders", cancel_pending_orders)
+        await self._run_step(report, "refresh_order_state", refresh_order_state)
+        await self._run_step(report, "reduce_risk", reduce_risk)
+        await self._run_step(report, "snapshot_state", snapshot_state)
+        await self._run_step(report, "notify", notify)
+
+        return report
+
+    def clear_block(self) -> None:
+        self.new_orders_blocked = False

src/core/order_policy.py

@@ -0,0 +1,93 @@
+"""Session-aware order policy guards.
+
+Default policy:
+- Low-liquidity sessions must reject market orders (price <= 0).
+"""
+
+from __future__ import annotations
+
+from dataclasses import dataclass
+from datetime import UTC, datetime, time
+from zoneinfo import ZoneInfo
+
+from src.markets.schedule import MarketInfo
+
+_LOW_LIQUIDITY_SESSIONS = {"NXT_AFTER", "US_PRE", "US_DAY", "US_AFTER"}
+
+
+class OrderPolicyRejected(Exception):
+    """Raised when an order violates session policy."""
+
+    def __init__(self, message: str, *, session_id: str, market_code: str) -> None:
+        super().__init__(message)
+        self.session_id = session_id
+        self.market_code = market_code
+
+
+@dataclass(frozen=True)
+class SessionInfo:
+    session_id: str
+    is_low_liquidity: bool
+
+
+def classify_session_id(market: MarketInfo, now: datetime | None = None) -> str:
+    """Classify current session by KST schedule used in v3 docs."""
+    now = now or datetime.now(UTC)
+    # v3 session tables are explicitly defined in KST perspective.
+    kst_time = now.astimezone(ZoneInfo("Asia/Seoul")).timetz().replace(tzinfo=None)
+
+    if market.code == "KR":
+        if time(8, 0) <= kst_time < time(8, 50):
+            return "NXT_PRE"
+        if time(9, 0) <= kst_time < time(15, 30):
+            return "KRX_REG"
+        if time(15, 30) <= kst_time < time(20, 0):
+            return "NXT_AFTER"
+        return "KR_OFF"
+
+    if market.code.startswith("US"):
+        if time(10, 0) <= kst_time < time(18, 0):
+            return "US_DAY"
+        if time(18, 0) <= kst_time < time(23, 30):
+            return "US_PRE"
+        if time(23, 30) <= kst_time or kst_time < time(6, 0):
+            return "US_REG"
+        if time(6, 0) <= kst_time < time(7, 0):
+            return "US_AFTER"
+        return "US_OFF"
+
+    return "GENERIC_REG"
+
+
+def get_session_info(market: MarketInfo, now: datetime | None = None) -> SessionInfo:
+    session_id = classify_session_id(market, now)
+    return SessionInfo(session_id=session_id, is_low_liquidity=session_id in _LOW_LIQUIDITY_SESSIONS)
+
+
+def validate_order_policy(
+    *,
+    market: MarketInfo,
+    order_type: str,
+    price: float,
+    now: datetime | None = None,
+) -> SessionInfo:
+    """Validate order against session policy and return resolved session info."""
+    info = get_session_info(market, now)
+
+    is_market_order = price <= 0
+    if info.is_low_liquidity and is_market_order:
+        raise OrderPolicyRejected(
+            f"Market order is forbidden in low-liquidity session ({info.session_id})",
+            session_id=info.session_id,
+            market_code=market.code,
+        )
+
+    # Guard against accidental unsupported actions.
+    if order_type not in {"BUY", "SELL"}:
+        raise OrderPolicyRejected(
+            f"Unsupported order_type={order_type}",
+            session_id=info.session_id,
+            market_code=market.code,
+        )
+
+    return info

src/core/priority_queue.py

@@ -0,0 +1,291 @@
+"""Priority-based task queue for latency control.
+
+Implements a thread-safe priority queue with timeout enforcement and metrics tracking.
+"""
+
+from __future__ import annotations
+
+import asyncio
+import heapq
+import logging
+import time
+from collections.abc import Callable, Coroutine
+from dataclasses import dataclass, field
+from typing import Any
+
+from src.core.criticality import CriticalityLevel
+
+logger = logging.getLogger(__name__)
+
+
+@dataclass(order=True)
+class PriorityTask:
+    """Task with priority and timestamp for queue ordering."""
+
+    # Lower priority value = higher urgency (CRITICAL=0, HIGH=1, NORMAL=2, LOW=3)
+    priority: int
+    timestamp: float
+    # Task data not used in comparison
+    task_id: str = field(compare=False)
+    task_data: dict[str, Any] = field(compare=False, default_factory=dict)
+    callback: Callable[[], Coroutine[Any, Any, Any]] | None = field(
+        compare=False, default=None
+    )
+
+
+@dataclass
+class QueueMetrics:
+    """Metrics for priority queue performance monitoring."""
+
+    total_enqueued: int = 0
+    total_dequeued: int = 0
+    total_timeouts: int = 0
+    total_errors: int = 0
+    current_size: int = 0
+    # Average wait time per criticality level (in seconds)
+    avg_wait_time: dict[CriticalityLevel, float] = field(default_factory=dict)
+    # P95 wait time per criticality level
+    p95_wait_time: dict[CriticalityLevel, float] = field(default_factory=dict)
+
+
+class PriorityTaskQueue:
+    """Thread-safe priority queue with timeout enforcement."""
+
+    # Priority mapping for criticality levels
+    PRIORITY_MAP = {
+        CriticalityLevel.CRITICAL: 0,
+        CriticalityLevel.HIGH: 1,
+        CriticalityLevel.NORMAL: 2,
+        CriticalityLevel.LOW: 3,
+    }
+
+    def __init__(self, max_size: int = 1000) -> None:
+        """Initialize the priority task queue.
+
+        Args:
+            max_size: Maximum queue size (default 1000)
+        """
+        self._queue: list[PriorityTask] = []
+        self._lock = asyncio.Lock()
+        self._max_size = max_size
+        self._metrics = QueueMetrics()
+        # Track wait times for metrics
+        self._wait_times: dict[CriticalityLevel, list[float]] = {
+            level: [] for level in CriticalityLevel
+        }
+
+    async def enqueue(
+        self,
+        task_id: str,
+        criticality: CriticalityLevel,
+        task_data: dict[str, Any],
+        callback: Callable[[], Coroutine[Any, Any, Any]] | None = None,
+    ) -> bool:
+        """Add a task to the priority queue.
+
+        Args:
+            task_id: Unique identifier for the task
+            criticality: Criticality level determining priority
+            task_data: Data associated with the task
+            callback: Optional async callback to execute
+
+        Returns:
+            True if enqueued successfully, False if queue is full
+        """
+        async with self._lock:
+            if len(self._queue) >= self._max_size:
+                logger.warning(
+                    "Priority queue full (size=%d), rejecting task %s",
+                    len(self._queue),
+                    task_id,
+                )
+                return False
+
+            priority = self.PRIORITY_MAP[criticality]
+            timestamp = time.time()
+
+            task = PriorityTask(
+                priority=priority,
+                timestamp=timestamp,
+                task_id=task_id,
+                task_data=task_data,
+                callback=callback,
+            )
+
+            heapq.heappush(self._queue, task)
+            self._metrics.total_enqueued += 1
+            self._metrics.current_size = len(self._queue)
+
+            logger.debug(
+                "Enqueued task %s with criticality %s (priority=%d, queue_size=%d)",
+                task_id,
+                criticality.value,
+                priority,
+                len(self._queue),
+            )
+
+            return True
+
+    async def dequeue(self, timeout: float | None = None) -> PriorityTask | None:
+        """Remove and return the highest priority task from the queue.
+
+        Args:
+            timeout: Maximum time to wait for a task (seconds)
+
+        Returns:
+            PriorityTask if available, None if queue is empty or timeout
+        """
+        start_time = time.time()
+        deadline = start_time + timeout if timeout else None
+
+        while True:
+            async with self._lock:
+                if self._queue:
+                    task = heapq.heappop(self._queue)
+                    self._metrics.total_dequeued += 1
+                    self._metrics.current_size = len(self._queue)
+
+                    # Calculate wait time
+                    wait_time = time.time() - task.timestamp
+                    criticality = self._get_criticality_from_priority(task.priority)
+                    self._wait_times[criticality].append(wait_time)
+                    self._update_wait_time_metrics()
+
+                    logger.debug(
+                        "Dequeued task %s (priority=%d, wait_time=%.2fs, queue_size=%d)",
+                        task.task_id,
+                        task.priority,
+                        wait_time,
+                        len(self._queue),
+                    )
+
+                    return task
+
+            # Queue is empty
+            if deadline and time.time() >= deadline:
+                return None
+
+            # Wait a bit before checking again
+            await asyncio.sleep(0.1)
+
+    async def execute_with_timeout(
+        self,
+        task: PriorityTask,
+        timeout: float | None,
+    ) -> Any:
+        """Execute a task with timeout enforcement.
+
+        Args:
+            task: Task to execute
+            timeout: Timeout in seconds (None = no timeout)
+
+        Returns:
+            Result from task callback
+
+        Raises:
+            asyncio.TimeoutError: If task exceeds timeout
+            Exception: Any exception raised by the task callback
+        """
+        if not task.callback:
+            logger.warning("Task %s has no callback, skipping execution", task.task_id)
+            return None
+
+        criticality = self._get_criticality_from_priority(task.priority)
+
+        try:
+            if timeout:
+                result = await asyncio.wait_for(task.callback(), timeout=timeout)
+            else:
+                result = await task.callback()
+
+            logger.debug(
+                "Task %s completed successfully (criticality=%s)",
+                task.task_id,
+                criticality.value,
+            )
+            return result
+
+        except TimeoutError:
+            self._metrics.total_timeouts += 1
+            logger.error(
+                "Task %s timed out after %.2fs (criticality=%s)",
+                task.task_id,
+                timeout or 0.0,
+                criticality.value,
+            )
+            raise
+
+        except Exception as exc:
+            self._metrics.total_errors += 1
+            logger.exception(
+                "Task %s failed with error (criticality=%s): %s",
+                task.task_id,
+                criticality.value,
+                exc,
+            )
+            raise
+
+    def _get_criticality_from_priority(self, priority: int) -> CriticalityLevel:
+        """Convert priority back to criticality level."""
+        for level, prio in self.PRIORITY_MAP.items():
+            if prio == priority:
+                return level
+        return CriticalityLevel.NORMAL
+
+    def _update_wait_time_metrics(self) -> None:
+        """Update average and p95 wait time metrics."""
+        for level, times in self._wait_times.items():
+            if not times:
+                continue
+
+            # Keep only last 1000 measurements to avoid memory bloat
+            if len(times) > 1000:
+                self._wait_times[level] = times[-1000:]
+                times = self._wait_times[level]
+
+            # Calculate average
+            self._metrics.avg_wait_time[level] = sum(times) / len(times)
+
+            # Calculate P95
+            sorted_times = sorted(times)
+            p95_idx = int(len(sorted_times) * 0.95)
+            self._metrics.p95_wait_time[level] = sorted_times[p95_idx]
+
+    async def get_metrics(self) -> QueueMetrics:
+        """Get current queue metrics.
+
+        Returns:
+            QueueMetrics with current statistics
+        """
+        async with self._lock:
+            return QueueMetrics(
+                total_enqueued=self._metrics.total_enqueued,
+                total_dequeued=self._metrics.total_dequeued,
+                total_timeouts=self._metrics.total_timeouts,
+                total_errors=self._metrics.total_errors,
+                current_size=self._metrics.current_size,
+                avg_wait_time=dict(self._metrics.avg_wait_time),
+                p95_wait_time=dict(self._metrics.p95_wait_time),
+            )
+
+    async def size(self) -> int:
+        """Get current queue size.
+
+        Returns:
+            Number of tasks in queue
+        """
+        async with self._lock:
+            return len(self._queue)
+
+    async def clear(self) -> int:
+        """Clear all tasks from the queue.
+
+        Returns:
+            Number of tasks cleared
+        """
+        async with self._lock:
+            count = len(self._queue)
+            self._queue.clear()
+            self._metrics.current_size = 0
+            logger.info("Cleared %d tasks from priority queue", count)
+            return count

src/dashboard/__init__.py

@@ -0,0 +1,5 @@
+"""FastAPI dashboard package for observability APIs."""
+
+from src.dashboard.app import create_dashboard_app
+
+__all__ = ["create_dashboard_app"]

src/dashboard/app.py

@@ -0,0 +1,498 @@
+"""FastAPI application for observability dashboard endpoints."""
+
+from __future__ import annotations
+
+import json
+import os
+import sqlite3
+from datetime import UTC, datetime, timezone
+from pathlib import Path
+from typing import Any
+
+from fastapi import FastAPI, HTTPException, Query
+from fastapi.responses import FileResponse
+
+
+def create_dashboard_app(db_path: str, mode: str = "paper") -> FastAPI:
+    """Create dashboard FastAPI app bound to a SQLite database path."""
+    app = FastAPI(title="The Ouroboros Dashboard", version="1.0.0")
+    app.state.db_path = db_path
+    app.state.mode = mode
+
+    @app.get("/")
+    def index() -> FileResponse:
+        index_path = Path(__file__).parent / "static" / "index.html"
+        return FileResponse(index_path)
+
+    @app.get("/api/status")
+    def get_status() -> dict[str, Any]:
+        today = datetime.now(UTC).date().isoformat()
+        with _connect(db_path) as conn:
+            market_rows = conn.execute(
+                """
+                SELECT DISTINCT market FROM (
+                    SELECT market FROM trades WHERE DATE(timestamp) = ?
+                    UNION
+                    SELECT market FROM decision_logs WHERE DATE(timestamp) = ?
+                    UNION
+                    SELECT market FROM playbooks WHERE date = ?
+                ) ORDER BY market
+                """,
+                (today, today, today),
+            ).fetchall()
+            markets = [row[0] for row in market_rows] if market_rows else []
+            market_status: dict[str, Any] = {}
+            total_trades = 0
+            total_pnl = 0.0
+            total_decisions = 0
+            for market in markets:
+                trade_row = conn.execute(
+                    """
+                    SELECT COUNT(*) AS c, COALESCE(SUM(pnl), 0.0) AS p
+                    FROM trades
+                    WHERE DATE(timestamp) = ? AND market = ?
+                    """,
+                    (today, market),
+                ).fetchone()
+                decision_row = conn.execute(
+                    """
+                    SELECT COUNT(*) AS c
+                    FROM decision_logs
+                    WHERE DATE(timestamp) = ? AND market = ?
+                    """,
+                    (today, market),
+                ).fetchone()
+                playbook_row = conn.execute(
+                    """
+                    SELECT status
+                    FROM playbooks
+                    WHERE date = ? AND market = ?
+                    LIMIT 1
+                    """,
+                    (today, market),
+                ).fetchone()
+                market_status[market] = {
+                    "trade_count": int(trade_row["c"] if trade_row else 0),
+                    "total_pnl": float(trade_row["p"] if trade_row else 0.0),
+                    "decision_count": int(decision_row["c"] if decision_row else 0),
+                    "playbook_status": playbook_row["status"] if playbook_row else None,
+                }
+                total_trades += market_status[market]["trade_count"]
+                total_pnl += market_status[market]["total_pnl"]
+                total_decisions += market_status[market]["decision_count"]
+
+            cb_threshold = float(os.getenv("CIRCUIT_BREAKER_PCT", "-3.0"))
+            pnl_pct_rows = conn.execute(
+                """
+                SELECT key, value
+                FROM system_metrics
+                WHERE key LIKE 'portfolio_pnl_pct_%'
+                ORDER BY updated_at DESC
+                LIMIT 20
+                """
+            ).fetchall()
+            current_pnl_pct: float | None = None
+            if pnl_pct_rows:
+                values = [
+                    json.loads(row["value"]).get("pnl_pct")
+                    for row in pnl_pct_rows
+                    if json.loads(row["value"]).get("pnl_pct") is not None
+                ]
+                if values:
+                    current_pnl_pct = round(min(values), 4)
+
+            if current_pnl_pct is None:
+                cb_status = "unknown"
+            elif current_pnl_pct <= cb_threshold:
+                cb_status = "tripped"
+            elif current_pnl_pct <= cb_threshold + 1.0:
+                cb_status = "warning"
+            else:
+                cb_status = "ok"
+
+            return {
+                "date": today,
+                "mode": mode,
+                "markets": market_status,
+                "totals": {
+                    "trade_count": total_trades,
+                    "total_pnl": round(total_pnl, 2),
+                    "decision_count": total_decisions,
+                },
+                "circuit_breaker": {
+                    "threshold_pct": cb_threshold,
+                    "current_pnl_pct": current_pnl_pct,
+                    "status": cb_status,
+                },
+            }
+
+    @app.get("/api/playbook/{date_str}")
+    def get_playbook(date_str: str, market: str = Query("KR")) -> dict[str, Any]:
+        with _connect(db_path) as conn:
+            row = conn.execute(
+                """
+                SELECT date, market, status, playbook_json, generated_at,
+                       token_count, scenario_count, match_count
+                FROM playbooks
+                WHERE date = ? AND market = ?
+                """,
+                (date_str, market),
+            ).fetchone()
+            if row is None:
+                raise HTTPException(status_code=404, detail="playbook not found")
+            return {
+                "date": row["date"],
+                "market": row["market"],
+                "status": row["status"],
+                "playbook": json.loads(row["playbook_json"]),
+                "generated_at": row["generated_at"],
+                "token_count": row["token_count"],
+                "scenario_count": row["scenario_count"],
+                "match_count": row["match_count"],
+            }
+
+    @app.get("/api/scorecard/{date_str}")
+    def get_scorecard(date_str: str, market: str = Query("KR")) -> dict[str, Any]:
+        key = f"scorecard_{market}"
+        with _connect(db_path) as conn:
+            row = conn.execute(
+                """
+                SELECT value
+                FROM contexts
+                WHERE layer = 'L6_DAILY' AND timeframe = ? AND key = ?
+                """,
+                (date_str, key),
+            ).fetchone()
+            if row is None:
+                raise HTTPException(status_code=404, detail="scorecard not found")
+            return {"date": date_str, "market": market, "scorecard": json.loads(row["value"])}
+
+    @app.get("/api/performance")
+    def get_performance(market: str = Query("all")) -> dict[str, Any]:
+        with _connect(db_path) as conn:
+            if market == "all":
+                by_market_rows = conn.execute(
+                    """
+                    SELECT market,
+                           COUNT(*) AS total_trades,
+                           SUM(CASE WHEN pnl > 0 THEN 1 ELSE 0 END) AS wins,
+                           SUM(CASE WHEN pnl < 0 THEN 1 ELSE 0 END) AS losses,
+                           COALESCE(SUM(pnl), 0.0) AS total_pnl,
+                           COALESCE(AVG(confidence), 0.0) AS avg_confidence
+                    FROM trades
+                    GROUP BY market
+                    ORDER BY market
+                    """
+                ).fetchall()
+                combined = _performance_from_rows(by_market_rows)
+                return {
+                    "market": "all",
+                    "combined": combined,
+                    "by_market": [
+                        _row_to_performance(row)
+                        for row in by_market_rows
+                    ],
+                }
+
+            row = conn.execute(
+                """
+                SELECT market,
+                       COUNT(*) AS total_trades,
+                       SUM(CASE WHEN pnl > 0 THEN 1 ELSE 0 END) AS wins,
+                       SUM(CASE WHEN pnl < 0 THEN 1 ELSE 0 END) AS losses,
+                       COALESCE(SUM(pnl), 0.0) AS total_pnl,
+                       COALESCE(AVG(confidence), 0.0) AS avg_confidence
+                FROM trades
+                WHERE market = ?
+                GROUP BY market
+                """,
+                (market,),
+            ).fetchone()
+            if row is None:
+                return {"market": market, "metrics": _empty_performance(market)}
+            return {"market": market, "metrics": _row_to_performance(row)}
+
+    @app.get("/api/context/{layer}")
+    def get_context_layer(
+        layer: str,
+        timeframe: str | None = Query(default=None),
+        limit: int = Query(default=100, ge=1, le=1000),
+    ) -> dict[str, Any]:
+        with _connect(db_path) as conn:
+            if timeframe is None:
+                rows = conn.execute(
+                    """
+                    SELECT timeframe, key, value, updated_at
+                    FROM contexts
+                    WHERE layer = ?
+                    ORDER BY updated_at DESC
+                    LIMIT ?
+                    """,
+                    (layer, limit),
+                ).fetchall()
+            else:
+                rows = conn.execute(
+                    """
+                    SELECT timeframe, key, value, updated_at
+                    FROM contexts
+                    WHERE layer = ? AND timeframe = ?
+                    ORDER BY key
+                    LIMIT ?
+                    """,
+                    (layer, timeframe, limit),
+                ).fetchall()
+
+            entries = [
+                {
+                    "timeframe": row["timeframe"],
+                    "key": row["key"],
+                    "value": json.loads(row["value"]),
+                    "updated_at": row["updated_at"],
+                }
+                for row in rows
+            ]
+            return {
+                "layer": layer,
+                "timeframe": timeframe,
+                "count": len(entries),
+                "entries": entries,
+            }
+
+    @app.get("/api/decisions")
+    def get_decisions(
+        market: str = Query("KR"),
+        limit: int = Query(default=50, ge=1, le=500),
+    ) -> dict[str, Any]:
+        with _connect(db_path) as conn:
+            rows = conn.execute(
+                """
+                SELECT decision_id, timestamp, stock_code, market, exchange_code,
+                       action, confidence, rationale, context_snapshot, input_data,
+                       outcome_pnl, outcome_accuracy
+                FROM decision_logs
+                WHERE market = ?
+                ORDER BY timestamp DESC
+                LIMIT ?
+                """,
+                (market, limit),
+            ).fetchall()
+            decisions = []
+            for row in rows:
+                decisions.append(
+                    {
+                        "decision_id": row["decision_id"],
+                        "timestamp": row["timestamp"],
+                        "stock_code": row["stock_code"],
+                        "market": row["market"],
+                        "exchange_code": row["exchange_code"],
+                        "action": row["action"],
+                        "confidence": row["confidence"],
+                        "rationale": row["rationale"],
+                        "context_snapshot": json.loads(row["context_snapshot"]),
+                        "input_data": json.loads(row["input_data"]),
+                        "outcome_pnl": row["outcome_pnl"],
+                        "outcome_accuracy": row["outcome_accuracy"],
+                    }
+                )
+            return {"market": market, "count": len(decisions), "decisions": decisions}
+
+    @app.get("/api/pnl/history")
+    def get_pnl_history(
+        days: int = Query(default=30, ge=1, le=365),
+        market: str = Query("all"),
+    ) -> dict[str, Any]:
+        """Return daily P&L history for charting."""
+        with _connect(db_path) as conn:
+            if market == "all":
+                rows = conn.execute(
+                    """
+                    SELECT DATE(timestamp) AS date,
+                           SUM(pnl) AS daily_pnl,
+                           COUNT(*) AS trade_count
+                    FROM trades
+                    WHERE pnl IS NOT NULL
+                      AND DATE(timestamp) >= DATE('now', ?)
+                    GROUP BY DATE(timestamp)
+                    ORDER BY DATE(timestamp)
+                    """,
+                    (f"-{days} days",),
+                ).fetchall()
+            else:
+                rows = conn.execute(
+                    """
+                    SELECT DATE(timestamp) AS date,
+                           SUM(pnl) AS daily_pnl,
+                           COUNT(*) AS trade_count
+                    FROM trades
+                    WHERE pnl IS NOT NULL
+                      AND market = ?
+                      AND DATE(timestamp) >= DATE('now', ?)
+                    GROUP BY DATE(timestamp)
+                    ORDER BY DATE(timestamp)
+                    """,
+                    (market, f"-{days} days"),
+                ).fetchall()
+            return {
+                "days": days,
+                "market": market,
+                "labels": [row["date"] for row in rows],
+                "pnl": [round(float(row["daily_pnl"]), 2) for row in rows],
+                "trades": [int(row["trade_count"]) for row in rows],
+            }
+
+    @app.get("/api/scenarios/active")
+    def get_active_scenarios(
+        market: str = Query("US"),
+        date_str: str | None = Query(default=None),
+        limit: int = Query(default=50, ge=1, le=500),
+    ) -> dict[str, Any]:
+        if date_str is None:
+            date_str = datetime.now(UTC).date().isoformat()
+
+        with _connect(db_path) as conn:
+            rows = conn.execute(
+                """
+                SELECT timestamp, stock_code, action, confidence, rationale, context_snapshot
+                FROM decision_logs
+                WHERE market = ? AND DATE(timestamp) = ?
+                ORDER BY timestamp DESC
+                LIMIT ?
+                """,
+                (market, date_str, limit),
+            ).fetchall()
+            matches: list[dict[str, Any]] = []
+            for row in rows:
+                snapshot = json.loads(row["context_snapshot"])
+                scenario_match = snapshot.get("scenario_match", {})
+                if not isinstance(scenario_match, dict) or not scenario_match:
+                    continue
+                matches.append(
+                    {
+                        "timestamp": row["timestamp"],
+                        "stock_code": row["stock_code"],
+                        "action": row["action"],
+                        "confidence": row["confidence"],
+                        "rationale": row["rationale"],
+                        "scenario_match": scenario_match,
+                    }
+                )
+            return {"market": market, "date": date_str, "count": len(matches), "matches": matches}
+
+    @app.get("/api/positions")
+    def get_positions() -> dict[str, Any]:
+        """Return all currently open positions (last trade per symbol is BUY)."""
+        with _connect(db_path) as conn:
+            rows = conn.execute(
+                """
+                SELECT stock_code, market, exchange_code,
+                       price AS entry_price, quantity, timestamp AS entry_time,
+                       decision_id
+                FROM (
+                    SELECT stock_code, market, exchange_code, price, quantity,
+                           timestamp, decision_id, action,
+                           ROW_NUMBER() OVER (
+                               PARTITION BY stock_code, market
+                               ORDER BY timestamp DESC
+                           ) AS rn
+                    FROM trades
+                )
+                WHERE rn = 1 AND action = 'BUY'
+                ORDER BY entry_time DESC
+                """
+            ).fetchall()
+
+            now = datetime.now(timezone.utc)
+            positions = []
+            for row in rows:
+                entry_time_str = row["entry_time"]
+                try:
+                    entry_dt = datetime.fromisoformat(entry_time_str.replace("Z", "+00:00"))
+                    held_seconds = int((now - entry_dt).total_seconds())
+                    held_hours = held_seconds // 3600
+                    held_minutes = (held_seconds % 3600) // 60
+                    if held_hours >= 1:
+                        held_display = f"{held_hours}h {held_minutes}m"
+                    else:
+                        held_display = f"{held_minutes}m"
+                except (ValueError, TypeError):
+                    held_display = "--"
+
+                positions.append(
+                    {
+                        "stock_code": row["stock_code"],
+                        "market": row["market"],
+                        "exchange_code": row["exchange_code"],
+                        "entry_price": row["entry_price"],
+                        "quantity": row["quantity"],
+                        "entry_time": entry_time_str,
+                        "held": held_display,
+                        "decision_id": row["decision_id"],
+                    }
+                )
+
+            return {"count": len(positions), "positions": positions}
+
+    return app
+
+
+def _connect(db_path: str) -> sqlite3.Connection:
+    conn = sqlite3.connect(db_path)
+    conn.row_factory = sqlite3.Row
+    conn.execute("PRAGMA journal_mode=WAL")
+    conn.execute("PRAGMA busy_timeout=8000")
+    return conn
+
+
+def _row_to_performance(row: sqlite3.Row) -> dict[str, Any]:
+    wins = int(row["wins"] or 0)
+    losses = int(row["losses"] or 0)
+    total = int(row["total_trades"] or 0)
+    win_rate = round((wins / (wins + losses) * 100), 2) if (wins + losses) > 0 else 0.0
+    return {
+        "market": row["market"],
+        "total_trades": total,
+        "wins": wins,
+        "losses": losses,
+        "win_rate": win_rate,
+        "total_pnl": round(float(row["total_pnl"] or 0.0), 2),
+        "avg_confidence": round(float(row["avg_confidence"] or 0.0), 2),
+    }
+
+
+def _performance_from_rows(rows: list[sqlite3.Row]) -> dict[str, Any]:
+    total_trades = 0
+    wins = 0
+    losses = 0
+    total_pnl = 0.0
+    confidence_weighted = 0.0
+    for row in rows:
+        market_total = int(row["total_trades"] or 0)
+        market_conf = float(row["avg_confidence"] or 0.0)
+        total_trades += market_total
+        wins += int(row["wins"] or 0)
+        losses += int(row["losses"] or 0)
+        total_pnl += float(row["total_pnl"] or 0.0)
+        confidence_weighted += market_total * market_conf
+    win_rate = round((wins / (wins + losses) * 100), 2) if (wins + losses) > 0 else 0.0
+    avg_confidence = round(confidence_weighted / total_trades, 2) if total_trades > 0 else 0.0
+    return {
+        "market": "all",
+        "total_trades": total_trades,
+        "wins": wins,
+        "losses": losses,
+        "win_rate": win_rate,
+        "total_pnl": round(total_pnl, 2),
+        "avg_confidence": avg_confidence,
+    }
+
+
+def _empty_performance(market: str) -> dict[str, Any]:
+    return {
+        "market": market,
+        "total_trades": 0,
+        "wins": 0,
+        "losses": 0,
+        "win_rate": 0.0,
+        "total_pnl": 0.0,
+        "avg_confidence": 0.0,
+    }

src/dashboard/static/index.html

@@ -0,0 +1,798 @@
+<!doctype html>
+<html lang="ko">
+  <head>
+    <meta charset="UTF-8" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
+    <title>The Ouroboros Dashboard</title>
+    <script src="https://cdn.jsdelivr.net/npm/chart.js@4.4.0/dist/chart.umd.min.js"></script>
+    <style>
+      :root {
+        --bg: #0b1724;
+        --panel: #12263a;
+        --fg: #e6eef7;
+        --muted: #9fb3c8;
+        --accent: #3cb371;
+        --red: #e05555;
+        --warn: #e8a040;
+        --border: #28455f;
+      }
+      * { box-sizing: border-box; margin: 0; padding: 0; }
+      body {
+        font-family: ui-monospace, SFMono-Regular, Menlo, monospace;
+        background: radial-gradient(circle at top left, #173b58, var(--bg));
+        color: var(--fg);
+        min-height: 100vh;
+        font-size: 13px;
+      }
+      .wrap { max-width: 1100px; margin: 0 auto; padding: 20px 16px; }
+
+      /* Header */
+      header {
+        display: flex;
+        align-items: center;
+        justify-content: space-between;
+        margin-bottom: 20px;
+        padding-bottom: 12px;
+        border-bottom: 1px solid var(--border);
+      }
+      header h1 { font-size: 18px; color: var(--accent); letter-spacing: 0.5px; }
+      .header-right { display: flex; align-items: center; gap: 12px; color: var(--muted); font-size: 12px; }
+      .refresh-btn {
+        background: none; border: 1px solid var(--border); color: var(--muted);
+        padding: 4px 10px; border-radius: 6px; cursor: pointer; font-family: inherit;
+        font-size: 12px; transition: border-color 0.2s;
+      }
+      .refresh-btn:hover { border-color: var(--accent); color: var(--accent); }
+      .mode-badge {
+        padding: 3px 10px; border-radius: 5px; font-size: 12px; font-weight: 700;
+        letter-spacing: 0.5px;
+      }
+      .mode-badge.live {
+        background: rgba(224, 85, 85, 0.15); color: var(--red);
+        border: 1px solid rgba(224, 85, 85, 0.4);
+        animation: pulse-warn 2s ease-in-out infinite;
+      }
+      .mode-badge.paper {
+        background: rgba(232, 160, 64, 0.15); color: var(--warn);
+        border: 1px solid rgba(232, 160, 64, 0.4);
+      }
+
+      /* CB Gauge */
+      .cb-gauge-wrap {
+        display: flex; align-items: center; gap: 8px;
+        font-size: 11px; color: var(--muted);
+      }
+      .cb-dot {
+        width: 8px; height: 8px; border-radius: 50%; flex-shrink: 0;
+      }
+      .cb-dot.ok      { background: var(--accent); }
+      .cb-dot.warning { background: var(--warn); animation: pulse-warn 1.2s ease-in-out infinite; }
+      .cb-dot.tripped { background: var(--red);  animation: pulse-warn 0.6s ease-in-out infinite; }
+      .cb-dot.unknown { background: var(--border); }
+      @keyframes pulse-warn {
+        0%, 100% { opacity: 1; }
+        50%       { opacity: 0.35; }
+      }
+      .cb-bar-wrap { width: 64px; height: 5px; background: rgba(255,255,255,0.08); border-radius: 3px; overflow: hidden; }
+      .cb-bar-fill { height: 100%; border-radius: 3px; transition: width 0.4s, background 0.4s; }
+
+      /* Summary cards */
+      .cards { display: grid; grid-template-columns: repeat(4, 1fr); gap: 12px; margin-bottom: 20px; }
+      @media (max-width: 700px) { .cards { grid-template-columns: repeat(2, 1fr); } }
+      .card {
+        background: var(--panel);
+        border: 1px solid var(--border);
+        border-radius: 10px;
+        padding: 16px;
+      }
+      .card-label { color: var(--muted); font-size: 11px; margin-bottom: 6px; text-transform: uppercase; letter-spacing: 0.5px; }
+      .card-value { font-size: 22px; font-weight: 700; }
+      .card-sub { color: var(--muted); font-size: 11px; margin-top: 4px; }
+      .positive { color: var(--accent); }
+      .negative { color: var(--red); }
+      .neutral { color: var(--fg); }
+
+      /* Chart panel */
+      .chart-panel {
+        background: var(--panel);
+        border: 1px solid var(--border);
+        border-radius: 10px;
+        padding: 16px;
+        margin-bottom: 20px;
+      }
+      .panel-header {
+        display: flex;
+        align-items: center;
+        justify-content: space-between;
+        margin-bottom: 16px;
+      }
+      .panel-title { font-size: 13px; color: var(--muted); font-weight: 600; }
+      .chart-container { position: relative; height: 180px; }
+      .chart-error { color: var(--muted); text-align: center; padding: 40px 0; font-size: 12px; }
+
+      /* Days selector */
+      .days-selector { display: flex; gap: 4px; }
+      .day-btn {
+        background: none; border: 1px solid var(--border); color: var(--muted);
+        padding: 3px 8px; border-radius: 4px; cursor: pointer; font-family: inherit; font-size: 11px;
+      }
+      .day-btn.active { border-color: var(--accent); color: var(--accent); background: rgba(60, 179, 113, 0.08); }
+
+      /* Decisions panel */
+      .decisions-panel {
+        background: var(--panel);
+        border: 1px solid var(--border);
+        border-radius: 10px;
+        padding: 16px;
+      }
+      .market-tabs { display: flex; gap: 6px; flex-wrap: wrap; }
+      .tab-btn {
+        background: none; border: 1px solid var(--border); color: var(--muted);
+        padding: 4px 10px; border-radius: 6px; cursor: pointer; font-family: inherit; font-size: 11px;
+      }
+      .tab-btn.active { border-color: var(--accent); color: var(--accent); background: rgba(60, 179, 113, 0.08); }
+      .decisions-table { width: 100%; border-collapse: collapse; margin-top: 14px; }
+      .decisions-table th {
+        text-align: left; color: var(--muted); font-size: 11px; font-weight: 600;
+        padding: 6px 8px; border-bottom: 1px solid var(--border); white-space: nowrap;
+      }
+      .decisions-table td {
+        padding: 8px 8px; border-bottom: 1px solid rgba(40, 69, 95, 0.5);
+        vertical-align: middle; white-space: nowrap;
+      }
+      .decisions-table tr:last-child td { border-bottom: none; }
+      .decisions-table tr:hover td { background: rgba(255,255,255,0.02); }
+      .badge {
+        display: inline-block; padding: 2px 7px; border-radius: 4px;
+        font-size: 11px; font-weight: 700; letter-spacing: 0.5px;
+      }
+      .badge-buy { background: rgba(60, 179, 113, 0.15); color: var(--accent); }
+      .badge-sell { background: rgba(224, 85, 85, 0.15); color: var(--red); }
+      .badge-hold { background: rgba(159, 179, 200, 0.12); color: var(--muted); }
+      .conf-bar-wrap { display: flex; align-items: center; gap: 6px; min-width: 90px; }
+      .conf-bar { flex: 1; height: 6px; background: rgba(255,255,255,0.08); border-radius: 3px; overflow: hidden; }
+      .conf-fill { height: 100%; border-radius: 3px; background: var(--accent); transition: width 0.3s; }
+      .conf-val { color: var(--muted); font-size: 11px; min-width: 26px; text-align: right; }
+      .rationale-cell { max-width: 200px; overflow: hidden; text-overflow: ellipsis; color: var(--muted); }
+      .empty-row td { text-align: center; color: var(--muted); padding: 24px; }
+
+      /* Positions panel */
+      .positions-panel {
+        background: var(--panel);
+        border: 1px solid var(--border);
+        border-radius: 10px;
+        padding: 16px;
+        margin-bottom: 20px;
+      }
+      .positions-table { width: 100%; border-collapse: collapse; margin-top: 14px; }
+      .positions-table th {
+        text-align: left; color: var(--muted); font-size: 11px; font-weight: 600;
+        padding: 6px 8px; border-bottom: 1px solid var(--border); white-space: nowrap;
+      }
+      .positions-table td {
+        padding: 8px 8px; border-bottom: 1px solid rgba(40, 69, 95, 0.5);
+        vertical-align: middle; white-space: nowrap;
+      }
+      .positions-table tr:last-child td { border-bottom: none; }
+      .positions-table tr:hover td { background: rgba(255,255,255,0.02); }
+      .pos-empty { color: var(--muted); text-align: center; padding: 20px 0; font-size: 12px; }
+      .pos-count {
+        display: inline-block; background: rgba(60, 179, 113, 0.12);
+        color: var(--accent); font-size: 11px; font-weight: 700;
+        padding: 2px 8px; border-radius: 10px; margin-left: 8px;
+      }
+
+      /* Spinner */
+      .spinner { display: inline-block; width: 12px; height: 12px; border: 2px solid var(--border); border-top-color: var(--accent); border-radius: 50%; animation: spin 0.8s linear infinite; }
+      @keyframes spin { to { transform: rotate(360deg); } }
+
+      /* Generic panel */
+      .panel {
+        background: var(--panel);
+        border: 1px solid var(--border);
+        border-radius: 10px;
+        padding: 16px;
+        margin-top: 20px;
+      }
+
+      /* Playbook panel - details/summary accordion */
+      .playbook-panel details { border: 1px solid var(--border); border-radius: 4px; margin-bottom: 6px; }
+      .playbook-panel summary { padding: 8px 12px; cursor: pointer; font-weight: 600; background: var(--bg); color: var(--fg); }
+      .playbook-panel summary:hover { color: var(--accent); }
+      .playbook-panel pre { margin: 0; padding: 12px; background: var(--bg); overflow-x: auto;
+                            font-size: 11px; color: #a0c4ff; white-space: pre-wrap; }
+
+      /* Scorecard KPI card grid */
+      .scorecard-grid { display: grid; grid-template-columns: repeat(auto-fill, minmax(140px, 1fr)); gap: 10px; }
+      .kpi-card { background: var(--bg); border: 1px solid var(--border); border-radius: 6px; padding: 12px; text-align: center; }
+      .kpi-card .kpi-label { font-size: 11px; color: var(--muted); margin-bottom: 4px; }
+      .kpi-card .kpi-value { font-size: 20px; font-weight: 700; color: var(--fg); }
+
+      /* Scenarios table */
+      .scenarios-table { width: 100%; border-collapse: collapse; font-size: 13px; }
+      .scenarios-table th { background: var(--bg); padding: 8px; text-align: left; border-bottom: 1px solid var(--border);
+                            color: var(--muted); font-size: 11px; font-weight: 600; white-space: nowrap; }
+      .scenarios-table td { padding: 7px 8px; border-bottom: 1px solid rgba(40,69,95,0.5); }
+      .scenarios-table tr:hover td { background: rgba(255,255,255,0.02); }
+
+      /* Context table */
+      .context-table { width: 100%; border-collapse: collapse; font-size: 12px; }
+      .context-table th { background: var(--bg); padding: 8px; text-align: left; border-bottom: 1px solid var(--border);
+                          color: var(--muted); font-size: 11px; font-weight: 600; white-space: nowrap; }
+      .context-table td { padding: 6px 8px; border-bottom: 1px solid rgba(40,69,95,0.5); vertical-align: top; }
+      .context-value { max-height: 60px; overflow-y: auto; color: #a0c4ff; word-break: break-all; }
+
+      /* Common panel select controls */
+      .panel-controls { display: flex; gap: 8px; align-items: center; flex-wrap: wrap; }
+      .panel-controls select, .panel-controls input[type="number"] {
+        background: var(--bg); color: var(--fg); border: 1px solid var(--border);
+        border-radius: 4px; padding: 4px 8px; font-size: 13px; font-family: inherit;
+      }
+      .panel-date { color: var(--muted); font-size: 12px; }
+      .empty-msg { color: var(--muted); text-align: center; padding: 20px 0; font-size: 12px; }
+    </style>
+  </head>
+  <body>
+    <div class="wrap">
+      <!-- Header -->
+      <header>
+        <h1>&#x1F40D; The Ouroboros</h1>
+        <div class="header-right">
+          <span class="mode-badge" id="mode-badge">--</span>
+          <div class="cb-gauge-wrap" id="cb-gauge" title="Circuit Breaker">
+            <span class="cb-dot unknown" id="cb-dot"></span>
+            <span id="cb-label">CB --</span>
+            <div class="cb-bar-wrap">
+              <div class="cb-bar-fill" id="cb-bar" style="width:0%;background:var(--accent)"></div>
+            </div>
+          </div>
+          <span id="last-updated">--</span>
+          <button class="refresh-btn" onclick="refreshAll()">&#x21BA; 새로고침</button>
+        </div>
+      </header>
+
+      <!-- Summary cards -->
+      <div class="cards">
+        <div class="card">
+          <div class="card-label">오늘 거래</div>
+          <div class="card-value neutral" id="card-trades">--</div>
+          <div class="card-sub" id="card-trades-sub">거래 건수</div>
+        </div>
+        <div class="card">
+          <div class="card-label">오늘 P&amp;L</div>
+          <div class="card-value" id="card-pnl">--</div>
+          <div class="card-sub" id="card-pnl-sub">실현 손익</div>
+        </div>
+        <div class="card">
+          <div class="card-label">승률</div>
+          <div class="card-value neutral" id="card-winrate">--</div>
+          <div class="card-sub">전체 누적</div>
+        </div>
+        <div class="card">
+          <div class="card-label">누적 거래</div>
+          <div class="card-value neutral" id="card-total">--</div>
+          <div class="card-sub">전체 기간</div>
+        </div>
+      </div>
+
+      <!-- Open Positions -->
+      <div class="positions-panel">
+        <div class="panel-header">
+          <span class="panel-title">
+            현재 보유 포지션
+            <span class="pos-count" id="positions-count">0</span>
+          </span>
+        </div>
+        <table class="positions-table">
+          <thead>
+            <tr>
+              <th>종목</th>
+              <th>시장</th>
+              <th>수량</th>
+              <th>진입가</th>
+              <th>보유 시간</th>
+            </tr>
+          </thead>
+          <tbody id="positions-body">
+            <tr><td colspan="5" class="pos-empty"><span class="spinner"></span></td></tr>
+          </tbody>
+        </table>
+      </div>
+
+      <!-- P&L Chart -->
+      <div class="chart-panel">
+        <div class="panel-header">
+          <span class="panel-title">P&amp;L 추이</span>
+          <div class="days-selector">
+            <button class="day-btn active" data-days="7" onclick="selectDays(this)">7일</button>
+            <button class="day-btn" data-days="30" onclick="selectDays(this)">30일</button>
+            <button class="day-btn" data-days="90" onclick="selectDays(this)">90일</button>
+          </div>
+        </div>
+        <div class="chart-container">
+          <canvas id="pnl-chart"></canvas>
+          <div class="chart-error" id="chart-error" style="display:none">데이터 없음</div>
+        </div>
+      </div>
+
+      <!-- Decisions log -->
+      <div class="decisions-panel">
+        <div class="panel-header">
+          <span class="panel-title">최근 결정 로그</span>
+          <div class="market-tabs" id="market-tabs">
+            <button class="tab-btn active" data-market="KR" onclick="selectMarket(this)">KR</button>
+            <button class="tab-btn" data-market="US_NASDAQ" onclick="selectMarket(this)">US_NASDAQ</button>
+            <button class="tab-btn" data-market="US_NYSE" onclick="selectMarket(this)">US_NYSE</button>
+            <button class="tab-btn" data-market="JP" onclick="selectMarket(this)">JP</button>
+            <button class="tab-btn" data-market="HK" onclick="selectMarket(this)">HK</button>
+          </div>
+        </div>
+        <table class="decisions-table">
+          <thead>
+            <tr>
+              <th>시각</th>
+              <th>종목</th>
+              <th>액션</th>
+              <th>신뢰도</th>
+              <th>사유</th>
+            </tr>
+          </thead>
+          <tbody id="decisions-body">
+            <tr class="empty-row"><td colspan="5"><span class="spinner"></span></td></tr>
+          </tbody>
+        </table>
+      </div>
+
+      <!-- playbook panel -->
+      <div class="panel playbook-panel">
+        <div class="panel-header">
+          <span class="panel-title">&#x1F4CB; 프리마켓 플레이북</span>
+          <div class="panel-controls">
+            <select id="pb-market-select" onchange="fetchPlaybook()">
+              <option value="KR">KR</option>
+              <option value="US_NASDAQ">US_NASDAQ</option>
+              <option value="US_NYSE">US_NYSE</option>
+            </select>
+            <span id="pb-date" class="panel-date"></span>
+          </div>
+        </div>
+        <div id="playbook-content"><p class="empty-msg">데이터 없음</p></div>
+      </div>
+
+      <!-- scorecard panel -->
+      <div class="panel">
+        <div class="panel-header">
+          <span class="panel-title">&#x1F4CA; 일간 스코어카드</span>
+          <div class="panel-controls">
+            <select id="sc-market-select" onchange="fetchScorecard()">
+              <option value="KR">KR</option>
+              <option value="US_NASDAQ">US_NASDAQ</option>
+            </select>
+            <span id="sc-date" class="panel-date"></span>
+          </div>
+        </div>
+        <div id="scorecard-grid" class="scorecard-grid"><p class="empty-msg">데이터 없음</p></div>
+      </div>
+
+      <!-- scenarios panel -->
+      <div class="panel">
+        <div class="panel-header">
+          <span class="panel-title">&#x1F3AF; 활성 시나리오 매칭</span>
+          <div class="panel-controls">
+            <select id="scen-market-select" onchange="fetchScenarios()">
+              <option value="KR">KR</option>
+              <option value="US_NASDAQ">US_NASDAQ</option>
+            </select>
+          </div>
+        </div>
+        <div id="scenarios-content"><p class="empty-msg">데이터 없음</p></div>
+      </div>
+
+      <!-- context layer panel -->
+      <div class="panel">
+        <div class="panel-header">
+          <span class="panel-title">&#x1F9E0; 컨텍스트 트리</span>
+          <div class="panel-controls">
+            <select id="ctx-layer-select" onchange="fetchContext()">
+              <option value="L7_REALTIME">L7_REALTIME</option>
+              <option value="L6_DAILY">L6_DAILY</option>
+              <option value="L5_WEEKLY">L5_WEEKLY</option>
+              <option value="L4_MONTHLY">L4_MONTHLY</option>
+              <option value="L3_QUARTERLY">L3_QUARTERLY</option>
+              <option value="L2_YEARLY">L2_YEARLY</option>
+              <option value="L1_LIFETIME">L1_LIFETIME</option>
+            </select>
+            <input id="ctx-limit" type="number" value="20" min="1" max="200"
+                   style="width:60px;" onchange="fetchContext()">
+          </div>
+        </div>
+        <div id="context-content"><p class="empty-msg">데이터 없음</p></div>
+      </div>
+    </div>
+
+    <script>
+      let pnlChart = null;
+      let currentDays = 7;
+      let currentMarket = 'KR';
+
+      function fmt(dt) {
+        try {
+          const d = new Date(dt);
+          return d.toLocaleTimeString('ko-KR', { hour: '2-digit', minute: '2-digit', hour12: false });
+        } catch { return dt || '--'; }
+      }
+
+      function fmtPnl(v) {
+        if (v === null || v === undefined) return '--';
+        const n = parseFloat(v);
+        const cls = n > 0 ? 'positive' : n < 0 ? 'negative' : 'neutral';
+        const sign = n > 0 ? '+' : '';
+        return `<span class="${cls}">${sign}${n.toFixed(2)}</span>`;
+      }
+
+      function badge(action) {
+        const a = (action || '').toUpperCase();
+        const cls = a === 'BUY' ? 'badge-buy' : a === 'SELL' ? 'badge-sell' : 'badge-hold';
+        return `<span class="badge ${cls}">${a}</span>`;
+      }
+
+      function confBar(conf) {
+        const pct = Math.min(Math.max(conf || 0, 0), 100);
+        return `<div class="conf-bar-wrap">
+          <div class="conf-bar"><div class="conf-fill" style="width:${pct}%"></div></div>
+          <span class="conf-val">${pct}</span>
+        </div>`;
+      }
+
+      function fmtPrice(v, market) {
+        if (v === null || v === undefined) return '--';
+        const n = parseFloat(v);
+        const sym = market === 'KR' ? '₩' : market === 'JP' ? '¥' : market === 'HK' ? 'HK$' : '$';
+        return sym + n.toLocaleString('en-US', { minimumFractionDigits: 0, maximumFractionDigits: 4 });
+      }
+
+      async function fetchPositions() {
+        const tbody = document.getElementById('positions-body');
+        const countEl = document.getElementById('positions-count');
+        try {
+          const r = await fetch('/api/positions');
+          if (!r.ok) throw new Error('fetch failed');
+          const d = await r.json();
+          countEl.textContent = d.count ?? 0;
+          if (!d.positions || d.positions.length === 0) {
+            tbody.innerHTML = '<tr><td colspan="5" class="pos-empty">현재 보유 중인 포지션 없음</td></tr>';
+            return;
+          }
+          tbody.innerHTML = d.positions.map(p => `
+            <tr>
+              <td><strong>${p.stock_code || '--'}</strong></td>
+              <td><span style="color:var(--muted);font-size:11px">${p.market || '--'}</span></td>
+              <td>${p.quantity ?? '--'}</td>
+              <td>${fmtPrice(p.entry_price, p.market)}</td>
+              <td style="color:var(--muted);font-size:11px">${p.held || '--'}</td>
+            </tr>
+          `).join('');
+        } catch {
+          tbody.innerHTML = '<tr><td colspan="5" class="pos-empty">데이터 로드 실패</td></tr>';
+        }
+      }
+
+      function renderCbGauge(cb) {
+        if (!cb) return;
+        const dot = document.getElementById('cb-dot');
+        const label = document.getElementById('cb-label');
+        const bar = document.getElementById('cb-bar');
+
+        const status = cb.status || 'unknown';
+        const threshold = cb.threshold_pct ?? -3.0;
+        const current = cb.current_pnl_pct;
+
+        // dot color
+        dot.className = `cb-dot ${status}`;
+
+        // label
+        if (current !== null && current !== undefined) {
+          const sign = current > 0 ? '+' : '';
+          label.textContent = `CB ${sign}${current.toFixed(2)}%`;
+        } else {
+          label.textContent = 'CB --';
+        }
+
+        // bar: fill = how much of the threshold has been consumed (0%=safe, 100%=tripped)
+        const colorMap = { ok: 'var(--accent)', warning: 'var(--warn)', tripped: 'var(--red)', unknown: 'var(--border)' };
+        bar.style.background = colorMap[status] || 'var(--border)';
+        if (current !== null && current !== undefined && threshold < 0) {
+          const fillPct = Math.min(Math.max((current / threshold) * 100, 0), 100);
+          bar.style.width = `${fillPct}%`;
+        } else {
+          bar.style.width = '0%';
+        }
+      }
+
+      async function fetchStatus() {
+        try {
+          const r = await fetch('/api/status');
+          if (!r.ok) return;
+          const d = await r.json();
+          const t = d.totals || {};
+          document.getElementById('card-trades').textContent = t.trade_count ?? '--';
+          const pnlEl = document.getElementById('card-pnl');
+          const pnlV = t.total_pnl;
+          if (pnlV !== undefined) {
+            const n = parseFloat(pnlV);
+            const sign = n > 0 ? '+' : '';
+            pnlEl.textContent = `${sign}${n.toFixed(2)}`;
+            pnlEl.className = `card-value ${n > 0 ? 'positive' : n < 0 ? 'negative' : 'neutral'}`;
+          }
+          document.getElementById('card-pnl-sub').textContent = `결정 ${t.decision_count ?? 0}건`;
+          renderCbGauge(d.circuit_breaker);
+          renderModeBadge(d.mode);
+        } catch {}
+      }
+
+      function renderModeBadge(mode) {
+        const el = document.getElementById('mode-badge');
+        if (!el) return;
+        if (mode === 'live') {
+          el.textContent = '🔴 실전투자';
+          el.className = 'mode-badge live';
+        } else {
+          el.textContent = '🟡 모의투자';
+          el.className = 'mode-badge paper';
+        }
+      }
+
+      async function fetchPerformance() {
+        try {
+          const r = await fetch('/api/performance?market=all');
+          if (!r.ok) return;
+          const d = await r.json();
+          const c = d.combined || {};
+          document.getElementById('card-winrate').textContent = c.win_rate !== undefined ? `${c.win_rate}%` : '--';
+          document.getElementById('card-total').textContent = c.total_trades ?? '--';
+        } catch {}
+      }
+
+      async function fetchPnlHistory(days) {
+        try {
+          const r = await fetch(`/api/pnl/history?days=${days}`);
+          if (!r.ok) throw new Error('fetch failed');
+          const d = await r.json();
+          renderChart(d);
+        } catch {
+          document.getElementById('chart-error').style.display = 'block';
+        }
+      }
+
+      function renderChart(data) {
+        const errEl = document.getElementById('chart-error');
+        if (!data.labels || data.labels.length === 0) {
+          errEl.style.display = 'block';
+          return;
+        }
+        errEl.style.display = 'none';
+
+        const colors = data.pnl.map(v => v >= 0 ? 'rgba(60,179,113,0.75)' : 'rgba(224,85,85,0.75)');
+        const borderColors = data.pnl.map(v => v >= 0 ? '#3cb371' : '#e05555');
+
+        if (pnlChart) { pnlChart.destroy(); pnlChart = null; }
+        const ctx = document.getElementById('pnl-chart').getContext('2d');
+        pnlChart = new Chart(ctx, {
+          type: 'bar',
+          data: {
+            labels: data.labels,
+            datasets: [{
+              label: 'Daily P&L',
+              data: data.pnl,
+              backgroundColor: colors,
+              borderColor: borderColors,
+              borderWidth: 1,
+              borderRadius: 3,
+            }]
+          },
+          options: {
+            responsive: true,
+            maintainAspectRatio: false,
+            plugins: {
+              legend: { display: false },
+              tooltip: {
+                callbacks: {
+                  label: ctx => {
+                    const v = ctx.parsed.y;
+                    const sign = v >= 0 ? '+' : '';
+                    const trades = data.trades[ctx.dataIndex];
+                    return [`P&L: ${sign}${v.toFixed(2)}`, `거래: ${trades}건`];
+                  }
+                }
+              }
+            },
+            scales: {
+              x: {
+                ticks: { color: '#9fb3c8', font: { size: 10 }, maxRotation: 0 },
+                grid: { color: 'rgba(40,69,95,0.4)' }
+              },
+              y: {
+                ticks: { color: '#9fb3c8', font: { size: 10 } },
+                grid: { color: 'rgba(40,69,95,0.4)' }
+              }
+            }
+          }
+        });
+      }
+
+      async function fetchDecisions(market) {
+        const tbody = document.getElementById('decisions-body');
+        tbody.innerHTML = '<tr class="empty-row"><td colspan="5"><span class="spinner"></span></td></tr>';
+        try {
+          const r = await fetch(`/api/decisions?market=${market}&limit=50`);
+          if (!r.ok) throw new Error('fetch failed');
+          const d = await r.json();
+          if (!d.decisions || d.decisions.length === 0) {
+            tbody.innerHTML = '<tr class="empty-row"><td colspan="5">결정 로그 없음</td></tr>';
+            return;
+          }
+          tbody.innerHTML = d.decisions.map(dec => `
+            <tr>
+              <td>${fmt(dec.timestamp)}</td>
+              <td>${dec.stock_code || '--'}</td>
+              <td>${badge(dec.action)}</td>
+              <td>${confBar(dec.confidence)}</td>
+              <td class="rationale-cell" title="${(dec.rationale || '').replace(/"/g, '&quot;')}">${dec.rationale || '--'}</td>
+            </tr>
+          `).join('');
+        } catch {
+          tbody.innerHTML = '<tr class="empty-row"><td colspan="5">데이터 로드 실패</td></tr>';
+        }
+      }
+
+      function selectDays(btn) {
+        document.querySelectorAll('.day-btn').forEach(b => b.classList.remove('active'));
+        btn.classList.add('active');
+        currentDays = parseInt(btn.dataset.days, 10);
+        fetchPnlHistory(currentDays);
+      }
+
+      function selectMarket(btn) {
+        document.querySelectorAll('.tab-btn').forEach(b => b.classList.remove('active'));
+        btn.classList.add('active');
+        currentMarket = btn.dataset.market;
+        fetchDecisions(currentMarket);
+      }
+
+      function todayStr() {
+        return new Date().toISOString().slice(0, 10);
+      }
+
+      function esc(s) {
+        return String(s ?? '').replace(/&/g, '&amp;').replace(/</g, '&lt;').replace(/>/g, '&gt;').replace(/"/g, '&quot;');
+      }
+
+      async function fetchJSON(url) {
+        const r = await fetch(url);
+        if (!r.ok) throw new Error(`HTTP ${r.status}`);
+        return r.json();
+      }
+
+      async function fetchPlaybook() {
+        const market = document.getElementById('pb-market-select').value;
+        const date = todayStr();
+        document.getElementById('pb-date').textContent = date;
+        const el = document.getElementById('playbook-content');
+        try {
+          const data = await fetchJSON(`/api/playbook/${date}?market=${market}`);
+          const stocks = data.stock_playbooks ?? [];
+          if (stocks.length === 0) {
+            el.innerHTML = '<p class="empty-msg">오늘 플레이북 없음</p>';
+            return;
+          }
+          el.innerHTML = stocks.map(sp =>
+            `<details><summary>${esc(sp.stock_code ?? '?')} — ${esc(sp.signal ?? '')}</summary>` +
+            `<pre>${esc(JSON.stringify(sp, null, 2))}</pre></details>`
+          ).join('');
+        } catch {
+          el.innerHTML = '<p class="empty-msg">플레이북 없음 (오늘 미생성 또는 API 오류)</p>';
+        }
+      }
+
+      async function fetchScorecard() {
+        const market = document.getElementById('sc-market-select').value;
+        const date = todayStr();
+        document.getElementById('sc-date').textContent = date;
+        const el = document.getElementById('scorecard-grid');
+        try {
+          const data = await fetchJSON(`/api/scorecard/${date}?market=${market}`);
+          const sc = data.scorecard ?? {};
+          const entries = Object.entries(sc);
+          if (entries.length === 0) {
+            el.innerHTML = '<p class="empty-msg">스코어카드 없음</p>';
+            return;
+          }
+          el.className = 'scorecard-grid';
+          el.innerHTML = entries.map(([k, v]) => `
+            <div class="kpi-card">
+              <div class="kpi-label">${esc(k)}</div>
+              <div class="kpi-value">${typeof v === 'number' ? v.toFixed(2) : esc(String(v))}</div>
+            </div>`).join('');
+        } catch {
+          el.innerHTML = '<p class="empty-msg">스코어카드 없음 (오늘 미생성 또는 API 오류)</p>';
+        }
+      }
+
+      async function fetchScenarios() {
+        const market = document.getElementById('scen-market-select').value;
+        const date = todayStr();
+        const el = document.getElementById('scenarios-content');
+        try {
+          const data = await fetchJSON(`/api/scenarios/active?market=${market}&date_str=${date}&limit=50`);
+          const matches = data.matches ?? [];
+          if (matches.length === 0) {
+            el.innerHTML = '<p class="empty-msg">활성 시나리오 없음</p>';
+            return;
+          }
+          el.innerHTML = `<table class="scenarios-table">
+            <thead><tr><th>종목</th><th>신호</th><th>신뢰도</th><th>매칭 조건</th></tr></thead>
+            <tbody>${matches.map(m => `
+              <tr>
+                <td>${esc(m.stock_code)}</td>
+                <td>${esc(m.signal ?? '-')}</td>
+                <td>${esc(m.confidence ?? '-')}</td>
+                <td><code style="font-size:11px">${esc(JSON.stringify(m.scenario_match ?? {}))}</code></td>
+              </tr>`).join('')}
+            </tbody></table>`;
+        } catch {
+          el.innerHTML = '<p class="empty-msg">데이터 없음</p>';
+        }
+      }
+
+      async function fetchContext() {
+        const layer = document.getElementById('ctx-layer-select').value;
+        const limit = Math.min(Math.max(parseInt(document.getElementById('ctx-limit').value, 10) || 20, 1), 200);
+        const el = document.getElementById('context-content');
+        try {
+          const data = await fetchJSON(`/api/context/${layer}?limit=${limit}`);
+          const entries = data.entries ?? [];
+          if (entries.length === 0) {
+            el.innerHTML = '<p class="empty-msg">컨텍스트 없음</p>';
+            return;
+          }
+          el.innerHTML = `<table class="context-table">
+            <thead><tr><th>timeframe</th><th>key</th><th>value</th><th>updated</th></tr></thead>
+            <tbody>${entries.map(e => `
+              <tr>
+                <td>${esc(e.timeframe)}</td>
+                <td>${esc(e.key)}</td>
+                <td><div class="context-value">${esc(JSON.stringify(e.value ?? e.raw_value))}</div></td>
+                <td style="font-size:11px;color:var(--muted)">${esc((e.updated_at ?? '').slice(0, 16))}</td>
+              </tr>`).join('')}
+            </tbody></table>`;
+        } catch {
+          el.innerHTML = '<p class="empty-msg">데이터 없음</p>';
+        }
+      }
+
+      async function refreshAll() {
+        document.getElementById('last-updated').textContent = '업데이트 중...';
+        await Promise.all([
+          fetchStatus(),
+          fetchPerformance(),
+          fetchPositions(),
+          fetchPnlHistory(currentDays),
+          fetchDecisions(currentMarket),
+          fetchPlaybook(),
+          fetchScorecard(),
+          fetchScenarios(),
+          fetchContext(),
+        ]);
+        const now = new Date();
+        const timeStr = now.toLocaleTimeString('ko-KR', { hour: '2-digit', minute: '2-digit', second: '2-digit', hour12: false });
+        document.getElementById('last-updated').textContent = `마지막 업데이트: ${timeStr}`;
+      }
+
+      // Initial load
+      refreshAll();
+
+      // Auto-refresh every 30 seconds
+      setInterval(refreshAll, 30000);
+    </script>
+  </body>
+</html>

src/data/README.md

@@ -0,0 +1,205 @@
+# External Data Integration
+
+This module provides objective external data sources to enhance trading decisions beyond just market prices and user input.
+
+## Modules
+
+### `news_api.py` - News Sentiment Analysis
+
+Fetches real-time news for stocks with sentiment scoring.
+
+**Features:**
+- Alpha Vantage and NewsAPI.org support
+- Sentiment scoring (-1.0 to +1.0)
+- 5-minute caching to minimize API quota usage
+- Graceful fallback when API unavailable
+
+**Usage:**
+```python
+from src.data.news_api import NewsAPI
+
+# Initialize with API key
+news_api = NewsAPI(api_key="your_key", provider="alphavantage")
+
+# Fetch news sentiment
+sentiment = await news_api.get_news_sentiment("AAPL")
+if sentiment:
+    print(f"Average sentiment: {sentiment.avg_sentiment}")
+    for article in sentiment.articles[:3]:
+        print(f"{article.title} ({article.sentiment_score})")
+```
+
+### `economic_calendar.py` - Major Economic Events
+
+Tracks FOMC meetings, GDP releases, CPI, earnings calendars, and other market-moving events.
+
+**Features:**
+- High-impact event tracking (FOMC, GDP, CPI)
+- Earnings calendar per stock
+- Event proximity checking
+- Hardcoded major events for 2026 (no API required)
+
+**Usage:**
+```python
+from src.data.economic_calendar import EconomicCalendar
+
+calendar = EconomicCalendar()
+calendar.load_hardcoded_events()
+
+# Get upcoming high-impact events
+upcoming = calendar.get_upcoming_events(days_ahead=7, min_impact="HIGH")
+print(f"High-impact events: {upcoming.high_impact_count}")
+
+# Check if near earnings
+earnings_date = calendar.get_earnings_date("AAPL")
+if earnings_date:
+    print(f"Next earnings: {earnings_date}")
+
+# Check for high volatility period
+if calendar.is_high_volatility_period(hours_ahead=24):
+    print("High-impact event imminent!")
+```
+
+### `market_data.py` - Market Indicators
+
+Provides market breadth, sector performance, and sentiment indicators.
+
+**Features:**
+- Market sentiment levels (Fear & Greed equivalent)
+- Market breadth (advancing/declining stocks)
+- Sector performance tracking
+- Fear/Greed score calculation
+
+**Usage:**
+```python
+from src.data.market_data import MarketData
+
+market_data = MarketData(api_key="your_key")
+
+# Get market sentiment
+sentiment = market_data.get_market_sentiment()
+print(f"Market sentiment: {sentiment.name}")
+
+# Get full indicators
+indicators = market_data.get_market_indicators("US")
+print(f"Sentiment: {indicators.sentiment.name}")
+print(f"A/D Ratio: {indicators.breadth.advance_decline_ratio}")
+```
+
+## Integration with GeminiClient
+
+The external data sources are seamlessly integrated into the AI decision engine:
+
+```python
+from src.brain.gemini_client import GeminiClient
+from src.data.news_api import NewsAPI
+from src.data.economic_calendar import EconomicCalendar
+from src.data.market_data import MarketData
+from src.config import Settings
+
+settings = Settings()
+
+# Initialize data sources
+news_api = NewsAPI(api_key=settings.NEWS_API_KEY, provider=settings.NEWS_API_PROVIDER)
+calendar = EconomicCalendar()
+calendar.load_hardcoded_events()
+market_data = MarketData(api_key=settings.MARKET_DATA_API_KEY)
+
+# Create enhanced client
+client = GeminiClient(
+    settings,
+    news_api=news_api,
+    economic_calendar=calendar,
+    market_data=market_data
+)
+
+# Make decision with external context
+market_data_dict = {
+    "stock_code": "AAPL",
+    "current_price": 180.0,
+    "market_name": "US stock market"
+}
+
+decision = await client.decide(market_data_dict)
+```
+
+The external data is automatically included in the prompt sent to Gemini:
+
+```
+Market: US stock market
+Stock Code: AAPL
+Current Price: 180.0
+
+EXTERNAL DATA:
+News Sentiment: 0.85 (from 10 articles)
+  1. [Reuters] Apple hits record high (sentiment: 0.92)
+  2. [Bloomberg] Strong iPhone sales (sentiment: 0.78)
+  3. [CNBC] Tech sector rallying (sentiment: 0.85)
+
+Upcoming High-Impact Events: 2 in next 7 days
+  Next: FOMC Meeting (FOMC) on 2026-03-18
+  Earnings: AAPL on 2026-02-10
+
+Market Sentiment: GREED
+Advance/Decline Ratio: 2.35
+```
+
+## Configuration
+
+Add these to your `.env` file:
+
+```bash
+# External Data APIs (optional)
+NEWS_API_KEY=your_alpha_vantage_key
+NEWS_API_PROVIDER=alphavantage  # or "newsapi"
+MARKET_DATA_API_KEY=your_market_data_key
+```
+
+## API Recommendations
+
+### Alpha Vantage (News)
+- **Free tier:** 5 calls/min, 500 calls/day
+- **Pros:** Provides sentiment scores, no credit card required
+- **URL:** https://www.alphavantage.co/
+
+### NewsAPI.org
+- **Free tier:** 100 requests/day
+- **Pros:** Large news coverage, easy to use
+- **Cons:** No sentiment scores (we use keyword heuristics)
+- **URL:** https://newsapi.org/
+
+## Caching Strategy
+
+To minimize API quota usage:
+
+1. **News:** 5-minute TTL cache per stock
+2. **Economic Calendar:** Loaded once at startup (hardcoded events)
+3. **Market Data:** Fetched per decision (lightweight)
+
+## Graceful Degradation
+
+The system works gracefully without external data:
+
+- If no API keys provided → decisions work with just market prices
+- If API fails → decision continues without external context
+- If cache expired → attempts refetch, falls back to no data
+- Errors are logged but never block trading decisions
+
+## Testing
+
+All modules have comprehensive test coverage (81%+):
+
+```bash
+pytest tests/test_data_integration.py -v --cov=src/data
+```
+
+Tests use mocks to avoid requiring real API keys.
+
+## Future Enhancements
+
+- Twitter/X sentiment analysis
+- Reddit WallStreetBets sentiment
+- Options flow data
+- Insider trading activity
+- Analyst upgrades/downgrades
+- Real-time economic data APIs

src/data/__init__.py

@@ -0,0 +1,5 @@
+"""External data integration for objective decision-making."""
+
+from __future__ import annotations
+
+__all__ = ["NewsAPI", "EconomicCalendar", "MarketData"]

src/data/economic_calendar.py

@@ -0,0 +1,219 @@
+"""Economic calendar integration for major market events.
+
+Tracks FOMC meetings, GDP releases, CPI, earnings calendars, and other
+market-moving events.
+"""
+
+from __future__ import annotations
+
+import logging
+from dataclasses import dataclass
+from datetime import datetime, timedelta
+from typing import Any
+
+logger = logging.getLogger(__name__)
+
+
+@dataclass
+class EconomicEvent:
+    """Single economic event."""
+
+    name: str
+    event_type: str  # "FOMC", "GDP", "CPI", "EARNINGS", etc.
+    datetime: datetime
+    impact: str  # "HIGH", "MEDIUM", "LOW"
+    country: str
+    description: str
+
+
+@dataclass
+class UpcomingEvents:
+    """Collection of upcoming economic events."""
+
+    events: list[EconomicEvent]
+    high_impact_count: int
+    next_major_event: EconomicEvent | None
+
+
+class EconomicCalendar:
+    """Economic calendar with event tracking and impact scoring."""
+
+    def __init__(self, api_key: str | None = None) -> None:
+        """Initialize economic calendar.
+
+        Args:
+            api_key: API key for calendar provider (None for testing/hardcoded)
+        """
+        self._api_key = api_key
+        # For now, use hardcoded major events (can be extended with API)
+        self._events: list[EconomicEvent] = []
+
+    # ------------------------------------------------------------------
+    # Public API
+    # ------------------------------------------------------------------
+
+    def get_upcoming_events(
+        self, days_ahead: int = 7, min_impact: str = "MEDIUM"
+    ) -> UpcomingEvents:
+        """Get upcoming economic events within specified timeframe.
+
+        Args:
+            days_ahead: Number of days to look ahead
+            min_impact: Minimum impact level ("LOW", "MEDIUM", "HIGH")
+
+        Returns:
+            UpcomingEvents with filtered events
+        """
+        now = datetime.now()
+        end_date = now + timedelta(days=days_ahead)
+
+        # Filter events by timeframe and impact
+        upcoming = [
+            event
+            for event in self._events
+            if now <= event.datetime <= end_date
+            and self._impact_level(event.impact) >= self._impact_level(min_impact)
+        ]
+
+        # Sort by datetime
+        upcoming.sort(key=lambda e: e.datetime)
+
+        # Count high-impact events
+        high_impact_count = sum(1 for e in upcoming if e.impact == "HIGH")
+
+        # Get next major event
+        next_major = None
+        for event in upcoming:
+            if event.impact == "HIGH":
+                next_major = event
+                break
+
+        return UpcomingEvents(
+            events=upcoming,
+            high_impact_count=high_impact_count,
+            next_major_event=next_major,
+        )
+
+    def add_event(self, event: EconomicEvent) -> None:
+        """Add an economic event to the calendar."""
+        self._events.append(event)
+
+    def clear_events(self) -> None:
+        """Clear all events (useful for testing)."""
+        self._events.clear()
+
+    def get_earnings_date(self, stock_code: str) -> datetime | None:
+        """Get next earnings date for a stock.
+
+        Args:
+            stock_code: Stock ticker symbol
+
+        Returns:
+            Next earnings datetime or None if not found
+        """
+        now = datetime.now()
+        earnings_events = [
+            event
+            for event in self._events
+            if event.event_type == "EARNINGS"
+            and stock_code.upper() in event.name.upper()
+            and event.datetime > now
+        ]
+
+        if not earnings_events:
+            return None
+
+        # Return earliest upcoming earnings
+        earnings_events.sort(key=lambda e: e.datetime)
+        return earnings_events[0].datetime
+
+    def load_hardcoded_events(self) -> None:
+        """Load hardcoded major economic events for 2026.
+
+        This is a fallback when no API is available.
+        """
+        # Major FOMC meetings in 2026 (estimated)
+        fomc_dates = [
+            datetime(2026, 3, 18),
+            datetime(2026, 5, 6),
+            datetime(2026, 6, 17),
+            datetime(2026, 7, 29),
+            datetime(2026, 9, 16),
+            datetime(2026, 11, 4),
+            datetime(2026, 12, 16),
+        ]
+
+        for date in fomc_dates:
+            self.add_event(
+                EconomicEvent(
+                    name="FOMC Meeting",
+                    event_type="FOMC",
+                    datetime=date,
+                    impact="HIGH",
+                    country="US",
+                    description="Federal Reserve interest rate decision",
+                )
+            )
+
+        # Quarterly GDP releases (estimated)
+        gdp_dates = [
+            datetime(2026, 4, 28),
+            datetime(2026, 7, 30),
+            datetime(2026, 10, 29),
+        ]
+
+        for date in gdp_dates:
+            self.add_event(
+                EconomicEvent(
+                    name="US GDP Release",
+                    event_type="GDP",
+                    datetime=date,
+                    impact="HIGH",
+                    country="US",
+                    description="Quarterly GDP growth rate",
+                )
+            )
+
+        # Monthly CPI releases (12th of each month, estimated)
+        for month in range(1, 13):
+            try:
+                cpi_date = datetime(2026, month, 12)
+                self.add_event(
+                    EconomicEvent(
+                        name="US CPI Release",
+                        event_type="CPI",
+                        datetime=cpi_date,
+                        impact="HIGH",
+                        country="US",
+                        description="Consumer Price Index inflation data",
+                    )
+                )
+            except ValueError:
+                continue
+
+    # ------------------------------------------------------------------
+    # Helpers
+    # ------------------------------------------------------------------
+
+    def _impact_level(self, impact: str) -> int:
+        """Convert impact string to numeric level."""
+        levels = {"LOW": 1, "MEDIUM": 2, "HIGH": 3}
+        return levels.get(impact.upper(), 0)
+
+    def is_high_volatility_period(self, hours_ahead: int = 24) -> bool:
+        """Check if we're near a high-impact event.
+
+        Args:
+            hours_ahead: Number of hours to look ahead
+
+        Returns:
+            True if high-impact event is imminent
+        """
+        now = datetime.now()
+        threshold = now + timedelta(hours=hours_ahead)
+
+        for event in self._events:
+            if event.impact == "HIGH" and now <= event.datetime <= threshold:
+                return True
+
+        return False

src/data/market_data.py

@@ -0,0 +1,198 @@
+"""Additional market data indicators beyond basic price data.
+
+Provides market breadth, sector performance, and market sentiment indicators.
+"""
+
+from __future__ import annotations
+
+import logging
+from dataclasses import dataclass
+from enum import Enum
+
+logger = logging.getLogger(__name__)
+
+
+class MarketSentiment(Enum):
+    """Overall market sentiment levels."""
+
+    EXTREME_FEAR = 1
+    FEAR = 2
+    NEUTRAL = 3
+    GREED = 4
+    EXTREME_GREED = 5
+
+
+@dataclass
+class SectorPerformance:
+    """Performance metrics for a market sector."""
+
+    sector_name: str
+    daily_change_pct: float
+    weekly_change_pct: float
+    leader_stock: str  # Best performing stock in sector
+    laggard_stock: str  # Worst performing stock in sector
+
+
+@dataclass
+class MarketBreadth:
+    """Market breadth indicators."""
+
+    advancing_stocks: int
+    declining_stocks: int
+    unchanged_stocks: int
+    new_highs: int
+    new_lows: int
+    advance_decline_ratio: float
+
+
+@dataclass
+class MarketIndicators:
+    """Aggregated market indicators."""
+
+    sentiment: MarketSentiment
+    breadth: MarketBreadth
+    sector_performance: list[SectorPerformance]
+    vix_level: float | None  # Volatility index if available
+
+
+class MarketData:
+    """Market data provider for additional indicators."""
+
+    def __init__(self, api_key: str | None = None) -> None:
+        """Initialize market data provider.
+
+        Args:
+            api_key: API key for data provider (None for testing)
+        """
+        self._api_key = api_key
+
+    # ------------------------------------------------------------------
+    # Public API
+    # ------------------------------------------------------------------
+
+    def get_market_sentiment(self) -> MarketSentiment:
+        """Get current market sentiment level.
+
+        This is a simplified version. In production, this would integrate
+        with Fear & Greed Index or similar sentiment indicators.
+
+        Returns:
+            MarketSentiment enum value
+        """
+        # Default to neutral when API not available
+        if self._api_key is None:
+            logger.debug("No market data API key — returning NEUTRAL sentiment")
+            return MarketSentiment.NEUTRAL
+
+        # TODO: Integrate with actual sentiment API
+        return MarketSentiment.NEUTRAL
+
+    def get_market_breadth(self, market: str = "US") -> MarketBreadth | None:
+        """Get market breadth indicators.
+
+        Args:
+            market: Market code ("US", "KR", etc.)
+
+        Returns:
+            MarketBreadth object or None if unavailable
+        """
+        if self._api_key is None:
+            logger.debug("No market data API key — returning None for breadth")
+            return None
+
+        # TODO: Integrate with actual market breadth API
+        return None
+
+    def get_sector_performance(
+        self, market: str = "US"
+    ) -> list[SectorPerformance]:
+        """Get sector performance rankings.
+
+        Args:
+            market: Market code ("US", "KR", etc.)
+
+        Returns:
+            List of SectorPerformance objects, sorted by daily change
+        """
+        if self._api_key is None:
+            logger.debug("No market data API key — returning empty sector list")
+            return []
+
+        # TODO: Integrate with actual sector performance API
+        return []
+
+    def get_market_indicators(self, market: str = "US") -> MarketIndicators:
+        """Get aggregated market indicators.
+
+        Args:
+            market: Market code ("US", "KR", etc.)
+
+        Returns:
+            MarketIndicators with all available data
+        """
+        sentiment = self.get_market_sentiment()
+        breadth = self.get_market_breadth(market)
+        sectors = self.get_sector_performance(market)
+
+        # Default breadth if unavailable
+        if breadth is None:
+            breadth = MarketBreadth(
+                advancing_stocks=0,
+                declining_stocks=0,
+                unchanged_stocks=0,
+                new_highs=0,
+                new_lows=0,
+                advance_decline_ratio=1.0,
+            )
+
+        return MarketIndicators(
+            sentiment=sentiment,
+            breadth=breadth,
+            sector_performance=sectors,
+            vix_level=None,  # TODO: Add VIX integration
+        )
+
+    # ------------------------------------------------------------------
+    # Helper Methods
+    # ------------------------------------------------------------------
+
+    def calculate_fear_greed_score(
+        self, breadth: MarketBreadth, vix: float | None = None
+    ) -> int:
+        """Calculate a simple fear/greed score (0-100).
+
+        Args:
+            breadth: Market breadth data
+            vix: VIX level (optional)
+
+        Returns:
+            Score from 0 (extreme fear) to 100 (extreme greed)
+        """
+        # Start at neutral
+        score = 50
+
+        # Adjust based on advance/decline ratio
+        if breadth.advance_decline_ratio > 1.5:
+            score += 20
+        elif breadth.advance_decline_ratio > 1.0:
+            score += 10
+        elif breadth.advance_decline_ratio < 0.5:
+            score -= 20
+        elif breadth.advance_decline_ratio < 1.0:
+            score -= 10
+
+        # Adjust based on new highs/lows
+        if breadth.new_highs > breadth.new_lows * 2:
+            score += 15
+        elif breadth.new_lows > breadth.new_highs * 2:
+            score -= 15
+
+        # Adjust based on VIX if available
+        if vix is not None:
+            if vix > 30:  # High volatility = fear
+                score -= 15
+            elif vix < 15:  # Low volatility = complacency/greed
+                score += 10
+
+        # Clamp to 0-100
+        return max(0, min(100, score))

src/data/news_api.py

@@ -0,0 +1,316 @@
+"""News API integration with sentiment analysis and caching.
+
+Fetches real-time news for stocks using free-tier APIs (Alpha Vantage or NewsAPI).
+Includes 5-minute caching to minimize API quota usage.
+"""
+
+from __future__ import annotations
+
+import logging
+import time
+from dataclasses import dataclass
+from typing import Any
+
+import aiohttp
+
+logger = logging.getLogger(__name__)
+
+# Cache entries expire after 5 minutes
+CACHE_TTL_SECONDS = 300
+
+
+@dataclass
+class NewsArticle:
+    """Single news article with sentiment."""
+
+    title: str
+    summary: str
+    source: str
+    published_at: str
+    sentiment_score: float  # -1.0 (negative) to +1.0 (positive)
+    url: str
+
+
+@dataclass
+class NewsSentiment:
+    """Aggregated news sentiment for a stock."""
+
+    stock_code: str
+    articles: list[NewsArticle]
+    avg_sentiment: float  # Average sentiment across all articles
+    article_count: int
+    fetched_at: float  # Unix timestamp
+
+
+class NewsAPI:
+    """News API client with sentiment analysis and caching."""
+
+    def __init__(
+        self,
+        api_key: str | None = None,
+        provider: str = "alphavantage",
+        cache_ttl: int = CACHE_TTL_SECONDS,
+    ) -> None:
+        """Initialize NewsAPI client.
+
+        Args:
+            api_key: API key for the news provider (None for testing)
+            provider: News provider ("alphavantage" or "newsapi")
+            cache_ttl: Cache time-to-live in seconds
+        """
+        self._api_key = api_key
+        self._provider = provider
+        self._cache_ttl = cache_ttl
+        self._cache: dict[str, NewsSentiment] = {}
+
+    # ------------------------------------------------------------------
+    # Public API
+    # ------------------------------------------------------------------
+
+    async def get_news_sentiment(self, stock_code: str) -> NewsSentiment | None:
+        """Fetch news sentiment for a stock with caching.
+
+        Args:
+            stock_code: Stock ticker symbol (e.g., "AAPL", "005930")
+
+        Returns:
+            NewsSentiment object or None if fetch fails or API unavailable
+        """
+        # Check cache first
+        cached = self._get_from_cache(stock_code)
+        if cached is not None:
+            logger.debug("News cache hit for %s", stock_code)
+            return cached
+
+        # API key required for real requests
+        if self._api_key is None:
+            logger.warning("No news API key provided — returning None")
+            return None
+
+        # Fetch from API
+        try:
+            sentiment = await self._fetch_news(stock_code)
+            if sentiment is not None:
+                self._cache[stock_code] = sentiment
+            return sentiment
+        except Exception as exc:
+            logger.error("Failed to fetch news for %s: %s", stock_code, exc)
+            return None
+
+    def clear_cache(self) -> None:
+        """Clear the news cache (useful for testing)."""
+        self._cache.clear()
+
+    # ------------------------------------------------------------------
+    # Cache Management
+    # ------------------------------------------------------------------
+
+    def _get_from_cache(self, stock_code: str) -> NewsSentiment | None:
+        """Retrieve cached sentiment if not expired."""
+        if stock_code not in self._cache:
+            return None
+
+        cached = self._cache[stock_code]
+        age = time.time() - cached.fetched_at
+
+        if age > self._cache_ttl:
+            logger.debug("News cache expired for %s (age: %.1fs)", stock_code, age)
+            del self._cache[stock_code]
+            return None
+
+        return cached
+
+    # ------------------------------------------------------------------
+    # API Fetching
+    # ------------------------------------------------------------------
+
+    async def _fetch_news(self, stock_code: str) -> NewsSentiment | None:
+        """Fetch news from the provider API."""
+        if self._provider == "alphavantage":
+            return await self._fetch_alphavantage(stock_code)
+        elif self._provider == "newsapi":
+            return await self._fetch_newsapi(stock_code)
+        else:
+            logger.error("Unknown news provider: %s", self._provider)
+            return None
+
+    async def _fetch_alphavantage(self, stock_code: str) -> NewsSentiment | None:
+        """Fetch news from Alpha Vantage News Sentiment API."""
+        url = "https://www.alphavantage.co/query"
+        params = {
+            "function": "NEWS_SENTIMENT",
+            "tickers": stock_code,
+            "apikey": self._api_key,
+            "limit": 10,  # Fetch top 10 articles
+        }
+
+        try:
+            async with aiohttp.ClientSession() as session:
+                async with session.get(url, params=params, timeout=10) as resp:
+                    if resp.status != 200:
+                        logger.error(
+                            "Alpha Vantage API error: HTTP %d", resp.status
+                        )
+                        return None
+
+                    data = await resp.json()
+                    return self._parse_alphavantage_response(stock_code, data)
+
+        except Exception as exc:
+            logger.error("Alpha Vantage request failed: %s", exc)
+            return None
+
+    async def _fetch_newsapi(self, stock_code: str) -> NewsSentiment | None:
+        """Fetch news from NewsAPI.org."""
+        url = "https://newsapi.org/v2/everything"
+        params = {
+            "q": stock_code,
+            "apiKey": self._api_key,
+            "pageSize": 10,
+            "sortBy": "publishedAt",
+            "language": "en",
+        }
+
+        try:
+            async with aiohttp.ClientSession() as session:
+                async with session.get(url, params=params, timeout=10) as resp:
+                    if resp.status != 200:
+                        logger.error("NewsAPI error: HTTP %d", resp.status)
+                        return None
+
+                    data = await resp.json()
+                    return self._parse_newsapi_response(stock_code, data)
+
+        except Exception as exc:
+            logger.error("NewsAPI request failed: %s", exc)
+            return None
+
+    # ------------------------------------------------------------------
+    # Response Parsing
+    # ------------------------------------------------------------------
+
+    def _parse_alphavantage_response(
+        self, stock_code: str, data: dict[str, Any]
+    ) -> NewsSentiment | None:
+        """Parse Alpha Vantage API response."""
+        if "feed" not in data:
+            logger.warning("No 'feed' key in Alpha Vantage response")
+            return None
+
+        articles: list[NewsArticle] = []
+        for item in data["feed"]:
+            # Extract sentiment for this specific ticker
+            ticker_sentiment = self._extract_ticker_sentiment(item, stock_code)
+
+            article = NewsArticle(
+                title=item.get("title", ""),
+                summary=item.get("summary", "")[:200],  # Truncate long summaries
+                source=item.get("source", "Unknown"),
+                published_at=item.get("time_published", ""),
+                sentiment_score=ticker_sentiment,
+                url=item.get("url", ""),
+            )
+            articles.append(article)
+
+        if not articles:
+            return None
+
+        avg_sentiment = sum(a.sentiment_score for a in articles) / len(articles)
+
+        return NewsSentiment(
+            stock_code=stock_code,
+            articles=articles,
+            avg_sentiment=avg_sentiment,
+            article_count=len(articles),
+            fetched_at=time.time(),
+        )
+
+    def _extract_ticker_sentiment(
+        self, item: dict[str, Any], stock_code: str
+    ) -> float:
+        """Extract sentiment score for specific ticker from article."""
+        ticker_sentiments = item.get("ticker_sentiment", [])
+        for ts in ticker_sentiments:
+            if ts.get("ticker", "").upper() == stock_code.upper():
+                # Alpha Vantage provides sentiment_score as string
+                score_str = ts.get("ticker_sentiment_score", "0")
+                try:
+                    return float(score_str)
+                except ValueError:
+                    return 0.0
+
+        # Fallback to overall sentiment if ticker-specific not found
+        overall_sentiment = item.get("overall_sentiment_score", "0")
+        try:
+            return float(overall_sentiment)
+        except ValueError:
+            return 0.0
+
+    def _parse_newsapi_response(
+        self, stock_code: str, data: dict[str, Any]
+    ) -> NewsSentiment | None:
+        """Parse NewsAPI.org response.
+
+        Note: NewsAPI doesn't provide sentiment scores, so we use a
+        simple heuristic based on title keywords.
+        """
+        if data.get("status") != "ok" or "articles" not in data:
+            logger.warning("Invalid NewsAPI response")
+            return None
+
+        articles: list[NewsArticle] = []
+        for item in data["articles"]:
+            # Simple sentiment heuristic based on keywords
+            sentiment = self._estimate_sentiment_from_text(
+                item.get("title", "") + " " + item.get("description", "")
+            )
+
+            article = NewsArticle(
+                title=item.get("title", ""),
+                summary=item.get("description", "")[:200],
+                source=item.get("source", {}).get("name", "Unknown"),
+                published_at=item.get("publishedAt", ""),
+                sentiment_score=sentiment,
+                url=item.get("url", ""),
+            )
+            articles.append(article)
+
+        if not articles:
+            return None
+
+        avg_sentiment = sum(a.sentiment_score for a in articles) / len(articles)
+
+        return NewsSentiment(
+            stock_code=stock_code,
+            articles=articles,
+            avg_sentiment=avg_sentiment,
+            article_count=len(articles),
+            fetched_at=time.time(),
+        )
+
+    def _estimate_sentiment_from_text(self, text: str) -> float:
+        """Simple keyword-based sentiment estimation.
+
+        This is a fallback for APIs that don't provide sentiment scores.
+        Returns a score between -1.0 and +1.0.
+        """
+        text_lower = text.lower()
+
+        positive_keywords = [
+            "surge", "jump", "gain", "rise", "soar", "rally", "profit",
+            "growth", "upgrade", "beat", "strong", "bullish", "breakthrough",
+        ]
+        negative_keywords = [
+            "plunge", "fall", "drop", "decline", "crash", "loss", "weak",
+            "downgrade", "miss", "bearish", "concern", "risk", "warning",
+        ]
+
+        positive_count = sum(1 for kw in positive_keywords if kw in text_lower)
+        negative_count = sum(1 for kw in negative_keywords if kw in text_lower)
+
+        total = positive_count + negative_count
+        if total == 0:
+            return 0.0
+
+        # Normalize to -1.0 to +1.0 range
+        return (positive_count - negative_count) / total

src/db.py

@@ -2,9 +2,11 @@
 
 from __future__ import annotations
 
+import json
 import sqlite3
 from datetime import UTC, datetime
 from pathlib import Path
+from typing import Any
 
 
 def init_db(db_path: str) -> sqlite3.Connection:
@@ -12,6 +14,11 @@ def init_db(db_path: str) -> sqlite3.Connection:
     if db_path != ":memory:":
         Path(db_path).parent.mkdir(parents=True, exist_ok=True)
     conn = sqlite3.connect(db_path)
+    # Enable WAL mode for concurrent read/write (dashboard + trading loop).
+    # WAL does not apply to in-memory databases.
+    if db_path != ":memory:":
+        conn.execute("PRAGMA journal_mode=WAL")
+        conn.execute("PRAGMA busy_timeout=5000")
     conn.execute(
         """
         CREATE TABLE IF NOT EXISTS trades (
@@ -25,12 +32,14 @@ def init_db(db_path: str) -> sqlite3.Connection:
             price REAL,
             pnl REAL DEFAULT 0.0,
             market TEXT DEFAULT 'KR',
-            exchange_code TEXT DEFAULT 'KRX'
+            exchange_code TEXT DEFAULT 'KRX',
+            decision_id TEXT,
+            mode TEXT DEFAULT 'paper'
         )
         """
     )
 
-    # Migration: Add market and exchange_code columns if they don't exist
+    # Migration: Add columns if they don't exist (backward-compatible schema upgrades)
     cursor = conn.execute("PRAGMA table_info(trades)")
     columns = {row[1] for row in cursor.fetchall()}
 
@@ -38,6 +47,116 @@ def init_db(db_path: str) -> sqlite3.Connection:
         conn.execute("ALTER TABLE trades ADD COLUMN market TEXT DEFAULT 'KR'")
     if "exchange_code" not in columns:
         conn.execute("ALTER TABLE trades ADD COLUMN exchange_code TEXT DEFAULT 'KRX'")
+    if "selection_context" not in columns:
+        conn.execute("ALTER TABLE trades ADD COLUMN selection_context TEXT")
+    if "decision_id" not in columns:
+        conn.execute("ALTER TABLE trades ADD COLUMN decision_id TEXT")
+    if "mode" not in columns:
+        conn.execute("ALTER TABLE trades ADD COLUMN mode TEXT DEFAULT 'paper'")
+
+    # Context tree tables for multi-layered memory management
+    conn.execute(
+        """
+        CREATE TABLE IF NOT EXISTS contexts (
+            id INTEGER PRIMARY KEY AUTOINCREMENT,
+            layer TEXT NOT NULL,
+            timeframe TEXT NOT NULL,
+            key TEXT NOT NULL,
+            value TEXT NOT NULL,
+            created_at TEXT NOT NULL,
+            updated_at TEXT NOT NULL,
+            UNIQUE(layer, timeframe, key)
+        )
+        """
+    )
+
+    # Decision logging table for comprehensive audit trail
+    conn.execute(
+        """
+        CREATE TABLE IF NOT EXISTS decision_logs (
+            decision_id TEXT PRIMARY KEY,
+            timestamp TEXT NOT NULL,
+            stock_code TEXT NOT NULL,
+            market TEXT NOT NULL,
+            exchange_code TEXT NOT NULL,
+            action TEXT NOT NULL,
+            confidence INTEGER NOT NULL,
+            rationale TEXT NOT NULL,
+            context_snapshot TEXT NOT NULL,
+            input_data TEXT NOT NULL,
+            outcome_pnl REAL,
+            outcome_accuracy INTEGER,
+            reviewed INTEGER DEFAULT 0,
+            review_notes TEXT
+        )
+        """
+    )
+
+    conn.execute(
+        """
+        CREATE TABLE IF NOT EXISTS context_metadata (
+            layer TEXT PRIMARY KEY,
+            description TEXT NOT NULL,
+            retention_days INTEGER,
+            aggregation_source TEXT
+        )
+        """
+    )
+
+    # Playbook storage for pre-market strategy persistence
+    conn.execute(
+        """
+        CREATE TABLE IF NOT EXISTS playbooks (
+            id INTEGER PRIMARY KEY AUTOINCREMENT,
+            date TEXT NOT NULL,
+            market TEXT NOT NULL,
+            status TEXT NOT NULL DEFAULT 'pending',
+            playbook_json TEXT NOT NULL,
+            generated_at TEXT NOT NULL,
+            token_count INTEGER DEFAULT 0,
+            scenario_count INTEGER DEFAULT 0,
+            match_count INTEGER DEFAULT 0,
+            UNIQUE(date, market)
+        )
+        """
+    )
+
+    conn.execute("CREATE INDEX IF NOT EXISTS idx_playbooks_date ON playbooks(date)")
+    conn.execute("CREATE INDEX IF NOT EXISTS idx_playbooks_market ON playbooks(market)")
+
+    # Create indices for efficient context queries
+    conn.execute("CREATE INDEX IF NOT EXISTS idx_contexts_layer ON contexts(layer)")
+    conn.execute("CREATE INDEX IF NOT EXISTS idx_contexts_timeframe ON contexts(timeframe)")
+    conn.execute("CREATE INDEX IF NOT EXISTS idx_contexts_updated ON contexts(updated_at)")
+
+    # Create indices for efficient decision log queries
+    conn.execute(
+        "CREATE INDEX IF NOT EXISTS idx_decision_logs_timestamp ON decision_logs(timestamp)"
+    )
+    conn.execute(
+        "CREATE INDEX IF NOT EXISTS idx_decision_logs_reviewed ON decision_logs(reviewed)"
+    )
+    conn.execute(
+        "CREATE INDEX IF NOT EXISTS idx_decision_logs_confidence ON decision_logs(confidence)"
+    )
+
+    # Index for open-position queries (partition by stock_code, market, ordered by timestamp)
+    conn.execute(
+        "CREATE INDEX IF NOT EXISTS idx_trades_stock_market_ts"
+        " ON trades (stock_code, market, timestamp DESC)"
+    )
+
+    # Lightweight key-value store for trading system runtime metrics (dashboard use only)
+    # Intentionally separate from the AI context tree to preserve separation of concerns.
+    conn.execute(
+        """
+        CREATE TABLE IF NOT EXISTS system_metrics (
+            key        TEXT PRIMARY KEY,
+            value      TEXT NOT NULL,
+            updated_at TEXT NOT NULL
+        )
+        """
+    )
 
     conn.commit()
     return conn
@@ -54,15 +173,38 @@ def log_trade(
     pnl: float = 0.0,
     market: str = "KR",
     exchange_code: str = "KRX",
+    selection_context: dict[str, any] | None = None,
+    decision_id: str | None = None,
+    mode: str = "paper",
 ) -> None:
-    """Insert a trade record into the database."""
+    """Insert a trade record into the database.
+
+    Args:
+        conn: Database connection
+        stock_code: Stock code
+        action: Trade action (BUY/SELL/HOLD)
+        confidence: Confidence level (0-100)
+        rationale: AI decision rationale
+        quantity: Number of shares
+        price: Trade price
+        pnl: Profit/loss
+        market: Market code
+        exchange_code: Exchange code
+        selection_context: Scanner selection data (RSI, volume_ratio, signal, score)
+        decision_id: Unique decision identifier for audit linking
+        mode: Trading mode ('paper' or 'live') for data separation
+    """
+    # Serialize selection context to JSON
+    context_json = json.dumps(selection_context) if selection_context else None
+
     conn.execute(
         """
         INSERT INTO trades (
             timestamp, stock_code, action, confidence, rationale,
-            quantity, price, pnl, market, exchange_code
+            quantity, price, pnl, market, exchange_code, selection_context, decision_id,
+            mode
         )
-        VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?)
+        VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)
         """,
         (
             datetime.now(UTC).isoformat(),
@@ -75,6 +217,72 @@ def log_trade(
             pnl,
             market,
             exchange_code,
+            context_json,
+            decision_id,
+            mode,
         ),
     )
     conn.commit()
+
+
+def get_latest_buy_trade(
+    conn: sqlite3.Connection, stock_code: str, market: str
+) -> dict[str, Any] | None:
+    """Fetch the most recent BUY trade for a stock and market."""
+    cursor = conn.execute(
+        """
+        SELECT decision_id, price, quantity
+        FROM trades
+        WHERE stock_code = ?
+          AND market = ?
+          AND action = 'BUY'
+          AND decision_id IS NOT NULL
+        ORDER BY timestamp DESC
+        LIMIT 1
+        """,
+        (stock_code, market),
+    )
+    row = cursor.fetchone()
+    if not row:
+        return None
+    return {"decision_id": row[0], "price": row[1], "quantity": row[2]}
+
+
+def get_open_position(
+    conn: sqlite3.Connection, stock_code: str, market: str
+) -> dict[str, Any] | None:
+    """Return open position if latest trade is BUY, else None."""
+    cursor = conn.execute(
+        """
+        SELECT action, decision_id, price, quantity, timestamp
+        FROM trades
+        WHERE stock_code = ?
+          AND market = ?
+          AND action IN ('BUY', 'SELL')
+        ORDER BY timestamp DESC
+        LIMIT 1
+        """,
+        (stock_code, market),
+    )
+    row = cursor.fetchone()
+    if not row or row[0] != "BUY":
+        return None
+    return {"decision_id": row[1], "price": row[2], "quantity": row[3], "timestamp": row[4]}
+
+
+def get_recent_symbols(
+    conn: sqlite3.Connection, market: str, limit: int = 30
+) -> list[str]:
+    """Return recent unique symbols for a market, newest first."""
+    cursor = conn.execute(
+        """
+        SELECT stock_code, MAX(timestamp) AS last_ts
+        FROM trades
+        WHERE market = ?
+        GROUP BY stock_code
+        ORDER BY last_ts DESC
+        LIMIT ?
+        """,
+        (market, limit),
+    )
+    return [row[0] for row in cursor.fetchall() if row and row[0]]

src/evolution/__init__.py

@@ -0,0 +1,23 @@
+"""Evolution engine for self-improving trading strategies."""
+
+from src.evolution.ab_test import ABTester, ABTestResult, StrategyPerformance
+from src.evolution.daily_review import DailyReviewer
+from src.evolution.optimizer import EvolutionOptimizer
+from src.evolution.performance_tracker import (
+    PerformanceDashboard,
+    PerformanceTracker,
+    StrategyMetrics,
+)
+from src.evolution.scorecard import DailyScorecard
+
+__all__ = [
+    "EvolutionOptimizer",
+    "ABTester",
+    "ABTestResult",
+    "StrategyPerformance",
+    "PerformanceTracker",
+    "PerformanceDashboard",
+    "StrategyMetrics",
+    "DailyScorecard",
+    "DailyReviewer",
+]

src/evolution/ab_test.py

@@ -0,0 +1,220 @@
+"""A/B Testing framework for strategy comparison.
+
+Runs multiple strategies in parallel, tracks their performance,
+and uses statistical significance testing to determine winners.
+"""
+
+from __future__ import annotations
+
+import logging
+from dataclasses import dataclass
+from typing import Any
+
+import scipy.stats as stats
+
+logger = logging.getLogger(__name__)
+
+
+@dataclass
+class StrategyPerformance:
+    """Performance metrics for a single strategy."""
+
+    strategy_name: str
+    total_trades: int
+    wins: int
+    losses: int
+    total_pnl: float
+    avg_pnl: float
+    win_rate: float
+    sharpe_ratio: float | None = None
+
+
+@dataclass
+class ABTestResult:
+    """Result of an A/B test between two strategies."""
+
+    strategy_a: str
+    strategy_b: str
+    winner: str | None
+    p_value: float
+    confidence_level: float
+    is_significant: bool
+    performance_a: StrategyPerformance
+    performance_b: StrategyPerformance
+
+
+class ABTester:
+    """A/B testing framework for comparing trading strategies."""
+
+    def __init__(self, significance_level: float = 0.05) -> None:
+        """Initialize A/B tester.
+
+        Args:
+            significance_level: P-value threshold for statistical significance (default 0.05)
+        """
+        self._significance_level = significance_level
+
+    def calculate_performance(
+        self, trades: list[dict[str, Any]], strategy_name: str
+    ) -> StrategyPerformance:
+        """Calculate performance metrics for a strategy.
+
+        Args:
+            trades: List of trade records with pnl values
+            strategy_name: Name of the strategy
+
+        Returns:
+            StrategyPerformance object with calculated metrics
+        """
+        if not trades:
+            return StrategyPerformance(
+                strategy_name=strategy_name,
+                total_trades=0,
+                wins=0,
+                losses=0,
+                total_pnl=0.0,
+                avg_pnl=0.0,
+                win_rate=0.0,
+                sharpe_ratio=None,
+            )
+
+        total_trades = len(trades)
+        wins = sum(1 for t in trades if t.get("pnl", 0) > 0)
+        losses = sum(1 for t in trades if t.get("pnl", 0) < 0)
+        pnls = [t.get("pnl", 0.0) for t in trades]
+        total_pnl = sum(pnls)
+        avg_pnl = total_pnl / total_trades if total_trades > 0 else 0.0
+        win_rate = (wins / total_trades * 100) if total_trades > 0 else 0.0
+
+        # Calculate Sharpe ratio (risk-adjusted return)
+        sharpe_ratio = None
+        if len(pnls) > 1:
+            mean_return = avg_pnl
+            std_return = (
+                sum((p - mean_return) ** 2 for p in pnls) / (len(pnls) - 1)
+            ) ** 0.5
+            if std_return > 0:
+                sharpe_ratio = mean_return / std_return
+
+        return StrategyPerformance(
+            strategy_name=strategy_name,
+            total_trades=total_trades,
+            wins=wins,
+            losses=losses,
+            total_pnl=round(total_pnl, 2),
+            avg_pnl=round(avg_pnl, 2),
+            win_rate=round(win_rate, 2),
+            sharpe_ratio=round(sharpe_ratio, 4) if sharpe_ratio else None,
+        )
+
+    def compare_strategies(
+        self,
+        trades_a: list[dict[str, Any]],
+        trades_b: list[dict[str, Any]],
+        strategy_a_name: str = "Strategy A",
+        strategy_b_name: str = "Strategy B",
+    ) -> ABTestResult:
+        """Compare two strategies using statistical testing.
+
+        Uses a two-sample t-test to determine if performance difference is significant.
+
+        Args:
+            trades_a: List of trades from strategy A
+            trades_b: List of trades from strategy B
+            strategy_a_name: Name of strategy A
+            strategy_b_name: Name of strategy B
+
+        Returns:
+            ABTestResult with comparison details
+        """
+        perf_a = self.calculate_performance(trades_a, strategy_a_name)
+        perf_b = self.calculate_performance(trades_b, strategy_b_name)
+
+        # Extract PnL arrays for statistical testing
+        pnls_a = [t.get("pnl", 0.0) for t in trades_a]
+        pnls_b = [t.get("pnl", 0.0) for t in trades_b]
+
+        # Perform two-sample t-test
+        if len(pnls_a) > 1 and len(pnls_b) > 1:
+            t_stat, p_value = stats.ttest_ind(pnls_a, pnls_b, equal_var=False)
+            is_significant = p_value < self._significance_level
+            confidence_level = (1 - p_value) * 100
+        else:
+            # Not enough data for statistical test
+            p_value = 1.0
+            is_significant = False
+            confidence_level = 0.0
+
+        # Determine winner based on average PnL
+        winner = None
+        if is_significant:
+            if perf_a.avg_pnl > perf_b.avg_pnl:
+                winner = strategy_a_name
+            elif perf_b.avg_pnl > perf_a.avg_pnl:
+                winner = strategy_b_name
+
+        return ABTestResult(
+            strategy_a=strategy_a_name,
+            strategy_b=strategy_b_name,
+            winner=winner,
+            p_value=round(p_value, 4),
+            confidence_level=round(confidence_level, 2),
+            is_significant=is_significant,
+            performance_a=perf_a,
+            performance_b=perf_b,
+        )
+
+    def should_deploy(
+        self,
+        result: ABTestResult,
+        min_win_rate: float = 60.0,
+        min_trades: int = 20,
+    ) -> bool:
+        """Determine if a winning strategy should be deployed.
+
+        Args:
+            result: A/B test result
+            min_win_rate: Minimum win rate percentage for deployment (default 60%)
+            min_trades: Minimum number of trades required (default 20)
+
+        Returns:
+            True if the winning strategy meets deployment criteria
+        """
+        if not result.is_significant or result.winner is None:
+            return False
+
+        # Get performance of winning strategy
+        if result.winner == result.strategy_a:
+            winning_perf = result.performance_a
+        else:
+            winning_perf = result.performance_b
+
+        # Check deployment criteria
+        has_enough_trades = winning_perf.total_trades >= min_trades
+        has_good_win_rate = winning_perf.win_rate >= min_win_rate
+        is_profitable = winning_perf.avg_pnl > 0
+
+        meets_criteria = has_enough_trades and has_good_win_rate and is_profitable
+
+        if meets_criteria:
+            logger.info(
+                "Strategy '%s' meets deployment criteria: "
+                "win_rate=%.2f%%, trades=%d, avg_pnl=%.2f",
+                result.winner,
+                winning_perf.win_rate,
+                winning_perf.total_trades,
+                winning_perf.avg_pnl,
+            )
+        else:
+            logger.info(
+                "Strategy '%s' does NOT meet deployment criteria: "
+                "win_rate=%.2f%% (min %.2f%%), trades=%d (min %d), avg_pnl=%.2f",
+                result.winner if result.winner else "unknown",
+                winning_perf.win_rate if result.winner else 0.0,
+                min_win_rate,
+                winning_perf.total_trades if result.winner else 0,
+                min_trades,
+                winning_perf.avg_pnl if result.winner else 0.0,
+            )
+
+        return meets_criteria

src/evolution/daily_review.py

@@ -0,0 +1,196 @@
+"""Daily review generator for market-scoped end-of-day scorecards."""
+
+from __future__ import annotations
+
+import json
+import logging
+import re
+import sqlite3
+from dataclasses import asdict
+
+from src.brain.gemini_client import GeminiClient
+from src.context.layer import ContextLayer
+from src.context.store import ContextStore
+from src.evolution.scorecard import DailyScorecard
+
+logger = logging.getLogger(__name__)
+
+
+class DailyReviewer:
+    """Builds daily scorecards and optional AI-generated lessons."""
+
+    def __init__(
+        self,
+        conn: sqlite3.Connection,
+        context_store: ContextStore,
+        gemini_client: GeminiClient | None = None,
+    ) -> None:
+        self._conn = conn
+        self._context_store = context_store
+        self._gemini = gemini_client
+
+    def generate_scorecard(self, date: str, market: str) -> DailyScorecard:
+        """Generate a market-scoped scorecard from decision logs and trades."""
+        decision_rows = self._conn.execute(
+            """
+            SELECT action, confidence, context_snapshot
+            FROM decision_logs
+            WHERE DATE(timestamp) = ? AND market = ?
+            """,
+            (date, market),
+        ).fetchall()
+
+        total_decisions = len(decision_rows)
+        buys = sum(1 for row in decision_rows if row[0] == "BUY")
+        sells = sum(1 for row in decision_rows if row[0] == "SELL")
+        holds = sum(1 for row in decision_rows if row[0] == "HOLD")
+        avg_confidence = (
+            round(sum(int(row[1]) for row in decision_rows) / total_decisions, 2)
+            if total_decisions > 0
+            else 0.0
+        )
+
+        matched = 0
+        for row in decision_rows:
+            try:
+                snapshot = json.loads(row[2]) if row[2] else {}
+            except json.JSONDecodeError:
+                snapshot = {}
+            scenario_match = snapshot.get("scenario_match", {})
+            if isinstance(scenario_match, dict) and scenario_match:
+                matched += 1
+        scenario_match_rate = (
+            round((matched / total_decisions) * 100, 2)
+            if total_decisions
+            else 0.0
+        )
+
+        trade_stats = self._conn.execute(
+            """
+            SELECT
+                COALESCE(SUM(pnl), 0.0),
+                SUM(CASE WHEN pnl > 0 THEN 1 ELSE 0 END),
+                SUM(CASE WHEN pnl < 0 THEN 1 ELSE 0 END)
+            FROM trades
+            WHERE DATE(timestamp) = ? AND market = ?
+            """,
+            (date, market),
+        ).fetchone()
+        total_pnl = round(float(trade_stats[0] or 0.0), 2) if trade_stats else 0.0
+        wins = int(trade_stats[1] or 0) if trade_stats else 0
+        losses = int(trade_stats[2] or 0) if trade_stats else 0
+        win_rate = round((wins / (wins + losses)) * 100, 2) if (wins + losses) > 0 else 0.0
+
+        top_winners = [
+            row[0]
+            for row in self._conn.execute(
+                """
+                SELECT stock_code, SUM(pnl) AS stock_pnl
+                FROM trades
+                WHERE DATE(timestamp) = ? AND market = ?
+                GROUP BY stock_code
+                HAVING stock_pnl > 0
+                ORDER BY stock_pnl DESC
+                LIMIT 3
+                """,
+                (date, market),
+            ).fetchall()
+        ]
+
+        top_losers = [
+            row[0]
+            for row in self._conn.execute(
+                """
+                SELECT stock_code, SUM(pnl) AS stock_pnl
+                FROM trades
+                WHERE DATE(timestamp) = ? AND market = ?
+                GROUP BY stock_code
+                HAVING stock_pnl < 0
+                ORDER BY stock_pnl ASC
+                LIMIT 3
+                """,
+                (date, market),
+            ).fetchall()
+        ]
+
+        return DailyScorecard(
+            date=date,
+            market=market,
+            total_decisions=total_decisions,
+            buys=buys,
+            sells=sells,
+            holds=holds,
+            total_pnl=total_pnl,
+            win_rate=win_rate,
+            avg_confidence=avg_confidence,
+            scenario_match_rate=scenario_match_rate,
+            top_winners=top_winners,
+            top_losers=top_losers,
+            lessons=[],
+            cross_market_note="",
+        )
+
+    async def generate_lessons(self, scorecard: DailyScorecard) -> list[str]:
+        """Generate concise lessons from scorecard metrics using Gemini."""
+        if self._gemini is None:
+            return []
+
+        prompt = (
+            "You are a trading performance reviewer.\n"
+            "Return ONLY a JSON array of 1-3 short lessons in English.\n"
+            f"Market: {scorecard.market}\n"
+            f"Date: {scorecard.date}\n"
+            f"Total decisions: {scorecard.total_decisions}\n"
+            f"Buys/Sells/Holds: {scorecard.buys}/{scorecard.sells}/{scorecard.holds}\n"
+            f"Total PnL: {scorecard.total_pnl}\n"
+            f"Win rate: {scorecard.win_rate}%\n"
+            f"Average confidence: {scorecard.avg_confidence}\n"
+            f"Scenario match rate: {scorecard.scenario_match_rate}%\n"
+            f"Top winners: {', '.join(scorecard.top_winners) or 'N/A'}\n"
+            f"Top losers: {', '.join(scorecard.top_losers) or 'N/A'}\n"
+        )
+
+        try:
+            decision = await self._gemini.decide(
+                {
+                    "stock_code": "REVIEW",
+                    "market_name": scorecard.market,
+                    "current_price": 0,
+                    "prompt_override": prompt,
+                }
+            )
+            return self._parse_lessons(decision.rationale)
+        except Exception as exc:
+            logger.warning("Failed to generate daily lessons: %s", exc)
+            return []
+
+    def store_scorecard_in_context(self, scorecard: DailyScorecard) -> None:
+        """Store scorecard in L6 using market-scoped key."""
+        self._context_store.set_context(
+            ContextLayer.L6_DAILY,
+            scorecard.date,
+            f"scorecard_{scorecard.market}",
+            asdict(scorecard),
+        )
+
+    def _parse_lessons(self, raw_text: str) -> list[str]:
+        """Parse lessons from JSON array response or fallback text."""
+        raw_text = raw_text.strip()
+        try:
+            parsed = json.loads(raw_text)
+            if isinstance(parsed, list):
+                return [str(item).strip() for item in parsed if str(item).strip()][:3]
+        except json.JSONDecodeError:
+            pass
+
+        match = re.search(r"\[.*\]", raw_text, re.DOTALL)
+        if match:
+            try:
+                parsed = json.loads(match.group(0))
+                if isinstance(parsed, list):
+                    return [str(item).strip() for item in parsed if str(item).strip()][:3]
+            except json.JSONDecodeError:
+                pass
+
+        lines = [line.strip("-* \t") for line in raw_text.splitlines() if line.strip()]
+        return lines[:3]

src/evolution/optimizer.py

@@ -1,10 +1,10 @@
 """Evolution Engine — analyzes trade logs and generates new strategies.
 
 This module:
-1. Reads trade_logs.db to identify failing patterns
-2. Asks Gemini to generate a new strategy class
-3. Runs pytest on the generated file
-4. Creates a simulated PR if tests pass
+1. Uses DecisionLogger.get_losing_decisions() to identify failing patterns
+2. Analyzes failure patterns by time, market conditions, stock characteristics
+3. Asks Gemini to generate improved strategy recommendations
+4. Generates new strategy classes with enhanced decision-making logic
 """
 
 from __future__ import annotations
@@ -14,6 +14,7 @@ import logging
 import sqlite3
 import subprocess
 import textwrap
+from collections import Counter
 from datetime import UTC, datetime
 from pathlib import Path
 from typing import Any
@@ -21,6 +22,8 @@ from typing import Any
 from google import genai
 
 from src.config import Settings
+from src.db import init_db
+from src.logging.decision_logger import DecisionLogger
 
 logger = logging.getLogger(__name__)
 
@@ -53,29 +56,105 @@ class EvolutionOptimizer:
         self._db_path = settings.DB_PATH
         self._client = genai.Client(api_key=settings.GEMINI_API_KEY)
         self._model_name = settings.GEMINI_MODEL
+        self._conn = init_db(self._db_path)
+        self._decision_logger = DecisionLogger(self._conn)
 
     # ------------------------------------------------------------------
     # Analysis
     # ------------------------------------------------------------------
 
     def analyze_failures(self, limit: int = 50) -> list[dict[str, Any]]:
-        """Find trades where high confidence led to losses."""
-        conn = sqlite3.connect(self._db_path)
-        conn.row_factory = sqlite3.Row
-        try:
-            rows = conn.execute(
-                """
-                SELECT stock_code, action, confidence, pnl, rationale, timestamp
-                FROM trades
-                WHERE confidence >= 80 AND pnl < 0
-                ORDER BY pnl ASC
-                LIMIT ?
-                """,
-                (limit,),
-            ).fetchall()
-            return [dict(r) for r in rows]
-        finally:
-            conn.close()
+        """Find high-confidence decisions that resulted in losses.
+
+        Uses DecisionLogger.get_losing_decisions() to retrieve failures.
+        """
+        losing_decisions = self._decision_logger.get_losing_decisions(
+            min_confidence=80, min_loss=-100.0
+        )
+
+        # Limit results
+        if len(losing_decisions) > limit:
+            losing_decisions = losing_decisions[:limit]
+
+        # Convert to dict format for analysis
+        failures = []
+        for decision in losing_decisions:
+            failures.append({
+                "decision_id": decision.decision_id,
+                "timestamp": decision.timestamp,
+                "stock_code": decision.stock_code,
+                "market": decision.market,
+                "exchange_code": decision.exchange_code,
+                "action": decision.action,
+                "confidence": decision.confidence,
+                "rationale": decision.rationale,
+                "outcome_pnl": decision.outcome_pnl,
+                "outcome_accuracy": decision.outcome_accuracy,
+                "context_snapshot": decision.context_snapshot,
+                "input_data": decision.input_data,
+            })
+
+        return failures
+
+    def identify_failure_patterns(
+        self, failures: list[dict[str, Any]]
+    ) -> dict[str, Any]:
+        """Identify patterns in losing decisions.
+
+        Analyzes:
+        - Time patterns (hour of day, day of week)
+        - Market conditions (volatility, volume)
+        - Stock characteristics (price range, market)
+        - Common failure modes in rationale
+        """
+        if not failures:
+            return {"pattern_count": 0, "patterns": {}}
+
+        patterns = {
+            "markets": Counter(),
+            "actions": Counter(),
+            "hours": Counter(),
+            "avg_confidence": 0.0,
+            "avg_loss": 0.0,
+            "total_failures": len(failures),
+        }
+
+        total_confidence = 0
+        total_loss = 0.0
+
+        for failure in failures:
+            # Market distribution
+            patterns["markets"][failure.get("market", "UNKNOWN")] += 1
+
+            # Action distribution
+            patterns["actions"][failure.get("action", "UNKNOWN")] += 1
+
+            # Time pattern (extract hour from ISO timestamp)
+            timestamp = failure.get("timestamp", "")
+            if timestamp:
+                try:
+                    dt = datetime.fromisoformat(timestamp)
+                    patterns["hours"][dt.hour] += 1
+                except (ValueError, AttributeError):
+                    pass
+
+            # Aggregate metrics
+            total_confidence += failure.get("confidence", 0)
+            total_loss += failure.get("outcome_pnl", 0.0)
+
+        patterns["avg_confidence"] = (
+            round(total_confidence / len(failures), 2) if failures else 0.0
+        )
+        patterns["avg_loss"] = (
+            round(total_loss / len(failures), 2) if failures else 0.0
+        )
+
+        # Convert Counters to regular dicts for JSON serialization
+        patterns["markets"] = dict(patterns["markets"])
+        patterns["actions"] = dict(patterns["actions"])
+        patterns["hours"] = dict(patterns["hours"])
+
+        return patterns
 
     def get_performance_summary(self) -> dict[str, Any]:
         """Return aggregate performance metrics from trade logs."""
@@ -109,14 +188,25 @@ class EvolutionOptimizer:
     async def generate_strategy(self, failures: list[dict[str, Any]]) -> Path | None:
         """Ask Gemini to generate a new strategy based on failure analysis.
 
+        Integrates failure patterns and market conditions to create improved strategies.
         Returns the path to the generated strategy file, or None on failure.
         """
+        # Identify failure patterns first
+        patterns = self.identify_failure_patterns(failures)
+
         prompt = (
             "You are a quantitative trading strategy developer.\n"
-            "Analyze these failed trades and generate an improved strategy.\n\n"
-            f"Failed trades:\n{json.dumps(failures, indent=2, default=str)}\n\n"
-            "Generate a Python class that inherits from BaseStrategy.\n"
-            "The class must have an `evaluate(self, market_data: dict) -> dict` method.\n"
+            "Analyze these failed trades and their patterns, then generate an improved strategy.\n\n"
+            f"Failure Patterns:\n{json.dumps(patterns, indent=2)}\n\n"
+            f"Sample Failed Trades (first 5):\n"
+            f"{json.dumps(failures[:5], indent=2, default=str)}\n\n"
+            "Based on these patterns, generate an improved trading strategy.\n"
+            "The strategy should:\n"
+            "1. Avoid the identified failure patterns\n"
+            "2. Consider market-specific conditions\n"
+            "3. Adjust confidence based on historical performance\n\n"
+            "Generate a Python method body that inherits from BaseStrategy.\n"
+            "The method signature is: evaluate(self, market_data: dict) -> dict\n"
             "The method must return a dict with keys: action, confidence, rationale.\n"
             "Respond with ONLY the method body (Python code), no class definition.\n"
         )
@@ -147,10 +237,15 @@ class EvolutionOptimizer:
         # Indent the body for the class method
         indented_body = textwrap.indent(body, "            ")
 
+        # Generate rationale from patterns
+        rationale = f"Auto-evolved from {len(failures)} failures. "
+        rationale += f"Primary failure markets: {list(patterns.get('markets', {}).keys())}. "
+        rationale += f"Average loss: {patterns.get('avg_loss', 0.0)}"
+
         content = STRATEGY_TEMPLATE.format(
             name=version,
             timestamp=datetime.now(UTC).isoformat(),
-            rationale="Auto-evolved from failure analysis",
+            rationale=rationale,
             class_name=class_name,
             body=indented_body.strip(),
         )

src/evolution/performance_tracker.py

@@ -0,0 +1,303 @@
+"""Performance tracking system for strategy monitoring.
+
+Tracks win rates, monitors improvement over time,
+and provides performance metrics dashboard.
+"""
+
+from __future__ import annotations
+
+import json
+import logging
+import sqlite3
+from dataclasses import asdict, dataclass
+from datetime import UTC, datetime, timedelta
+from typing import Any
+
+logger = logging.getLogger(__name__)
+
+
+@dataclass
+class StrategyMetrics:
+    """Performance metrics for a strategy over a time period."""
+
+    strategy_name: str
+    period_start: str
+    period_end: str
+    total_trades: int
+    wins: int
+    losses: int
+    holds: int
+    win_rate: float
+    avg_pnl: float
+    total_pnl: float
+    best_trade: float
+    worst_trade: float
+    avg_confidence: float
+
+
+@dataclass
+class PerformanceDashboard:
+    """Comprehensive performance dashboard."""
+
+    generated_at: str
+    overall_metrics: StrategyMetrics
+    daily_metrics: list[StrategyMetrics]
+    weekly_metrics: list[StrategyMetrics]
+    improvement_trend: dict[str, Any]
+
+
+class PerformanceTracker:
+    """Tracks and monitors strategy performance over time."""
+
+    def __init__(self, db_path: str) -> None:
+        """Initialize performance tracker.
+
+        Args:
+            db_path: Path to the trade logs database
+        """
+        self._db_path = db_path
+
+    def get_strategy_metrics(
+        self,
+        strategy_name: str | None = None,
+        start_date: str | None = None,
+        end_date: str | None = None,
+    ) -> StrategyMetrics:
+        """Get performance metrics for a strategy over a time period.
+
+        Args:
+            strategy_name: Name of the strategy (None = all strategies)
+            start_date: Start date in ISO format (None = beginning of time)
+            end_date: End date in ISO format (None = now)
+
+        Returns:
+            StrategyMetrics object with performance data
+        """
+        conn = sqlite3.connect(self._db_path)
+        conn.row_factory = sqlite3.Row
+
+        try:
+            # Build query with optional filters
+            query = """
+                SELECT
+                    COUNT(*) as total_trades,
+                    SUM(CASE WHEN pnl > 0 THEN 1 ELSE 0 END) as wins,
+                    SUM(CASE WHEN pnl < 0 THEN 1 ELSE 0 END) as losses,
+                    SUM(CASE WHEN action = 'HOLD' THEN 1 ELSE 0 END) as holds,
+                    COALESCE(AVG(CASE WHEN pnl IS NOT NULL THEN pnl END), 0) as avg_pnl,
+                    COALESCE(SUM(CASE WHEN pnl IS NOT NULL THEN pnl ELSE 0 END), 0) as total_pnl,
+                    COALESCE(MAX(pnl), 0) as best_trade,
+                    COALESCE(MIN(pnl), 0) as worst_trade,
+                    COALESCE(AVG(confidence), 0) as avg_confidence,
+                    MIN(timestamp) as period_start,
+                    MAX(timestamp) as period_end
+                FROM trades
+                WHERE 1=1
+            """
+            params: list[Any] = []
+
+            if start_date:
+                query += " AND timestamp >= ?"
+                params.append(start_date)
+
+            if end_date:
+                query += " AND timestamp <= ?"
+                params.append(end_date)
+
+            # Note: Currently trades table doesn't have strategy_name column
+            # This is a placeholder for future extension
+
+            row = conn.execute(query, params).fetchone()
+
+            total_trades = row["total_trades"] or 0
+            wins = row["wins"] or 0
+            win_rate = (wins / total_trades * 100) if total_trades > 0 else 0.0
+
+            return StrategyMetrics(
+                strategy_name=strategy_name or "default",
+                period_start=row["period_start"] or "",
+                period_end=row["period_end"] or "",
+                total_trades=total_trades,
+                wins=wins,
+                losses=row["losses"] or 0,
+                holds=row["holds"] or 0,
+                win_rate=round(win_rate, 2),
+                avg_pnl=round(row["avg_pnl"], 2),
+                total_pnl=round(row["total_pnl"], 2),
+                best_trade=round(row["best_trade"], 2),
+                worst_trade=round(row["worst_trade"], 2),
+                avg_confidence=round(row["avg_confidence"], 2),
+            )
+        finally:
+            conn.close()
+
+    def get_daily_metrics(
+        self, days: int = 7, strategy_name: str | None = None
+    ) -> list[StrategyMetrics]:
+        """Get daily performance metrics for the last N days.
+
+        Args:
+            days: Number of days to retrieve (default 7)
+            strategy_name: Name of the strategy (None = all strategies)
+
+        Returns:
+            List of StrategyMetrics, one per day
+        """
+        metrics = []
+        end_date = datetime.now(UTC)
+
+        for i in range(days):
+            day_end = end_date - timedelta(days=i)
+            day_start = day_end - timedelta(days=1)
+
+            day_metrics = self.get_strategy_metrics(
+                strategy_name=strategy_name,
+                start_date=day_start.isoformat(),
+                end_date=day_end.isoformat(),
+            )
+            metrics.append(day_metrics)
+
+        return metrics
+
+    def get_weekly_metrics(
+        self, weeks: int = 4, strategy_name: str | None = None
+    ) -> list[StrategyMetrics]:
+        """Get weekly performance metrics for the last N weeks.
+
+        Args:
+            weeks: Number of weeks to retrieve (default 4)
+            strategy_name: Name of the strategy (None = all strategies)
+
+        Returns:
+            List of StrategyMetrics, one per week
+        """
+        metrics = []
+        end_date = datetime.now(UTC)
+
+        for i in range(weeks):
+            week_end = end_date - timedelta(weeks=i)
+            week_start = week_end - timedelta(weeks=1)
+
+            week_metrics = self.get_strategy_metrics(
+                strategy_name=strategy_name,
+                start_date=week_start.isoformat(),
+                end_date=week_end.isoformat(),
+            )
+            metrics.append(week_metrics)
+
+        return metrics
+
+    def calculate_improvement_trend(
+        self, metrics_history: list[StrategyMetrics]
+    ) -> dict[str, Any]:
+        """Calculate improvement trend from historical metrics.
+
+        Args:
+            metrics_history: List of StrategyMetrics ordered from oldest to newest
+
+        Returns:
+            Dictionary with trend analysis
+        """
+        if len(metrics_history) < 2:
+            return {
+                "trend": "insufficient_data",
+                "win_rate_change": 0.0,
+                "pnl_change": 0.0,
+                "confidence_change": 0.0,
+            }
+
+        oldest = metrics_history[0]
+        newest = metrics_history[-1]
+
+        win_rate_change = newest.win_rate - oldest.win_rate
+        pnl_change = newest.avg_pnl - oldest.avg_pnl
+        confidence_change = newest.avg_confidence - oldest.avg_confidence
+
+        # Determine overall trend
+        if win_rate_change > 5.0 and pnl_change > 0:
+            trend = "improving"
+        elif win_rate_change < -5.0 or pnl_change < 0:
+            trend = "declining"
+        else:
+            trend = "stable"
+
+        return {
+            "trend": trend,
+            "win_rate_change": round(win_rate_change, 2),
+            "pnl_change": round(pnl_change, 2),
+            "confidence_change": round(confidence_change, 2),
+            "period_count": len(metrics_history),
+        }
+
+    def generate_dashboard(
+        self, strategy_name: str | None = None
+    ) -> PerformanceDashboard:
+        """Generate a comprehensive performance dashboard.
+
+        Args:
+            strategy_name: Name of the strategy (None = all strategies)
+
+        Returns:
+            PerformanceDashboard with all metrics
+        """
+        # Get overall metrics
+        overall_metrics = self.get_strategy_metrics(strategy_name=strategy_name)
+
+        # Get daily metrics (last 7 days)
+        daily_metrics = self.get_daily_metrics(days=7, strategy_name=strategy_name)
+
+        # Get weekly metrics (last 4 weeks)
+        weekly_metrics = self.get_weekly_metrics(weeks=4, strategy_name=strategy_name)
+
+        # Calculate improvement trend
+        improvement_trend = self.calculate_improvement_trend(weekly_metrics[::-1])
+
+        return PerformanceDashboard(
+            generated_at=datetime.now(UTC).isoformat(),
+            overall_metrics=overall_metrics,
+            daily_metrics=daily_metrics,
+            weekly_metrics=weekly_metrics,
+            improvement_trend=improvement_trend,
+        )
+
+    def export_dashboard_json(
+        self, dashboard: PerformanceDashboard
+    ) -> str:
+        """Export dashboard as JSON string.
+
+        Args:
+            dashboard: PerformanceDashboard object
+
+        Returns:
+            JSON string representation
+        """
+        data = {
+            "generated_at": dashboard.generated_at,
+            "overall_metrics": asdict(dashboard.overall_metrics),
+            "daily_metrics": [asdict(m) for m in dashboard.daily_metrics],
+            "weekly_metrics": [asdict(m) for m in dashboard.weekly_metrics],
+            "improvement_trend": dashboard.improvement_trend,
+        }
+        return json.dumps(data, indent=2)
+
+    def log_dashboard(self, dashboard: PerformanceDashboard) -> None:
+        """Log dashboard summary to logger.
+
+        Args:
+            dashboard: PerformanceDashboard object
+        """
+        logger.info("=" * 60)
+        logger.info("PERFORMANCE DASHBOARD")
+        logger.info("=" * 60)
+        logger.info("Generated: %s", dashboard.generated_at)
+        logger.info("")
+        logger.info("Overall Performance:")
+        logger.info("  Total Trades: %d", dashboard.overall_metrics.total_trades)
+        logger.info("  Win Rate: %.2f%%", dashboard.overall_metrics.win_rate)
+        logger.info("  Average P&L: %.2f", dashboard.overall_metrics.avg_pnl)
+        logger.info("  Total P&L: %.2f", dashboard.overall_metrics.total_pnl)
+        logger.info("")
+        logger.info("Improvement Trend (%s):", dashboard.improvement_trend["trend"])
+        logger.info("  Win Rate Change: %+.2f%%", dashboard.improvement_trend["win_rate_change"])
+        logger.info("  P&L Change: %+.2f", dashboard.improvement_trend["pnl_change"])
+        logger.info("=" * 60)

src/evolution/scorecard.py

@@ -0,0 +1,25 @@
+"""Daily scorecard model for end-of-day performance review."""
+
+from __future__ import annotations
+
+from dataclasses import dataclass, field
+
+
+@dataclass
+class DailyScorecard:
+    """Structured daily performance snapshot for a single market."""
+
+    date: str
+    market: str
+    total_decisions: int
+    buys: int
+    sells: int
+    holds: int
+    total_pnl: float
+    win_rate: float
+    avg_confidence: float
+    scenario_match_rate: float
+    top_winners: list[str] = field(default_factory=list)
+    top_losers: list[str] = field(default_factory=list)
+    lessons: list[str] = field(default_factory=list)
+    cross_market_note: str = ""

src/logging/__init__.py

@@ -0,0 +1,5 @@
+"""Decision logging and audit trail for trade decisions."""
+
+from src.logging.decision_logger import DecisionLog, DecisionLogger
+
+__all__ = ["DecisionLog", "DecisionLogger"]

src/logging/decision_logger.py

@@ -0,0 +1,235 @@
+"""Decision logging system with context snapshots for comprehensive audit trail."""
+
+from __future__ import annotations
+
+import json
+import sqlite3
+import uuid
+from dataclasses import dataclass
+from datetime import UTC, datetime
+from typing import Any
+
+
+@dataclass
+class DecisionLog:
+    """A logged trading decision with context and outcome."""
+
+    decision_id: str
+    timestamp: str
+    stock_code: str
+    market: str
+    exchange_code: str
+    action: str
+    confidence: int
+    rationale: str
+    context_snapshot: dict[str, Any]
+    input_data: dict[str, Any]
+    outcome_pnl: float | None = None
+    outcome_accuracy: int | None = None
+    reviewed: bool = False
+    review_notes: str | None = None
+
+
+class DecisionLogger:
+    """Logs trading decisions with full context for review and evolution."""
+
+    def __init__(self, conn: sqlite3.Connection) -> None:
+        """Initialize the decision logger with a database connection."""
+        self.conn = conn
+
+    def log_decision(
+        self,
+        stock_code: str,
+        market: str,
+        exchange_code: str,
+        action: str,
+        confidence: int,
+        rationale: str,
+        context_snapshot: dict[str, Any],
+        input_data: dict[str, Any],
+    ) -> str:
+        """Log a trading decision with full context.
+
+        Args:
+            stock_code: Stock symbol
+            market: Market code (e.g., "KR", "US_NASDAQ")
+            exchange_code: Exchange code (e.g., "KRX", "NASDAQ")
+            action: Trading action (BUY/SELL/HOLD)
+            confidence: Confidence level (0-100)
+            rationale: Reasoning for the decision
+            context_snapshot: L1-L7 context snapshot at decision time
+            input_data: Market data inputs (price, volume, orderbook, etc.)
+
+        Returns:
+            decision_id: Unique identifier for this decision
+        """
+        decision_id = str(uuid.uuid4())
+        timestamp = datetime.now(UTC).isoformat()
+
+        self.conn.execute(
+            """
+            INSERT INTO decision_logs (
+                decision_id, timestamp, stock_code, market, exchange_code,
+                action, confidence, rationale, context_snapshot, input_data
+            )
+            VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?)
+            """,
+            (
+                decision_id,
+                timestamp,
+                stock_code,
+                market,
+                exchange_code,
+                action,
+                confidence,
+                rationale,
+                json.dumps(context_snapshot),
+                json.dumps(input_data),
+            ),
+        )
+        self.conn.commit()
+
+        return decision_id
+
+    def get_unreviewed_decisions(
+        self, min_confidence: int = 80, limit: int | None = None
+    ) -> list[DecisionLog]:
+        """Get unreviewed decisions with high confidence.
+
+        Args:
+            min_confidence: Minimum confidence threshold (default 80)
+            limit: Maximum number of results (None = unlimited)
+
+        Returns:
+            List of unreviewed DecisionLog objects
+        """
+        query = """
+            SELECT
+                decision_id, timestamp, stock_code, market, exchange_code,
+                action, confidence, rationale, context_snapshot, input_data,
+                outcome_pnl, outcome_accuracy, reviewed, review_notes
+            FROM decision_logs
+            WHERE reviewed = 0 AND confidence >= ?
+            ORDER BY timestamp DESC
+        """
+        if limit is not None:
+            query += f" LIMIT {limit}"
+
+        cursor = self.conn.execute(query, (min_confidence,))
+        return [self._row_to_decision_log(row) for row in cursor.fetchall()]
+
+    def mark_reviewed(self, decision_id: str, notes: str) -> None:
+        """Mark a decision as reviewed with notes.
+
+        Args:
+            decision_id: Decision identifier
+            notes: Review notes and insights
+        """
+        self.conn.execute(
+            """
+            UPDATE decision_logs
+            SET reviewed = 1, review_notes = ?
+            WHERE decision_id = ?
+            """,
+            (notes, decision_id),
+        )
+        self.conn.commit()
+
+    def update_outcome(
+        self, decision_id: str, pnl: float, accuracy: int
+    ) -> None:
+        """Update the outcome of a decision after trade execution.
+
+        Args:
+            decision_id: Decision identifier
+            pnl: Actual profit/loss realized
+            accuracy: 1 if decision was correct, 0 if wrong
+        """
+        self.conn.execute(
+            """
+            UPDATE decision_logs
+            SET outcome_pnl = ?, outcome_accuracy = ?
+            WHERE decision_id = ?
+            """,
+            (pnl, accuracy, decision_id),
+        )
+        self.conn.commit()
+
+    def get_decision_by_id(self, decision_id: str) -> DecisionLog | None:
+        """Get a specific decision by ID.
+
+        Args:
+            decision_id: Decision identifier
+
+        Returns:
+            DecisionLog object or None if not found
+        """
+        cursor = self.conn.execute(
+            """
+            SELECT
+                decision_id, timestamp, stock_code, market, exchange_code,
+                action, confidence, rationale, context_snapshot, input_data,
+                outcome_pnl, outcome_accuracy, reviewed, review_notes
+            FROM decision_logs
+            WHERE decision_id = ?
+            """,
+            (decision_id,),
+        )
+        row = cursor.fetchone()
+        return self._row_to_decision_log(row) if row else None
+
+    def get_losing_decisions(
+        self, min_confidence: int = 80, min_loss: float = -100.0
+    ) -> list[DecisionLog]:
+        """Get high-confidence decisions that resulted in losses.
+
+        Useful for identifying patterns in failed predictions.
+
+        Args:
+            min_confidence: Minimum confidence threshold (default 80)
+            min_loss: Minimum loss amount (default -100.0, i.e., loss >= 100)
+
+        Returns:
+            List of losing DecisionLog objects
+        """
+        cursor = self.conn.execute(
+            """
+            SELECT
+                decision_id, timestamp, stock_code, market, exchange_code,
+                action, confidence, rationale, context_snapshot, input_data,
+                outcome_pnl, outcome_accuracy, reviewed, review_notes
+            FROM decision_logs
+            WHERE confidence >= ?
+              AND outcome_pnl IS NOT NULL
+              AND outcome_pnl <= ?
+            ORDER BY outcome_pnl ASC
+            """,
+            (min_confidence, min_loss),
+        )
+        return [self._row_to_decision_log(row) for row in cursor.fetchall()]
+
+    def _row_to_decision_log(self, row: tuple[Any, ...]) -> DecisionLog:
+        """Convert a database row to a DecisionLog object.
+
+        Args:
+            row: Database row tuple
+
+        Returns:
+            DecisionLog object
+        """
+        return DecisionLog(
+            decision_id=row[0],
+            timestamp=row[1],
+            stock_code=row[2],
+            market=row[3],
+            exchange_code=row[4],
+            action=row[5],
+            confidence=row[6],
+            rationale=row[7],
+            context_snapshot=json.loads(row[8]),
+            input_data=json.loads(row[9]),
+            outcome_pnl=row[10],
+            outcome_accuracy=row[11],
+            reviewed=bool(row[12]),
+            review_notes=row[13],
+        )

src/markets/schedule.py

@@ -123,6 +123,23 @@ MARKETS: dict[str, MarketInfo] = {
     ),
 }
 
+MARKET_SHORTHAND: dict[str, list[str]] = {
+    "US": ["US_NASDAQ", "US_NYSE", "US_AMEX"],
+    "CN": ["CN_SHA", "CN_SZA"],
+    "VN": ["VN_HAN", "VN_HCM"],
+}
+
+
+def expand_market_codes(codes: list[str]) -> list[str]:
+    """Expand shorthand market codes into concrete exchange market codes."""
+    expanded: list[str] = []
+    for code in codes:
+        if code in MARKET_SHORTHAND:
+            expanded.extend(MARKET_SHORTHAND[code])
+        else:
+            expanded.append(code)
+    return expanded
+
 
 def is_market_open(market: MarketInfo, now: datetime | None = None) -> bool:
     """

src/notifications/README.md

@@ -0,0 +1,350 @@
+# Telegram Notifications
+
+Real-time trading event notifications via Telegram Bot API.
+
+## Setup
+
+### 1. Create a Telegram Bot
+
+1. Open Telegram and message [@BotFather](https://t.me/BotFather)
+2. Send `/newbot` command
+3. Follow prompts to name your bot
+4. Save the **bot token** (looks like `1234567890:ABCdefGHIjklMNOpqrsTUVwxyz`)
+
+### 2. Get Your Chat ID
+
+**Option A: Using @userinfobot**
+1. Message [@userinfobot](https://t.me/userinfobot) on Telegram
+2. Send `/start`
+3. Save your numeric **chat ID** (e.g., `123456789`)
+
+**Option B: Using @RawDataBot**
+1. Message [@RawDataBot](https://t.me/rawdatabot) on Telegram
+2. Look for `"id":` in the JSON response
+3. Save your numeric **chat ID**
+
+### 3. Configure Environment
+
+Add to your `.env` file:
+
+```bash
+TELEGRAM_BOT_TOKEN=1234567890:ABCdefGHIjklMNOpqrsTUVwxyz
+TELEGRAM_CHAT_ID=123456789
+TELEGRAM_ENABLED=true
+```
+
+### 4. Test the Bot
+
+Start a conversation with your bot on Telegram first (send `/start`), then run:
+
+```bash
+python -m src.main --mode=paper
+```
+
+You should receive a startup notification.
+
+## Message Examples
+
+### Trade Execution
+```
+🟢 BUY
+Symbol: AAPL (United States)
+Quantity: 10 shares
+Price: 150.25
+Confidence: 85%
+```
+
+### Circuit Breaker
+```
+🚨 CIRCUIT BREAKER TRIPPED
+P&L: -3.15% (threshold: -3.0%)
+Trading halted for safety
+```
+
+### Fat-Finger Protection
+```
+⚠️ Fat-Finger Protection
+Order rejected: TSLA
+Attempted: 45.0% of cash
+Max allowed: 30%
+Amount: 45,000 / 100,000
+```
+
+### Market Open/Close
+```
+ℹ️ Market Open
+Korea trading session started
+
+ℹ️ Market Close
+Korea trading session ended
+📈 P&L: +1.25%
+```
+
+### System Status
+```
+📝 System Started
+Mode: PAPER
+Markets: KRX, NASDAQ
+
+System Shutdown
+Normal shutdown
+```
+
+## Notification Priorities
+
+| Priority | Emoji | Use Case |
+|----------|-------|----------|
+| LOW | ℹ️ | Market open/close |
+| MEDIUM | 📊 | Trade execution, system start/stop |
+| HIGH | ⚠️ | Fat-finger protection, errors |
+| CRITICAL | 🚨 | Circuit breaker trips |
+
+## Rate Limiting
+
+- Default: 1 message per second
+- Prevents hitting Telegram's global rate limits
+- Configurable via `rate_limit` parameter
+
+## Troubleshooting
+
+### No notifications received
+
+1. **Check bot configuration**
+   ```bash
+   # Verify env variables are set
+   grep TELEGRAM .env
+   ```
+
+2. **Start conversation with bot**
+   - Open bot in Telegram
+   - Send `/start` command
+   - Bot cannot message users who haven't started a conversation
+
+3. **Check logs**
+   ```bash
+   # Look for Telegram-related errors
+   python -m src.main --mode=paper 2>&1 | grep -i telegram
+   ```
+
+4. **Verify bot token**
+   ```bash
+   curl https://api.telegram.org/bot<YOUR_TOKEN>/getMe
+   # Should return bot info (not 401 error)
+   ```
+
+5. **Verify chat ID**
+   ```bash
+   curl -X POST https://api.telegram.org/bot<YOUR_TOKEN>/sendMessage \
+     -H 'Content-Type: application/json' \
+     -d '{"chat_id": "<YOUR_CHAT_ID>", "text": "Test"}'
+   # Should send a test message
+   ```
+
+### Notifications delayed
+
+- Check rate limiter settings
+- Verify network connection
+- Look for timeout errors in logs
+
+### "Chat not found" error
+
+- Incorrect chat ID
+- Bot blocked by user
+- Need to send `/start` to bot first
+
+### "Unauthorized" error
+
+- Invalid bot token
+- Token revoked (regenerate with @BotFather)
+
+## Graceful Degradation
+
+The system works without Telegram notifications:
+
+- Missing credentials → notifications disabled automatically
+- API errors → logged but trading continues
+- Network timeouts → trading loop unaffected
+- Rate limiting → messages queued, trading proceeds
+
+**Notifications never crash the trading system.**
+
+## Security Notes
+
+- Never commit `.env` file with credentials
+- Bot token grants full bot control
+- Chat ID is not sensitive (just a number)
+- Messages are sent over HTTPS
+- No trading credentials in notifications
+
+## Advanced Usage
+
+### Group Notifications
+
+1. Add bot to Telegram group
+2. Get group chat ID (negative number like `-123456789`)
+3. Use group chat ID in `TELEGRAM_CHAT_ID`
+
+### Multiple Recipients
+
+Create multiple bots or use a broadcast group with multiple members.
+
+### Custom Rate Limits
+
+Not currently exposed in config, but can be modified in code:
+
+```python
+telegram = TelegramClient(
+    bot_token=settings.TELEGRAM_BOT_TOKEN,
+    chat_id=settings.TELEGRAM_CHAT_ID,
+    rate_limit=2.0,  # 2 messages per second
+)
+```
+
+## Bidirectional Commands
+
+Control your trading bot remotely via Telegram commands. The bot not only sends notifications but also accepts commands for real-time control.
+
+### Available Commands
+
+| Command | Description |
+|---------|-------------|
+| `/start` | Welcome message with quick start guide |
+| `/help` | List all available commands |
+| `/status` | Current trading status (mode, markets, P&L, circuit breaker) |
+| `/positions` | View current holdings grouped by market |
+| `/stop` | Pause all trading operations |
+| `/resume` | Resume trading operations |
+
+### Command Examples
+
+**Check Trading Status**
+```
+You: /status
+
+Bot:
+📊 Trading Status
+
+Mode: PAPER
+Markets: Korea, United States
+Trading: Active
+
+Current P&L: +2.50%
+Circuit Breaker: -3.0%
+```
+
+**View Holdings**
+```
+You: /positions
+
+Bot:
+💼 Current Holdings
+
+🇰🇷 Korea
+• 005930: 10 shares @ 70,000
+• 035420: 5 shares @ 200,000
+
+🇺🇸 Overseas
+• AAPL: 15 shares @ 175
+• TSLA: 8 shares @ 245
+
+Cash: ₩5,000,000
+```
+
+**Pause Trading**
+```
+You: /stop
+
+Bot:
+⏸️ Trading Paused
+
+All trading operations have been suspended.
+Use /resume to restart trading.
+```
+
+**Resume Trading**
+```
+You: /resume
+
+Bot:
+▶️ Trading Resumed
+
+Trading operations have been restarted.
+```
+
+### Security
+
+**Chat ID Verification**
+- Commands are only accepted from the configured `TELEGRAM_CHAT_ID`
+- Unauthorized users receive no response
+- Command attempts from wrong chat IDs are logged
+
+**Authorization Required**
+- Only the bot owner (chat ID in `.env`) can control trading
+- No way for unauthorized users to discover or use commands
+- All command executions are logged for audit
+
+### Configuration
+
+Add to your `.env` file:
+
+```bash
+# Commands are enabled by default
+TELEGRAM_COMMANDS_ENABLED=true
+
+# Polling interval (seconds) - how often to check for commands
+TELEGRAM_POLLING_INTERVAL=1.0
+```
+
+To disable commands but keep notifications:
+```bash
+TELEGRAM_COMMANDS_ENABLED=false
+```
+
+### How It Works
+
+1. **Long Polling**: Bot checks Telegram API every second for new messages
+2. **Command Parsing**: Messages starting with `/` are parsed as commands
+3. **Authentication**: Chat ID is verified before executing any command
+4. **Execution**: Command handler is called with current bot state
+5. **Response**: Result is sent back via Telegram
+
+### Error Handling
+
+- Command parsing errors → "Unknown command" response
+- API failures → Graceful degradation, error logged
+- Invalid state → Appropriate message (e.g., "Trading is already paused")
+- Trading loop isolation → Command errors never crash trading
+
+### Troubleshooting Commands
+
+**Commands not responding**
+1. Check `TELEGRAM_COMMANDS_ENABLED=true` in `.env`
+2. Verify you started conversation with `/start`
+3. Check logs for command handler errors
+4. Confirm chat ID matches `.env` configuration
+
+**Wrong chat ID**
+- Commands from unauthorized chats are silently ignored
+- Check logs for "unauthorized chat_id" warnings
+
+**Delayed responses**
+- Polling interval is 1 second by default
+- Network latency may add delay
+- Check `TELEGRAM_POLLING_INTERVAL` setting
+
+## API Reference
+
+See `telegram_client.py` for full API documentation.
+
+### Notification Methods
+- `notify_trade_execution()` - Trade alerts
+- `notify_circuit_breaker()` - Emergency stops
+- `notify_fat_finger()` - Order rejections
+- `notify_market_open/close()` - Session tracking
+- `notify_system_start/shutdown()` - Lifecycle events
+- `notify_error()` - Error alerts
+
+### Command Handler
+- `TelegramCommandHandler` - Bidirectional command processing
+- `register_command()` - Register custom command handlers
+- `start_polling()` / `stop_polling()` - Lifecycle management

src/notifications/__init__.py

@@ -0,0 +1,5 @@
+"""Real-time notifications for trading events."""
+
+from src.notifications.telegram_client import TelegramClient
+
+__all__ = ["TelegramClient"]

src/notifications/telegram_client.py

@@ -0,0 +1,734 @@
+"""Telegram notification client for real-time trading alerts."""
+
+import asyncio
+import logging
+import time
+from collections.abc import Awaitable, Callable
+from dataclasses import dataclass, fields
+from enum import Enum
+from typing import ClassVar
+
+import aiohttp
+
+logger = logging.getLogger(__name__)
+
+
+class NotificationPriority(Enum):
+    """Priority levels for notifications with emoji indicators."""
+
+    LOW = ("ℹ️", "info")
+    MEDIUM = ("📊", "medium")
+    HIGH = ("⚠️", "warning")
+    CRITICAL = ("🚨", "critical")
+
+    def __init__(self, emoji: str, label: str) -> None:
+        self.emoji = emoji
+        self.label = label
+
+
+class LeakyBucket:
+    """Rate limiter using leaky bucket algorithm."""
+
+    def __init__(self, rate: float, capacity: int = 1) -> None:
+        """
+        Initialize rate limiter.
+
+        Args:
+            rate: Maximum requests per second
+            capacity: Bucket capacity (burst size)
+        """
+        self._rate = rate
+        self._capacity = capacity
+        self._tokens = float(capacity)
+        self._last_update = time.monotonic()
+        self._lock = asyncio.Lock()
+
+    async def acquire(self) -> None:
+        """Wait until a token is available, then consume it."""
+        async with self._lock:
+            now = time.monotonic()
+            elapsed = now - self._last_update
+            self._tokens = min(self._capacity, self._tokens + elapsed * self._rate)
+            self._last_update = now
+
+            if self._tokens < 1.0:
+                wait_time = (1.0 - self._tokens) / self._rate
+                await asyncio.sleep(wait_time)
+                self._tokens = 0.0
+            else:
+                self._tokens -= 1.0
+
+
+@dataclass
+class NotificationFilter:
+    """Granular on/off flags for each notification type.
+
+    circuit_breaker is intentionally omitted — it is always sent regardless.
+    """
+
+    # Maps user-facing command keys to dataclass field names
+    KEYS: ClassVar[dict[str, str]] = {
+        "trades": "trades",
+        "market": "market_open_close",
+        "fatfinger": "fat_finger",
+        "system": "system_events",
+        "playbook": "playbook",
+        "scenario": "scenario_match",
+        "errors": "errors",
+    }
+
+    trades: bool = True
+    market_open_close: bool = True
+    fat_finger: bool = True
+    system_events: bool = True
+    playbook: bool = True
+    scenario_match: bool = True
+    errors: bool = True
+
+    def set_flag(self, key: str, value: bool) -> bool:
+        """Set a filter flag by user-facing key. Returns False if key is unknown."""
+        field = self.KEYS.get(key.lower())
+        if field is None:
+            return False
+        setattr(self, field, value)
+        return True
+
+    def as_dict(self) -> dict[str, bool]:
+        """Return {user_key: current_value} for display."""
+        return {k: getattr(self, field) for k, field in self.KEYS.items()}
+
+
+@dataclass
+class NotificationMessage:
+    """Internal notification message structure."""
+
+    priority: NotificationPriority
+    message: str
+
+
+class TelegramClient:
+    """Telegram Bot API client for sending trading notifications."""
+
+    API_BASE = "https://api.telegram.org/bot{token}"
+    DEFAULT_TIMEOUT = 5.0  # seconds
+    DEFAULT_RATE = 1.0  # messages per second
+
+    def __init__(
+        self,
+        bot_token: str | None = None,
+        chat_id: str | None = None,
+        enabled: bool = True,
+        rate_limit: float = DEFAULT_RATE,
+        notification_filter: NotificationFilter | None = None,
+    ) -> None:
+        """
+        Initialize Telegram client.
+
+        Args:
+            bot_token: Telegram bot token from @BotFather
+            chat_id: Target chat ID (user or group)
+            enabled: Enable/disable notifications globally
+            rate_limit: Maximum messages per second
+            notification_filter: Granular per-type on/off flags
+        """
+        self._bot_token = bot_token
+        self._chat_id = chat_id
+        self._enabled = enabled
+        self._rate_limiter = LeakyBucket(rate=rate_limit)
+        self._session: aiohttp.ClientSession | None = None
+        self._filter = notification_filter if notification_filter is not None else NotificationFilter()
+
+        if not enabled:
+            logger.info("Telegram notifications disabled via configuration")
+        elif bot_token is None or chat_id is None:
+            logger.warning(
+                "Telegram notifications disabled (missing bot_token or chat_id)"
+            )
+            self._enabled = False
+        else:
+            logger.info("Telegram notifications enabled for chat_id=%s", chat_id)
+
+    def _get_session(self) -> aiohttp.ClientSession:
+        """Get or create aiohttp session."""
+        if self._session is None or self._session.closed:
+            self._session = aiohttp.ClientSession(
+                timeout=aiohttp.ClientTimeout(total=self.DEFAULT_TIMEOUT)
+            )
+        return self._session
+
+    async def close(self) -> None:
+        """Close HTTP session."""
+        if self._session is not None and not self._session.closed:
+            await self._session.close()
+
+    def set_notification(self, key: str, value: bool) -> bool:
+        """Toggle a notification type by user-facing key at runtime.
+
+        Args:
+            key: User-facing key (e.g. "scenario", "market", "all")
+            value: True to enable, False to disable
+
+        Returns:
+            True if key was valid, False if unknown.
+        """
+        if key == "all":
+            for k in NotificationFilter.KEYS:
+                self._filter.set_flag(k, value)
+            return True
+        return self._filter.set_flag(key, value)
+
+    def filter_status(self) -> dict[str, bool]:
+        """Return current per-type filter state keyed by user-facing names."""
+        return self._filter.as_dict()
+
+    async def send_message(self, text: str, parse_mode: str = "HTML") -> bool:
+        """
+        Send a generic text message to Telegram.
+
+        Args:
+            text: Message text to send
+            parse_mode: Parse mode for formatting (HTML or Markdown)
+
+        Returns:
+            True if message was sent successfully, False otherwise
+        """
+        if not self._enabled:
+            return False
+
+        try:
+            await self._rate_limiter.acquire()
+
+            url = f"{self.API_BASE.format(token=self._bot_token)}/sendMessage"
+            payload = {
+                "chat_id": self._chat_id,
+                "text": text,
+                "parse_mode": parse_mode,
+            }
+
+            session = self._get_session()
+            async with session.post(url, json=payload) as resp:
+                if resp.status != 200:
+                    error_text = await resp.text()
+                    logger.error(
+                        "Telegram API error (status=%d): %s", resp.status, error_text
+                    )
+                    return False
+                logger.debug("Telegram message sent: %s", text[:50])
+                return True
+
+        except asyncio.TimeoutError:
+            logger.error("Telegram message timeout")
+            return False
+        except aiohttp.ClientError as exc:
+            logger.error("Telegram message failed: %s", exc)
+            return False
+        except Exception as exc:
+            logger.error("Unexpected error sending message: %s", exc)
+            return False
+
+    async def _send_notification(self, msg: NotificationMessage) -> None:
+        """
+        Send notification to Telegram with graceful degradation.
+
+        Args:
+            msg: Notification message to send
+        """
+        formatted_message = f"{msg.priority.emoji} {msg.message}"
+        await self.send_message(formatted_message)
+
+    async def notify_trade_execution(
+        self,
+        stock_code: str,
+        market: str,
+        action: str,
+        quantity: int,
+        price: float,
+        confidence: float,
+    ) -> None:
+        """
+        Notify trade execution.
+
+        Args:
+            stock_code: Stock ticker symbol
+            market: Market name (e.g., "Korea", "United States")
+            action: "BUY" or "SELL"
+            quantity: Number of shares
+            price: Execution price
+            confidence: AI confidence level (0-100)
+        """
+        if not self._filter.trades:
+            return
+        emoji = "🟢" if action == "BUY" else "🔴"
+        message = (
+            f"<b>{emoji} {action}</b>\n"
+            f"Symbol: <code>{stock_code}</code> ({market})\n"
+            f"Quantity: {quantity:,} shares\n"
+            f"Price: {price:,.2f}\n"
+            f"Confidence: {confidence:.0f}%"
+        )
+        await self._send_notification(
+            NotificationMessage(priority=NotificationPriority.MEDIUM, message=message)
+        )
+
+    async def notify_market_open(self, market_name: str) -> None:
+        """
+        Notify market opening.
+
+        Args:
+            market_name: Name of the market (e.g., "Korea", "United States")
+        """
+        if not self._filter.market_open_close:
+            return
+        message = f"<b>Market Open</b>\n{market_name} trading session started"
+        await self._send_notification(
+            NotificationMessage(priority=NotificationPriority.LOW, message=message)
+        )
+
+    async def notify_market_close(self, market_name: str, pnl_pct: float) -> None:
+        """
+        Notify market closing.
+
+        Args:
+            market_name: Name of the market
+            pnl_pct: Final P&L percentage for the session
+        """
+        if not self._filter.market_open_close:
+            return
+        pnl_sign = "+" if pnl_pct >= 0 else ""
+        pnl_emoji = "📈" if pnl_pct >= 0 else "📉"
+        message = (
+            f"<b>Market Close</b>\n"
+            f"{market_name} trading session ended\n"
+            f"{pnl_emoji} P&L: {pnl_sign}{pnl_pct:.2f}%"
+        )
+        await self._send_notification(
+            NotificationMessage(priority=NotificationPriority.LOW, message=message)
+        )
+
+    async def notify_circuit_breaker(
+        self, pnl_pct: float, threshold: float
+    ) -> None:
+        """
+        Notify circuit breaker activation.
+
+        Args:
+            pnl_pct: Current P&L percentage
+            threshold: Circuit breaker threshold
+        """
+        message = (
+            f"<b>CIRCUIT BREAKER TRIPPED</b>\n"
+            f"P&L: {pnl_pct:.2f}% (threshold: {threshold:.1f}%)\n"
+            f"Trading halted for safety"
+        )
+        await self._send_notification(
+            NotificationMessage(priority=NotificationPriority.CRITICAL, message=message)
+        )
+
+    async def notify_fat_finger(
+        self,
+        stock_code: str,
+        order_amount: float,
+        total_cash: float,
+        max_pct: float,
+    ) -> None:
+        """
+        Notify fat-finger protection rejection.
+
+        Args:
+            stock_code: Stock ticker symbol
+            order_amount: Attempted order amount
+            total_cash: Total available cash
+            max_pct: Maximum allowed percentage
+        """
+        if not self._filter.fat_finger:
+            return
+        attempted_pct = (order_amount / total_cash) * 100 if total_cash > 0 else 0
+        message = (
+            f"<b>Fat-Finger Protection</b>\n"
+            f"Order rejected: <code>{stock_code}</code>\n"
+            f"Attempted: {attempted_pct:.1f}% of cash\n"
+            f"Max allowed: {max_pct:.0f}%\n"
+            f"Amount: {order_amount:,.0f} / {total_cash:,.0f}"
+        )
+        await self._send_notification(
+            NotificationMessage(priority=NotificationPriority.HIGH, message=message)
+        )
+
+    async def notify_system_start(
+        self, mode: str, enabled_markets: list[str]
+    ) -> None:
+        """
+        Notify system startup.
+
+        Args:
+            mode: Trading mode ("paper" or "live")
+            enabled_markets: List of enabled market codes
+        """
+        if not self._filter.system_events:
+            return
+        mode_emoji = "📝" if mode == "paper" else "💰"
+        markets_str = ", ".join(enabled_markets)
+        message = (
+            f"<b>{mode_emoji} System Started</b>\n"
+            f"Mode: {mode.upper()}\n"
+            f"Markets: {markets_str}"
+        )
+        await self._send_notification(
+            NotificationMessage(priority=NotificationPriority.MEDIUM, message=message)
+        )
+
+    async def notify_playbook_generated(
+        self,
+        market: str,
+        stock_count: int,
+        scenario_count: int,
+        token_count: int,
+    ) -> None:
+        """
+        Notify that a daily playbook was generated.
+
+        Args:
+            market: Market code (e.g., "KR", "US")
+            stock_count: Number of stocks in the playbook
+            scenario_count: Total number of scenarios
+            token_count: Gemini token usage for the playbook
+        """
+        if not self._filter.playbook:
+            return
+        message = (
+            f"<b>Playbook Generated</b>\n"
+            f"Market: {market}\n"
+            f"Stocks: {stock_count}\n"
+            f"Scenarios: {scenario_count}\n"
+            f"Tokens: {token_count}"
+        )
+        await self._send_notification(
+            NotificationMessage(priority=NotificationPriority.MEDIUM, message=message)
+        )
+
+    async def notify_scenario_matched(
+        self,
+        stock_code: str,
+        action: str,
+        condition_summary: str,
+        confidence: float,
+    ) -> None:
+        """
+        Notify that a scenario matched for a stock.
+
+        Args:
+            stock_code: Stock ticker symbol
+            action: Scenario action (BUY/SELL/HOLD/REDUCE_ALL)
+            condition_summary: Short summary of the matched condition
+            confidence: Scenario confidence (0-100)
+        """
+        if not self._filter.scenario_match:
+            return
+        message = (
+            f"<b>Scenario Matched</b>\n"
+            f"Symbol: <code>{stock_code}</code>\n"
+            f"Action: {action}\n"
+            f"Condition: {condition_summary}\n"
+            f"Confidence: {confidence:.0f}%"
+        )
+        await self._send_notification(
+            NotificationMessage(priority=NotificationPriority.HIGH, message=message)
+        )
+
+    async def notify_playbook_failed(self, market: str, reason: str) -> None:
+        """
+        Notify that playbook generation failed.
+
+        Args:
+            market: Market code (e.g., "KR", "US")
+            reason: Failure reason summary
+        """
+        if not self._filter.playbook:
+            return
+        message = (
+            f"<b>Playbook Failed</b>\n"
+            f"Market: {market}\n"
+            f"Reason: {reason[:200]}"
+        )
+        await self._send_notification(
+            NotificationMessage(priority=NotificationPriority.HIGH, message=message)
+        )
+
+    async def notify_system_shutdown(self, reason: str) -> None:
+        """
+        Notify system shutdown.
+
+        Args:
+            reason: Reason for shutdown (e.g., "Normal shutdown", "Circuit breaker")
+        """
+        if not self._filter.system_events:
+            return
+        message = f"<b>System Shutdown</b>\n{reason}"
+        priority = (
+            NotificationPriority.CRITICAL
+            if "circuit breaker" in reason.lower()
+            else NotificationPriority.MEDIUM
+        )
+        await self._send_notification(
+            NotificationMessage(priority=priority, message=message)
+        )
+
+    async def notify_unfilled_order(
+        self,
+        stock_code: str,
+        market: str,
+        action: str,
+        quantity: int,
+        outcome: str,
+        new_price: float | None = None,
+    ) -> None:
+        """Notify about an unfilled overseas order that was cancelled or resubmitted.
+
+        Args:
+            stock_code: Stock ticker symbol.
+            market: Exchange/market code (e.g., "NASD", "SEHK").
+            action: "BUY" or "SELL".
+            quantity: Unfilled quantity.
+            outcome: "cancelled" or "resubmitted".
+            new_price: New order price if resubmitted (None if only cancelled).
+        """
+        if not self._filter.trades:
+            return
+        # SELL resubmit is high priority — position liquidation at risk.
+        # BUY cancel is medium priority — only cash is freed.
+        priority = (
+            NotificationPriority.HIGH
+            if action == "SELL"
+            else NotificationPriority.MEDIUM
+        )
+        outcome_emoji = "🔄" if outcome == "resubmitted" else "❌"
+        outcome_label = "재주문" if outcome == "resubmitted" else "취소됨"
+        action_emoji = "🔴" if action == "SELL" else "🟢"
+        lines = [
+            f"<b>{outcome_emoji} 미체결 주문 {outcome_label}</b>",
+            f"Symbol: <code>{stock_code}</code> ({market})",
+            f"Action: {action_emoji} {action}",
+            f"Quantity: {quantity:,} shares",
+        ]
+        if new_price is not None:
+            lines.append(f"New Price: {new_price:.4f}")
+        message = "\n".join(lines)
+        await self._send_notification(NotificationMessage(priority=priority, message=message))
+
+    async def notify_error(
+        self, error_type: str, error_msg: str, context: str
+    ) -> None:
+        """
+        Notify system error.
+
+        Args:
+            error_type: Type of error (e.g., "Connection Error")
+            error_msg: Error message
+            context: Error context (e.g., stock code, market)
+        """
+        if not self._filter.errors:
+            return
+        message = (
+            f"<b>Error: {error_type}</b>\n"
+            f"Context: {context}\n"
+            f"Message: {error_msg[:200]}"  # Truncate long errors
+        )
+        await self._send_notification(
+            NotificationMessage(priority=NotificationPriority.HIGH, message=message)
+        )
+
+
+class TelegramCommandHandler:
+    """Handles incoming Telegram commands via long polling."""
+
+    def __init__(
+        self, client: TelegramClient, polling_interval: float = 1.0
+    ) -> None:
+        """
+        Initialize command handler.
+
+        Args:
+            client: TelegramClient instance for sending responses
+            polling_interval: Polling interval in seconds
+        """
+        self._client = client
+        self._polling_interval = polling_interval
+        self._commands: dict[str, Callable[[], Awaitable[None]]] = {}
+        self._commands_with_args: dict[str, Callable[[list[str]], Awaitable[None]]] = {}
+        self._last_update_id = 0
+        self._polling_task: asyncio.Task[None] | None = None
+        self._running = False
+
+    def register_command(
+        self, command: str, handler: Callable[[], Awaitable[None]]
+    ) -> None:
+        """
+        Register a command handler (no arguments).
+
+        Args:
+            command: Command name (without leading slash, e.g., "start")
+            handler: Async function to handle the command
+        """
+        self._commands[command] = handler
+        logger.debug("Registered command handler: /%s", command)
+
+    def register_command_with_args(
+        self, command: str, handler: Callable[[list[str]], Awaitable[None]]
+    ) -> None:
+        """
+        Register a command handler that receives trailing arguments.
+
+        Args:
+            command: Command name (without leading slash, e.g., "notify")
+            handler: Async function receiving list of argument tokens
+        """
+        self._commands_with_args[command] = handler
+        logger.debug("Registered command handler (with args): /%s", command)
+
+    async def start_polling(self) -> None:
+        """Start long polling for commands."""
+        if self._running:
+            logger.warning("Command handler already running")
+            return
+
+        if not self._client._enabled:
+            logger.info("Command handler disabled (TelegramClient disabled)")
+            return
+
+        self._running = True
+        self._polling_task = asyncio.create_task(self._poll_loop())
+        logger.info("Started Telegram command polling")
+
+    async def stop_polling(self) -> None:
+        """Stop polling and cancel pending tasks."""
+        if not self._running:
+            return
+
+        self._running = False
+        if self._polling_task:
+            self._polling_task.cancel()
+            try:
+                await self._polling_task
+            except asyncio.CancelledError:
+                pass
+        logger.info("Stopped Telegram command polling")
+
+    async def _poll_loop(self) -> None:
+        """Main polling loop that fetches updates."""
+        while self._running:
+            try:
+                updates = await self._get_updates()
+                for update in updates:
+                    await self._handle_update(update)
+            except asyncio.CancelledError:
+                break
+            except Exception as exc:
+                logger.error("Error in polling loop: %s", exc)
+
+            await asyncio.sleep(self._polling_interval)
+
+    async def _get_updates(self) -> list[dict]:
+        """
+        Fetch updates from Telegram API.
+
+        Returns:
+            List of update objects
+        """
+        try:
+            url = f"{self._client.API_BASE.format(token=self._client._bot_token)}/getUpdates"
+            payload = {
+                "offset": self._last_update_id + 1,
+                "timeout": int(self._polling_interval),
+                "allowed_updates": ["message"],
+            }
+
+            session = self._client._get_session()
+            async with session.post(url, json=payload) as resp:
+                if resp.status != 200:
+                    error_text = await resp.text()
+                    if resp.status == 409:
+                        # Another bot instance is already polling — stop this poller entirely.
+                        # Retrying would keep conflicting with the other instance.
+                        self._running = False
+                        logger.warning(
+                            "Telegram conflict (409): another instance is already polling. "
+                            "Disabling Telegram commands for this process. "
+                            "Ensure only one instance of The Ouroboros is running at a time.",
+                        )
+                    else:
+                        logger.error(
+                            "getUpdates API error (status=%d): %s", resp.status, error_text
+                        )
+                    return []
+
+                data = await resp.json()
+                if not data.get("ok"):
+                    logger.error("getUpdates returned ok=false: %s", data)
+                    return []
+
+                updates = data.get("result", [])
+                if updates:
+                    self._last_update_id = updates[-1]["update_id"]
+
+                return updates
+
+        except asyncio.TimeoutError:
+            logger.debug("getUpdates timeout (normal)")
+            return []
+        except aiohttp.ClientError as exc:
+            logger.error("getUpdates failed: %s", exc)
+            return []
+        except Exception as exc:
+            logger.error("Unexpected error in _get_updates: %s", exc)
+            return []
+
+    async def _handle_update(self, update: dict) -> None:
+        """
+        Parse and handle a single update.
+
+        Args:
+            update: Update object from Telegram API
+        """
+        try:
+            message = update.get("message")
+            if not message:
+                return
+
+            # Verify chat_id matches configured chat
+            chat_id = str(message.get("chat", {}).get("id", ""))
+            if chat_id != self._client._chat_id:
+                logger.warning(
+                    "Ignoring command from unauthorized chat_id: %s", chat_id
+                )
+                return
+
+            # Extract command text
+            text = message.get("text", "").strip()
+            if not text.startswith("/"):
+                return
+
+            # Parse command (remove leading slash and extract command name)
+            command_parts = text[1:].split()
+            if not command_parts:
+                return
+
+            # Remove @botname suffix if present (for group chats)
+            command_name = command_parts[0].split("@")[0]
+
+            # Execute handler (args-aware handlers take priority)
+            args_handler = self._commands_with_args.get(command_name)
+            if args_handler:
+                logger.info("Executing command: /%s %s", command_name, command_parts[1:])
+                await args_handler(command_parts[1:])
+            elif command_name in self._commands:
+                logger.info("Executing command: /%s", command_name)
+                await self._commands[command_name]()
+            else:
+                logger.debug("Unknown command: /%s", command_name)
+                await self._client.send_message(
+                    f"Unknown command: /{command_name}\nUse /help to see available commands."
+                )
+
+        except Exception as exc:
+            logger.error("Error handling update: %s", exc)
+            # Don't crash the polling loop on handler errors

src/strategy/exit_rules.py

@@ -0,0 +1,104 @@
+"""Composite exit rules: hard stop, break-even lock, ATR trailing, model assist."""
+
+from __future__ import annotations
+
+from dataclasses import dataclass
+
+from src.strategy.position_state_machine import PositionState, StateTransitionInput, promote_state
+
+
+@dataclass(frozen=True)
+class ExitRuleConfig:
+    hard_stop_pct: float = -2.0
+    be_arm_pct: float = 1.2
+    arm_pct: float = 3.0
+    atr_multiplier_k: float = 2.2
+    model_prob_threshold: float = 0.62
+
+
+@dataclass(frozen=True)
+class ExitRuleInput:
+    current_price: float
+    entry_price: float
+    peak_price: float
+    atr_value: float = 0.0
+    pred_down_prob: float = 0.0
+    liquidity_weak: bool = False
+
+
+@dataclass(frozen=True)
+class ExitEvaluation:
+    state: PositionState
+    should_exit: bool
+    reason: str
+    unrealized_pnl_pct: float
+    trailing_stop_price: float | None
+
+
+def evaluate_exit(
+    *,
+    current_state: PositionState,
+    config: ExitRuleConfig,
+    inp: ExitRuleInput,
+) -> ExitEvaluation:
+    """Evaluate composite exit logic and return updated state."""
+    if inp.entry_price <= 0 or inp.current_price <= 0:
+        return ExitEvaluation(
+            state=current_state,
+            should_exit=False,
+            reason="invalid_price",
+            unrealized_pnl_pct=0.0,
+            trailing_stop_price=None,
+        )
+
+    unrealized = (inp.current_price - inp.entry_price) / inp.entry_price * 100.0
+    hard_stop_hit = unrealized <= config.hard_stop_pct
+    take_profit_hit = unrealized >= config.arm_pct
+
+    trailing_stop_price: float | None = None
+    trailing_stop_hit = False
+    if inp.atr_value > 0 and inp.peak_price > 0:
+        trailing_stop_price = inp.peak_price - (config.atr_multiplier_k * inp.atr_value)
+        trailing_stop_hit = inp.current_price <= trailing_stop_price
+
+    be_lock_threat = current_state in (PositionState.BE_LOCK, PositionState.ARMED) and (
+        inp.current_price <= inp.entry_price
+    )
+    model_exit_signal = inp.pred_down_prob >= config.model_prob_threshold and inp.liquidity_weak
+
+    next_state = promote_state(
+        current=current_state,
+        inp=StateTransitionInput(
+            unrealized_pnl_pct=unrealized,
+            be_arm_pct=config.be_arm_pct,
+            arm_pct=config.arm_pct,
+            hard_stop_hit=hard_stop_hit,
+            trailing_stop_hit=trailing_stop_hit,
+            model_exit_signal=model_exit_signal,
+            be_lock_threat=be_lock_threat,
+        ),
+    )
+
+    if hard_stop_hit:
+        reason = "hard_stop"
+    elif trailing_stop_hit:
+        reason = "atr_trailing_stop"
+    elif be_lock_threat:
+        reason = "be_lock_threat"
+    elif model_exit_signal:
+        reason = "model_liquidity_exit"
+    elif take_profit_hit:
+        # Backward-compatible immediate profit-taking path.
+        reason = "arm_take_profit"
+    else:
+        reason = "hold"
+
+    should_exit = next_state == PositionState.EXITED or take_profit_hit
+
+    return ExitEvaluation(
+        state=next_state,
+        should_exit=should_exit,
+        reason=reason,
+        unrealized_pnl_pct=unrealized,
+        trailing_stop_price=trailing_stop_price,
+    )

src/strategy/models.py

@@ -0,0 +1,184 @@
+"""Pydantic models for pre-market scenario planning.
+
+Defines the data contracts for the proactive strategy system:
+- AI generates DayPlaybook before market open (structured JSON scenarios)
+- Local ScenarioEngine matches conditions during market hours (no API calls)
+"""
+
+from __future__ import annotations
+
+from datetime import UTC, date, datetime
+from enum import Enum
+
+from pydantic import BaseModel, Field, field_validator
+
+
+class ScenarioAction(str, Enum):
+    """Actions that can be taken by scenarios."""
+
+    BUY = "BUY"
+    SELL = "SELL"
+    HOLD = "HOLD"
+    REDUCE_ALL = "REDUCE_ALL"
+
+
+class MarketOutlook(str, Enum):
+    """AI's assessment of market direction."""
+
+    BULLISH = "bullish"
+    NEUTRAL_TO_BULLISH = "neutral_to_bullish"
+    NEUTRAL = "neutral"
+    NEUTRAL_TO_BEARISH = "neutral_to_bearish"
+    BEARISH = "bearish"
+
+
+class PlaybookStatus(str, Enum):
+    """Lifecycle status of a playbook."""
+
+    PENDING = "pending"
+    READY = "ready"
+    FAILED = "failed"
+    EXPIRED = "expired"
+
+
+class StockCondition(BaseModel):
+    """Condition fields for scenario matching (all optional, AND-combined).
+
+    The ScenarioEngine evaluates all non-None fields as AND conditions.
+    A condition matches only if ALL specified fields are satisfied.
+
+    Technical indicator fields:
+        rsi_below / rsi_above — RSI threshold
+        volume_ratio_above / volume_ratio_below — volume vs previous day
+        price_above / price_below — absolute price level
+        price_change_pct_above / price_change_pct_below — intraday % change
+
+    Position-aware fields (require market_data enrichment from open position):
+        unrealized_pnl_pct_above — matches if unrealized P&L > threshold (e.g. 3.0 → +3%)
+        unrealized_pnl_pct_below — matches if unrealized P&L < threshold (e.g. -2.0 → -2%)
+        holding_days_above — matches if position held for more than N days
+        holding_days_below — matches if position held for fewer than N days
+    """
+
+    rsi_below: float | None = None
+    rsi_above: float | None = None
+    volume_ratio_above: float | None = None
+    volume_ratio_below: float | None = None
+    price_above: float | None = None
+    price_below: float | None = None
+    price_change_pct_above: float | None = None
+    price_change_pct_below: float | None = None
+    unrealized_pnl_pct_above: float | None = None
+    unrealized_pnl_pct_below: float | None = None
+    holding_days_above: int | None = None
+    holding_days_below: int | None = None
+
+    def has_any_condition(self) -> bool:
+        """Check if at least one condition field is set."""
+        return any(
+            v is not None
+            for v in (
+                self.rsi_below,
+                self.rsi_above,
+                self.volume_ratio_above,
+                self.volume_ratio_below,
+                self.price_above,
+                self.price_below,
+                self.price_change_pct_above,
+                self.price_change_pct_below,
+                self.unrealized_pnl_pct_above,
+                self.unrealized_pnl_pct_below,
+                self.holding_days_above,
+                self.holding_days_below,
+            )
+        )
+
+
+class StockScenario(BaseModel):
+    """A single condition-action rule for one stock."""
+
+    condition: StockCondition
+    action: ScenarioAction
+    confidence: int = Field(ge=0, le=100)
+    allocation_pct: float = Field(ge=0, le=100, default=10.0)
+    stop_loss_pct: float = Field(le=0, default=-2.0)
+    take_profit_pct: float = Field(ge=0, default=3.0)
+    rationale: str = ""
+
+
+class StockPlaybook(BaseModel):
+    """All scenarios for a single stock (ordered by priority)."""
+
+    stock_code: str
+    stock_name: str = ""
+    scenarios: list[StockScenario] = Field(min_length=1)
+
+
+class GlobalRule(BaseModel):
+    """Portfolio-level rule (checked before stock-level scenarios)."""
+
+    condition: str  # e.g. "portfolio_pnl_pct < -2.0"
+    action: ScenarioAction
+    rationale: str = ""
+
+
+class CrossMarketContext(BaseModel):
+    """Summary of another market's state for cross-market awareness."""
+
+    market: str  # e.g. "US" or "KR"
+    date: str
+    total_pnl: float = 0.0
+    win_rate: float = 0.0
+    index_change_pct: float = 0.0  # e.g. KOSPI or S&P500 change
+    key_events: list[str] = Field(default_factory=list)
+    lessons: list[str] = Field(default_factory=list)
+
+
+class DayPlaybook(BaseModel):
+    """Complete playbook for a single trading day in a single market.
+
+    Generated by PreMarketPlanner (1 Gemini call per market per day).
+    Consumed by ScenarioEngine during market hours (0 API calls).
+    """
+
+    date: date
+    market: str  # "KR" or "US"
+    market_outlook: MarketOutlook = MarketOutlook.NEUTRAL
+    generated_at: str = ""  # ISO timestamp
+    gemini_model: str = ""
+    token_count: int = 0
+    global_rules: list[GlobalRule] = Field(default_factory=list)
+    stock_playbooks: list[StockPlaybook] = Field(default_factory=list)
+    default_action: ScenarioAction = ScenarioAction.HOLD
+    context_summary: dict = Field(default_factory=dict)
+    cross_market: CrossMarketContext | None = None
+
+    @field_validator("stock_playbooks")
+    @classmethod
+    def validate_unique_stocks(cls, v: list[StockPlaybook]) -> list[StockPlaybook]:
+        codes = [pb.stock_code for pb in v]
+        if len(codes) != len(set(codes)):
+            raise ValueError("Duplicate stock codes in playbook")
+        return v
+
+    def get_stock_playbook(self, stock_code: str) -> StockPlaybook | None:
+        """Find the playbook for a specific stock."""
+        for pb in self.stock_playbooks:
+            if pb.stock_code == stock_code:
+                return pb
+        return None
+
+    @property
+    def scenario_count(self) -> int:
+        """Total number of scenarios across all stocks."""
+        return sum(len(pb.scenarios) for pb in self.stock_playbooks)
+
+    @property
+    def stock_count(self) -> int:
+        """Number of stocks with scenarios."""
+        return len(self.stock_playbooks)
+
+    def model_post_init(self, __context: object) -> None:
+        """Set generated_at if not provided."""
+        if not self.generated_at:
+            self.generated_at = datetime.now(UTC).isoformat()

src/strategy/playbook_store.py

@@ -0,0 +1,184 @@
+"""Playbook persistence layer — CRUD for DayPlaybook in SQLite.
+
+Stores and retrieves market-specific daily playbooks with JSON serialization.
+Designed for the pre-market strategy system (one playbook per market per day).
+"""
+
+from __future__ import annotations
+
+import json
+import logging
+import sqlite3
+from datetime import date
+
+from src.strategy.models import DayPlaybook, PlaybookStatus
+
+logger = logging.getLogger(__name__)
+
+
+class PlaybookStore:
+    """CRUD operations for DayPlaybook persistence."""
+
+    def __init__(self, conn: sqlite3.Connection) -> None:
+        self._conn = conn
+
+    def save(self, playbook: DayPlaybook) -> int:
+        """Save or replace a playbook for a given date+market.
+
+        Uses INSERT OR REPLACE to enforce UNIQUE(date, market).
+
+        Returns:
+            The row id of the inserted/replaced record.
+        """
+        playbook_json = playbook.model_dump_json()
+        cursor = self._conn.execute(
+            """
+            INSERT OR REPLACE INTO playbooks
+                (date, market, status, playbook_json, generated_at,
+                 token_count, scenario_count, match_count)
+            VALUES (?, ?, ?, ?, ?, ?, ?, ?)
+            """,
+            (
+                playbook.date.isoformat(),
+                playbook.market,
+                PlaybookStatus.READY.value,
+                playbook_json,
+                playbook.generated_at,
+                playbook.token_count,
+                playbook.scenario_count,
+                0,
+            ),
+        )
+        self._conn.commit()
+        row_id = cursor.lastrowid or 0
+        logger.info(
+            "Saved playbook for %s/%s (%d stocks, %d scenarios)",
+            playbook.date, playbook.market,
+            playbook.stock_count, playbook.scenario_count,
+        )
+        return row_id
+
+    def load(self, target_date: date, market: str) -> DayPlaybook | None:
+        """Load a playbook for a specific date and market.
+
+        Returns:
+            DayPlaybook if found, None otherwise.
+        """
+        row = self._conn.execute(
+            "SELECT playbook_json FROM playbooks WHERE date = ? AND market = ?",
+            (target_date.isoformat(), market),
+        ).fetchone()
+        if row is None:
+            return None
+        return DayPlaybook.model_validate_json(row[0])
+
+    def get_status(self, target_date: date, market: str) -> PlaybookStatus | None:
+        """Get the status of a playbook without deserializing the full JSON."""
+        row = self._conn.execute(
+            "SELECT status FROM playbooks WHERE date = ? AND market = ?",
+            (target_date.isoformat(), market),
+        ).fetchone()
+        if row is None:
+            return None
+        return PlaybookStatus(row[0])
+
+    def update_status(self, target_date: date, market: str, status: PlaybookStatus) -> bool:
+        """Update the status of a playbook.
+
+        Returns:
+            True if a row was updated, False if not found.
+        """
+        cursor = self._conn.execute(
+            "UPDATE playbooks SET status = ? WHERE date = ? AND market = ?",
+            (status.value, target_date.isoformat(), market),
+        )
+        self._conn.commit()
+        return cursor.rowcount > 0
+
+    def increment_match_count(self, target_date: date, market: str) -> bool:
+        """Increment the match_count for tracking scenario hits during the day.
+
+        Returns:
+            True if a row was updated, False if not found.
+        """
+        cursor = self._conn.execute(
+            "UPDATE playbooks SET match_count = match_count + 1 WHERE date = ? AND market = ?",
+            (target_date.isoformat(), market),
+        )
+        self._conn.commit()
+        return cursor.rowcount > 0
+
+    def get_stats(self, target_date: date, market: str) -> dict | None:
+        """Get playbook stats without full deserialization.
+
+        Returns:
+            Dict with status, token_count, scenario_count, match_count, or None.
+        """
+        row = self._conn.execute(
+            """
+            SELECT status, token_count, scenario_count, match_count, generated_at
+            FROM playbooks WHERE date = ? AND market = ?
+            """,
+            (target_date.isoformat(), market),
+        ).fetchone()
+        if row is None:
+            return None
+        return {
+            "status": row[0],
+            "token_count": row[1],
+            "scenario_count": row[2],
+            "match_count": row[3],
+            "generated_at": row[4],
+        }
+
+    def list_recent(self, market: str | None = None, limit: int = 7) -> list[dict]:
+        """List recent playbooks with summary info.
+
+        Args:
+            market: Filter by market code. None for all markets.
+            limit: Max number of results.
+
+        Returns:
+            List of dicts with date, market, status, scenario_count, match_count.
+        """
+        if market is not None:
+            rows = self._conn.execute(
+                """
+                SELECT date, market, status, scenario_count, match_count
+                FROM playbooks WHERE market = ?
+                ORDER BY date DESC LIMIT ?
+                """,
+                (market, limit),
+            ).fetchall()
+        else:
+            rows = self._conn.execute(
+                """
+                SELECT date, market, status, scenario_count, match_count
+                FROM playbooks
+                ORDER BY date DESC LIMIT ?
+                """,
+                (limit,),
+            ).fetchall()
+        return [
+            {
+                "date": row[0],
+                "market": row[1],
+                "status": row[2],
+                "scenario_count": row[3],
+                "match_count": row[4],
+            }
+            for row in rows
+        ]
+
+    def delete(self, target_date: date, market: str) -> bool:
+        """Delete a playbook.
+
+        Returns:
+            True if a row was deleted, False if not found.
+        """
+        cursor = self._conn.execute(
+            "DELETE FROM playbooks WHERE date = ? AND market = ?",
+            (target_date.isoformat(), market),
+        )
+        self._conn.commit()
+        return cursor.rowcount > 0

src/strategy/position_state_machine.py

@@ -0,0 +1,70 @@
+"""Position state machine for staged exit control.
+
+State progression is monotonic (promotion-only) except terminal EXITED.
+"""
+
+from __future__ import annotations
+
+from dataclasses import dataclass
+from enum import Enum
+
+
+class PositionState(str, Enum):
+    HOLDING = "HOLDING"
+    BE_LOCK = "BE_LOCK"
+    ARMED = "ARMED"
+    EXITED = "EXITED"
+
+
+_STATE_RANK: dict[PositionState, int] = {
+    PositionState.HOLDING: 0,
+    PositionState.BE_LOCK: 1,
+    PositionState.ARMED: 2,
+    PositionState.EXITED: 3,
+}
+
+
+@dataclass(frozen=True)
+class StateTransitionInput:
+    unrealized_pnl_pct: float
+    be_arm_pct: float
+    arm_pct: float
+    hard_stop_hit: bool = False
+    trailing_stop_hit: bool = False
+    model_exit_signal: bool = False
+    be_lock_threat: bool = False
+
+
+def evaluate_exit_first(inp: StateTransitionInput) -> bool:
+    """Return True when terminal exit conditions are met.
+
+    EXITED must be evaluated before any promotion.
+    """
+    return (
+        inp.hard_stop_hit
+        or inp.trailing_stop_hit
+        or inp.model_exit_signal
+        or inp.be_lock_threat
+    )
+
+
+def promote_state(current: PositionState, inp: StateTransitionInput) -> PositionState:
+    """Promote to highest admissible state for current tick/bar.
+
+    Rules:
+    - EXITED has highest precedence and is terminal.
+    - Promotions are monotonic (no downgrade).
+    """
+    if current == PositionState.EXITED:
+        return PositionState.EXITED
+
+    if evaluate_exit_first(inp):
+        return PositionState.EXITED
+
+    target = PositionState.HOLDING
+    if inp.unrealized_pnl_pct >= inp.arm_pct:
+        target = PositionState.ARMED
+    elif inp.unrealized_pnl_pct >= inp.be_arm_pct:
+        target = PositionState.BE_LOCK
+
+    return target if _STATE_RANK[target] > _STATE_RANK[current] else current

src/strategy/pre_market_planner.py

@@ -0,0 +1,620 @@
+"""Pre-market planner — generates DayPlaybook via Gemini before market open.
+
+One Gemini API call per market per day. Candidates come from SmartVolatilityScanner.
+On failure, returns a smart rule-based fallback playbook that uses scanner signals
+(momentum/oversold) to generate BUY conditions, avoiding the all-HOLD problem.
+"""
+
+from __future__ import annotations
+
+import json
+import logging
+from datetime import date, timedelta
+from typing import Any
+
+from src.analysis.smart_scanner import ScanCandidate
+from src.brain.context_selector import ContextSelector, DecisionType
+from src.brain.gemini_client import GeminiClient
+from src.config import Settings
+from src.context.store import ContextLayer, ContextStore
+from src.strategy.models import (
+    CrossMarketContext,
+    DayPlaybook,
+    GlobalRule,
+    MarketOutlook,
+    ScenarioAction,
+    StockCondition,
+    StockPlaybook,
+    StockScenario,
+)
+
+logger = logging.getLogger(__name__)
+
+# Mapping from string to MarketOutlook enum
+_OUTLOOK_MAP: dict[str, MarketOutlook] = {
+    "bullish": MarketOutlook.BULLISH,
+    "neutral_to_bullish": MarketOutlook.NEUTRAL_TO_BULLISH,
+    "neutral": MarketOutlook.NEUTRAL,
+    "neutral_to_bearish": MarketOutlook.NEUTRAL_TO_BEARISH,
+    "bearish": MarketOutlook.BEARISH,
+}
+
+_ACTION_MAP: dict[str, ScenarioAction] = {
+    "BUY": ScenarioAction.BUY,
+    "SELL": ScenarioAction.SELL,
+    "HOLD": ScenarioAction.HOLD,
+    "REDUCE_ALL": ScenarioAction.REDUCE_ALL,
+}
+
+
+class PreMarketPlanner:
+    """Generates a DayPlaybook by calling Gemini once before market open.
+
+    Flow:
+    1. Collect strategic context (L5-L7) + cross-market context
+    2. Build a structured prompt with scan candidates
+    3. Call Gemini for JSON scenario generation
+    4. Parse and validate response into DayPlaybook
+    5. On failure → defensive playbook (HOLD everything)
+    """
+
+    def __init__(
+        self,
+        gemini_client: GeminiClient,
+        context_store: ContextStore,
+        context_selector: ContextSelector,
+        settings: Settings,
+    ) -> None:
+        self._gemini = gemini_client
+        self._context_store = context_store
+        self._context_selector = context_selector
+        self._settings = settings
+
+    async def generate_playbook(
+        self,
+        market: str,
+        candidates: list[ScanCandidate],
+        today: date | None = None,
+        current_holdings: list[dict] | None = None,
+    ) -> DayPlaybook:
+        """Generate a DayPlaybook for a market using Gemini.
+
+        Args:
+            market: Market code ("KR" or "US")
+            candidates: Stock candidates from SmartVolatilityScanner
+            today: Override date (defaults to date.today()). Use market-local date.
+            current_holdings: Currently held positions with entry_price and unrealized_pnl_pct.
+                Each dict: {"stock_code": str, "name": str, "qty": int,
+                            "entry_price": float, "unrealized_pnl_pct": float,
+                            "holding_days": int}
+
+        Returns:
+            DayPlaybook with scenarios. Empty/defensive if no candidates or failure.
+        """
+        if today is None:
+            today = date.today()
+
+        if not candidates:
+            logger.info("No candidates for %s — returning empty playbook", market)
+            return self._empty_playbook(today, market)
+
+        try:
+            # 1. Gather context
+            context_data = self._gather_context()
+            self_market_scorecard = self.build_self_market_scorecard(market, today)
+            cross_market = self.build_cross_market_context(market, today)
+
+            # 2. Build prompt
+            prompt = self._build_prompt(
+                market,
+                candidates,
+                context_data,
+                self_market_scorecard,
+                cross_market,
+                current_holdings=current_holdings,
+            )
+
+            # 3. Call Gemini
+            market_data = {
+                "stock_code": "PLANNER",
+                "current_price": 0,
+                "prompt_override": prompt,
+            }
+            decision = await self._gemini.decide(market_data)
+
+            # 4. Parse response
+            playbook = self._parse_response(
+                decision.rationale, today, market, candidates, cross_market,
+                current_holdings=current_holdings,
+            )
+            playbook_with_tokens = playbook.model_copy(
+                update={"token_count": decision.token_count}
+            )
+            logger.info(
+                "Generated playbook for %s: %d stocks, %d scenarios, %d tokens",
+                market,
+                playbook_with_tokens.stock_count,
+                playbook_with_tokens.scenario_count,
+                playbook_with_tokens.token_count,
+            )
+            return playbook_with_tokens
+
+        except Exception:
+            logger.exception("Playbook generation failed for %s", market)
+            if self._settings.DEFENSIVE_PLAYBOOK_ON_FAILURE:
+                return self._smart_fallback_playbook(today, market, candidates, self._settings)
+            return self._empty_playbook(today, market)
+
+    def build_cross_market_context(
+        self, target_market: str, today: date | None = None,
+    ) -> CrossMarketContext | None:
+        """Build cross-market context from the other market's L6 data.
+
+        KR planner → reads US scorecard from previous night.
+        US planner → reads KR scorecard from today.
+
+        Args:
+            target_market: The market being planned ("KR" or "US")
+            today: Override date (defaults to date.today()). Use market-local date.
+        """
+        other_market = "US" if target_market == "KR" else "KR"
+        if today is None:
+            today = date.today()
+        timeframe_date = today - timedelta(days=1) if target_market == "KR" else today
+        timeframe = timeframe_date.isoformat()
+
+        scorecard_key = f"scorecard_{other_market}"
+        scorecard_data = self._context_store.get_context(
+            ContextLayer.L6_DAILY, timeframe, scorecard_key
+        )
+
+        if scorecard_data is None:
+            logger.debug("No cross-market scorecard found for %s", other_market)
+            return None
+
+        if isinstance(scorecard_data, str):
+            try:
+                scorecard_data = json.loads(scorecard_data)
+            except (json.JSONDecodeError, TypeError):
+                return None
+
+        if not isinstance(scorecard_data, dict):
+            return None
+
+        return CrossMarketContext(
+            market=other_market,
+            date=timeframe,
+            total_pnl=float(scorecard_data.get("total_pnl", 0.0)),
+            win_rate=float(scorecard_data.get("win_rate", 0.0)),
+            index_change_pct=float(scorecard_data.get("index_change_pct", 0.0)),
+            key_events=scorecard_data.get("key_events", []),
+            lessons=scorecard_data.get("lessons", []),
+        )
+
+    def build_self_market_scorecard(
+        self, market: str, today: date | None = None,
+    ) -> dict[str, Any] | None:
+        """Build previous-day scorecard for the same market."""
+        if today is None:
+            today = date.today()
+        timeframe = (today - timedelta(days=1)).isoformat()
+        scorecard_key = f"scorecard_{market}"
+        scorecard_data = self._context_store.get_context(
+            ContextLayer.L6_DAILY, timeframe, scorecard_key
+        )
+
+        if scorecard_data is None:
+            return None
+
+        if isinstance(scorecard_data, str):
+            try:
+                scorecard_data = json.loads(scorecard_data)
+            except (json.JSONDecodeError, TypeError):
+                return None
+
+        if not isinstance(scorecard_data, dict):
+            return None
+
+        return {
+            "date": timeframe,
+            "total_pnl": float(scorecard_data.get("total_pnl", 0.0)),
+            "win_rate": float(scorecard_data.get("win_rate", 0.0)),
+            "lessons": scorecard_data.get("lessons", []),
+        }
+
+    def _gather_context(self) -> dict[str, Any]:
+        """Gather strategic context using ContextSelector."""
+        layers = self._context_selector.select_layers(
+            decision_type=DecisionType.STRATEGIC,
+            include_realtime=True,
+        )
+        return self._context_selector.get_context_data(layers, max_items_per_layer=10)
+
+    def _build_prompt(
+        self,
+        market: str,
+        candidates: list[ScanCandidate],
+        context_data: dict[str, Any],
+        self_market_scorecard: dict[str, Any] | None,
+        cross_market: CrossMarketContext | None,
+        current_holdings: list[dict] | None = None,
+    ) -> str:
+        """Build a structured prompt for Gemini to generate scenario JSON."""
+        max_scenarios = self._settings.MAX_SCENARIOS_PER_STOCK
+
+        candidates_text = "\n".join(
+            f"  - {c.stock_code} ({c.name}): price={c.price}, "
+            f"RSI={c.rsi:.1f}, volume_ratio={c.volume_ratio:.1f}, "
+            f"signal={c.signal}, score={c.score:.1f}"
+            for c in candidates
+        )
+
+        holdings_text = ""
+        if current_holdings:
+            lines = []
+            for h in current_holdings:
+                code = h.get("stock_code", "")
+                name = h.get("name", "")
+                qty = h.get("qty", 0)
+                entry_price = h.get("entry_price", 0.0)
+                pnl_pct = h.get("unrealized_pnl_pct", 0.0)
+                holding_days = h.get("holding_days", 0)
+                lines.append(
+                    f"  - {code} ({name}): {qty}주 @ {entry_price:,.0f}, "
+                    f"미실현손익 {pnl_pct:+.2f}%, 보유 {holding_days}일"
+                )
+            holdings_text = (
+                "\n## Current Holdings (보유 중 — SELL/HOLD 전략 고려 필요)\n"
+                + "\n".join(lines)
+                + "\n"
+            )
+
+        cross_market_text = ""
+        if cross_market:
+            cross_market_text = (
+                f"\n## Other Market ({cross_market.market}) Summary\n"
+                f"- P&L: {cross_market.total_pnl:+.2f}%\n"
+                f"- Win Rate: {cross_market.win_rate:.0f}%\n"
+                f"- Index Change: {cross_market.index_change_pct:+.2f}%\n"
+            )
+            if cross_market.lessons:
+                cross_market_text += f"- Lessons: {'; '.join(cross_market.lessons[:3])}\n"
+
+        self_market_text = ""
+        if self_market_scorecard:
+            self_market_text = (
+                f"\n## My Market Previous Day ({market})\n"
+                f"- Date: {self_market_scorecard['date']}\n"
+                f"- P&L: {self_market_scorecard['total_pnl']:+.2f}%\n"
+                f"- Win Rate: {self_market_scorecard['win_rate']:.0f}%\n"
+            )
+            lessons = self_market_scorecard.get("lessons", [])
+            if lessons:
+                self_market_text += f"- Lessons: {'; '.join(lessons[:3])}\n"
+
+        context_text = ""
+        if context_data:
+            context_text = "\n## Strategic Context\n"
+            for layer_name, layer_data in context_data.items():
+                if layer_data:
+                    context_text += f"### {layer_name}\n"
+                    for key, value in list(layer_data.items())[:5]:
+                        context_text += f"  - {key}: {value}\n"
+
+        holdings_instruction = ""
+        if current_holdings:
+            holding_codes = [h.get("stock_code", "") for h in current_holdings]
+            holdings_instruction = (
+                f"- Also include SELL/HOLD scenarios for held stocks: "
+                f"{', '.join(holding_codes)} "
+                f"(even if not in candidates list)\n"
+            )
+
+        return (
+            f"You are a pre-market trading strategist for the {market} market.\n"
+            f"Generate structured trading scenarios for today.\n\n"
+            f"## Candidates (from volatility scanner)\n{candidates_text}\n"
+            f"{holdings_text}"
+            f"{self_market_text}"
+            f"{cross_market_text}"
+            f"{context_text}\n"
+            f"## Instructions\n"
+            f"Return a JSON object with this exact structure:\n"
+            f'{{\n'
+            f'  "market_outlook": "bullish|neutral_to_bullish|neutral'
+            f'|neutral_to_bearish|bearish",\n'
+            f'  "global_rules": [\n'
+            f'    {{"condition": "portfolio_pnl_pct < -2.0",'
+            f' "action": "REDUCE_ALL", "rationale": "..."}}\n'
+            f'  ],\n'
+            f'  "stocks": [\n'
+            f'    {{\n'
+            f'      "stock_code": "...",\n'
+            f'      "scenarios": [\n'
+            f'        {{\n'
+            f'          "condition": {{"rsi_below": 30, "volume_ratio_above": 2.0,'
+            f' "unrealized_pnl_pct_above": 3.0, "holding_days_above": 5}},\n'
+            f'          "action": "BUY|SELL|HOLD",\n'
+            f'          "confidence": 85,\n'
+            f'          "allocation_pct": 10.0,\n'
+            f'          "stop_loss_pct": -2.0,\n'
+            f'          "take_profit_pct": 3.0,\n'
+            f'          "rationale": "..."\n'
+            f'        }}\n'
+            f'      ]\n'
+            f'    }}\n'
+            f'  ]\n'
+            f'}}\n\n'
+            f"Rules:\n"
+            f"- Max {max_scenarios} scenarios per stock\n"
+            f"- Candidates list is the primary source for BUY candidates\n"
+            f"{holdings_instruction}"
+            f"- Confidence 0-100 (80+ for actionable trades)\n"
+            f"- stop_loss_pct must be <= 0, take_profit_pct must be >= 0\n"
+            f"- Return ONLY the JSON, no markdown fences or explanation\n"
+        )
+
+    def _parse_response(
+        self,
+        response_text: str,
+        today: date,
+        market: str,
+        candidates: list[ScanCandidate],
+        cross_market: CrossMarketContext | None,
+        current_holdings: list[dict] | None = None,
+    ) -> DayPlaybook:
+        """Parse Gemini's JSON response into a validated DayPlaybook."""
+        cleaned = self._extract_json(response_text)
+        data = json.loads(cleaned)
+
+        valid_codes = {c.stock_code for c in candidates}
+        # Holdings are also valid — AI may generate SELL/HOLD scenarios for them
+        if current_holdings:
+            for h in current_holdings:
+                code = h.get("stock_code", "")
+                if code:
+                    valid_codes.add(code)
+
+        # Parse market outlook
+        outlook_str = data.get("market_outlook", "neutral")
+        market_outlook = _OUTLOOK_MAP.get(outlook_str, MarketOutlook.NEUTRAL)
+
+        # Parse global rules
+        global_rules = []
+        for rule_data in data.get("global_rules", []):
+            action_str = rule_data.get("action", "HOLD")
+            action = _ACTION_MAP.get(action_str, ScenarioAction.HOLD)
+            global_rules.append(
+                GlobalRule(
+                    condition=rule_data.get("condition", ""),
+                    action=action,
+                    rationale=rule_data.get("rationale", ""),
+                )
+            )
+
+        # Parse stock playbooks
+        stock_playbooks = []
+        max_scenarios = self._settings.MAX_SCENARIOS_PER_STOCK
+        for stock_data in data.get("stocks", []):
+            code = stock_data.get("stock_code", "")
+            if code not in valid_codes:
+                logger.warning("Gemini returned unknown stock %s — skipping", code)
+                continue
+
+            scenarios = []
+            for sc_data in stock_data.get("scenarios", [])[:max_scenarios]:
+                scenario = self._parse_scenario(sc_data)
+                if scenario:
+                    scenarios.append(scenario)
+
+            if scenarios:
+                stock_playbooks.append(
+                    StockPlaybook(
+                        stock_code=code,
+                        scenarios=scenarios,
+                    )
+                )
+
+        return DayPlaybook(
+            date=today,
+            market=market,
+            market_outlook=market_outlook,
+            global_rules=global_rules,
+            stock_playbooks=stock_playbooks,
+            cross_market=cross_market,
+        )
+
+    def _parse_scenario(self, sc_data: dict) -> StockScenario | None:
+        """Parse a single scenario from JSON data. Returns None if invalid."""
+        try:
+            cond_data = sc_data.get("condition", {})
+            condition = StockCondition(
+                rsi_below=cond_data.get("rsi_below"),
+                rsi_above=cond_data.get("rsi_above"),
+                volume_ratio_above=cond_data.get("volume_ratio_above"),
+                volume_ratio_below=cond_data.get("volume_ratio_below"),
+                price_above=cond_data.get("price_above"),
+                price_below=cond_data.get("price_below"),
+                price_change_pct_above=cond_data.get("price_change_pct_above"),
+                price_change_pct_below=cond_data.get("price_change_pct_below"),
+                unrealized_pnl_pct_above=cond_data.get("unrealized_pnl_pct_above"),
+                unrealized_pnl_pct_below=cond_data.get("unrealized_pnl_pct_below"),
+                holding_days_above=cond_data.get("holding_days_above"),
+                holding_days_below=cond_data.get("holding_days_below"),
+            )
+
+            if not condition.has_any_condition():
+                logger.warning("Scenario has no conditions — skipping")
+                return None
+
+            action_str = sc_data.get("action", "HOLD")
+            action = _ACTION_MAP.get(action_str, ScenarioAction.HOLD)
+
+            return StockScenario(
+                condition=condition,
+                action=action,
+                confidence=int(sc_data.get("confidence", 50)),
+                allocation_pct=float(sc_data.get("allocation_pct", 10.0)),
+                stop_loss_pct=float(sc_data.get("stop_loss_pct", -2.0)),
+                take_profit_pct=float(sc_data.get("take_profit_pct", 3.0)),
+                rationale=sc_data.get("rationale", ""),
+            )
+        except (ValueError, TypeError) as e:
+            logger.warning("Failed to parse scenario: %s", e)
+            return None
+
+    @staticmethod
+    def _extract_json(text: str) -> str:
+        """Extract JSON from response, stripping markdown fences if present."""
+        stripped = text.strip()
+        if stripped.startswith("```"):
+            # Remove first line (```json or ```) and last line (```)
+            lines = stripped.split("\n")
+            lines = lines[1:]  # Remove opening fence
+            if lines and lines[-1].strip() == "```":
+                lines = lines[:-1]
+            stripped = "\n".join(lines)
+        return stripped.strip()
+
+    @staticmethod
+    def _empty_playbook(today: date, market: str) -> DayPlaybook:
+        """Return an empty playbook (no stocks, no scenarios)."""
+        return DayPlaybook(
+            date=today,
+            market=market,
+            market_outlook=MarketOutlook.NEUTRAL,
+            stock_playbooks=[],
+        )
+
+    @staticmethod
+    def _defensive_playbook(
+        today: date,
+        market: str,
+        candidates: list[ScanCandidate],
+    ) -> DayPlaybook:
+        """Return a defensive playbook — HOLD everything with stop-loss ready."""
+        stock_playbooks = [
+            StockPlaybook(
+                stock_code=c.stock_code,
+                scenarios=[
+                    StockScenario(
+                        condition=StockCondition(price_change_pct_below=-3.0),
+                        action=ScenarioAction.SELL,
+                        confidence=90,
+                        stop_loss_pct=-3.0,
+                        rationale="Defensive stop-loss (planner failure)",
+                    ),
+                ],
+            )
+            for c in candidates
+        ]
+        return DayPlaybook(
+            date=today,
+            market=market,
+            market_outlook=MarketOutlook.NEUTRAL_TO_BEARISH,
+            default_action=ScenarioAction.HOLD,
+            stock_playbooks=stock_playbooks,
+            global_rules=[
+                GlobalRule(
+                    condition="portfolio_pnl_pct < -2.0",
+                    action=ScenarioAction.REDUCE_ALL,
+                    rationale="Defensive: reduce on loss threshold",
+                ),
+            ],
+        )
+
+    @staticmethod
+    def _smart_fallback_playbook(
+        today: date,
+        market: str,
+        candidates: list[ScanCandidate],
+        settings: Settings,
+    ) -> DayPlaybook:
+        """Rule-based fallback playbook when Gemini is unavailable.
+
+        Uses scanner signals (RSI, volume_ratio) to generate meaningful BUY
+        conditions instead of the all-SELL defensive playbook.  Candidates are
+        already pre-qualified by SmartVolatilityScanner, so we trust their
+        signals and build actionable scenarios from them.
+
+        Scenario logic per candidate:
+        - momentum signal: BUY when volume_ratio exceeds scanner threshold
+        - oversold signal: BUY when RSI is below oversold threshold
+        - always: SELL stop-loss at -3.0% as guard
+        """
+        stock_playbooks = []
+        for c in candidates:
+            scenarios: list[StockScenario] = []
+
+            if c.signal == "momentum":
+                scenarios.append(
+                    StockScenario(
+                        condition=StockCondition(
+                            volume_ratio_above=settings.VOL_MULTIPLIER,
+                        ),
+                        action=ScenarioAction.BUY,
+                        confidence=80,
+                        allocation_pct=10.0,
+                        stop_loss_pct=-3.0,
+                        take_profit_pct=5.0,
+                        rationale=(
+                            f"Rule-based BUY: momentum signal, "
+                            f"volume={c.volume_ratio:.1f}x (fallback planner)"
+                        ),
+                    )
+                )
+            elif c.signal == "oversold":
+                scenarios.append(
+                    StockScenario(
+                        condition=StockCondition(
+                            rsi_below=settings.RSI_OVERSOLD_THRESHOLD,
+                        ),
+                        action=ScenarioAction.BUY,
+                        confidence=80,
+                        allocation_pct=10.0,
+                        stop_loss_pct=-3.0,
+                        take_profit_pct=5.0,
+                        rationale=(
+                            f"Rule-based BUY: oversold signal, "
+                            f"RSI={c.rsi:.0f} (fallback planner)"
+                        ),
+                    )
+                )
+
+            # Always add stop-loss guard
+            scenarios.append(
+                StockScenario(
+                    condition=StockCondition(price_change_pct_below=-3.0),
+                    action=ScenarioAction.SELL,
+                    confidence=90,
+                    stop_loss_pct=-3.0,
+                    rationale="Rule-based stop-loss (fallback planner)",
+                )
+            )
+
+            stock_playbooks.append(
+                StockPlaybook(
+                    stock_code=c.stock_code,
+                    scenarios=scenarios,
+                )
+            )
+
+        logger.info(
+            "Smart fallback playbook for %s: %d stocks with rule-based BUY/SELL conditions",
+            market,
+            len(stock_playbooks),
+        )
+        return DayPlaybook(
+            date=today,
+            market=market,
+            market_outlook=MarketOutlook.NEUTRAL,
+            default_action=ScenarioAction.HOLD,
+            stock_playbooks=stock_playbooks,
+            global_rules=[
+                GlobalRule(
+                    condition="portfolio_pnl_pct < -2.0",
+                    action=ScenarioAction.REDUCE_ALL,
+                    rationale="Defensive: reduce on loss threshold",
+                ),
+            ],
+        )

src/strategy/scenario_engine.py

@@ -0,0 +1,305 @@
+"""Local scenario engine for playbook execution.
+
+Matches real-time market conditions against pre-defined scenarios
+without any API calls. Designed for sub-100ms execution.
+"""
+
+from __future__ import annotations
+
+import logging
+from dataclasses import dataclass, field
+from typing import Any
+
+from src.strategy.models import (
+    DayPlaybook,
+    GlobalRule,
+    ScenarioAction,
+    StockCondition,
+    StockScenario,
+)
+
+logger = logging.getLogger(__name__)
+
+
+@dataclass
+class ScenarioMatch:
+    """Result of matching market conditions against scenarios."""
+
+    stock_code: str
+    matched_scenario: StockScenario | None
+    action: ScenarioAction
+    confidence: int
+    rationale: str
+    global_rule_triggered: GlobalRule | None = None
+    match_details: dict[str, Any] = field(default_factory=dict)
+
+
+class ScenarioEngine:
+    """Evaluates playbook scenarios against real-time market data.
+
+    No API calls — pure Python condition matching.
+
+    Expected market_data keys: "rsi", "volume_ratio", "current_price", "price_change_pct".
+    Callers must normalize data source keys to match this contract.
+    """
+
+    def __init__(self) -> None:
+        self._warned_keys: set[str] = set()
+
+    @staticmethod
+    def _safe_float(value: Any) -> float | None:
+        """Safely cast a value to float. Returns None on failure."""
+        if value is None:
+            return None
+        try:
+            return float(value)
+        except (ValueError, TypeError):
+            return None
+
+    def _warn_missing_key(self, key: str) -> None:
+        """Log a missing-key warning once per key per engine instance."""
+        if key not in self._warned_keys:
+            self._warned_keys.add(key)
+            logger.warning("Condition requires '%s' but key missing from market_data", key)
+
+    def evaluate(
+        self,
+        playbook: DayPlaybook,
+        stock_code: str,
+        market_data: dict[str, Any],
+        portfolio_data: dict[str, Any],
+    ) -> ScenarioMatch:
+        """Match market conditions to scenarios and return a decision.
+
+        Algorithm:
+        1. Check global rules first (portfolio-level circuit breakers)
+        2. Find the StockPlaybook for the given stock_code
+        3. Iterate scenarios in order (first match wins)
+        4. If no match, return playbook.default_action (HOLD)
+
+        Args:
+            playbook: Today's DayPlaybook for this market
+            stock_code: Stock ticker to evaluate
+            market_data: Real-time market data (price, rsi, volume_ratio, etc.)
+            portfolio_data: Portfolio state (pnl_pct, total_cash, etc.)
+
+        Returns:
+            ScenarioMatch with the decision
+        """
+        # 1. Check global rules
+        triggered_rule = self.check_global_rules(playbook, portfolio_data)
+        if triggered_rule is not None:
+            logger.info(
+                "Global rule triggered for %s: %s -> %s",
+                stock_code,
+                triggered_rule.condition,
+                triggered_rule.action.value,
+            )
+            return ScenarioMatch(
+                stock_code=stock_code,
+                matched_scenario=None,
+                action=triggered_rule.action,
+                confidence=100,
+                rationale=f"Global rule: {triggered_rule.rationale or triggered_rule.condition}",
+                global_rule_triggered=triggered_rule,
+            )
+
+        # 2. Find stock playbook
+        stock_pb = playbook.get_stock_playbook(stock_code)
+        if stock_pb is None:
+            logger.debug("No playbook for %s — defaulting to %s", stock_code, playbook.default_action)
+            return ScenarioMatch(
+                stock_code=stock_code,
+                matched_scenario=None,
+                action=playbook.default_action,
+                confidence=0,
+                rationale=f"No scenarios defined for {stock_code}",
+            )
+
+        # 3. Iterate scenarios (first match wins)
+        for scenario in stock_pb.scenarios:
+            if self.evaluate_condition(scenario.condition, market_data):
+                logger.info(
+                    "Scenario matched for %s: %s (confidence=%d)",
+                    stock_code,
+                    scenario.action.value,
+                    scenario.confidence,
+                )
+                return ScenarioMatch(
+                    stock_code=stock_code,
+                    matched_scenario=scenario,
+                    action=scenario.action,
+                    confidence=scenario.confidence,
+                    rationale=scenario.rationale,
+                    match_details=self._build_match_details(scenario.condition, market_data),
+                )
+
+        # 4. No match — default action
+        logger.debug("No scenario matched for %s — defaulting to %s", stock_code, playbook.default_action)
+        return ScenarioMatch(
+            stock_code=stock_code,
+            matched_scenario=None,
+            action=playbook.default_action,
+            confidence=0,
+            rationale="No scenario conditions met — holding position",
+        )
+
+    def check_global_rules(
+        self,
+        playbook: DayPlaybook,
+        portfolio_data: dict[str, Any],
+    ) -> GlobalRule | None:
+        """Check portfolio-level rules. Returns first triggered rule or None."""
+        for rule in playbook.global_rules:
+            if self._evaluate_global_condition(rule.condition, portfolio_data):
+                return rule
+        return None
+
+    def evaluate_condition(
+        self,
+        condition: StockCondition,
+        market_data: dict[str, Any],
+    ) -> bool:
+        """Evaluate all non-None fields in condition as AND.
+
+        Returns True only if ALL specified conditions are met.
+        Empty condition (no fields set) returns False for safety.
+        """
+        if not condition.has_any_condition():
+            return False
+
+        checks: list[bool] = []
+
+        rsi = self._safe_float(market_data.get("rsi"))
+        if condition.rsi_below is not None or condition.rsi_above is not None:
+            if "rsi" not in market_data:
+                self._warn_missing_key("rsi")
+        if condition.rsi_below is not None:
+            checks.append(rsi is not None and rsi < condition.rsi_below)
+        if condition.rsi_above is not None:
+            checks.append(rsi is not None and rsi > condition.rsi_above)
+
+        volume_ratio = self._safe_float(market_data.get("volume_ratio"))
+        if condition.volume_ratio_above is not None or condition.volume_ratio_below is not None:
+            if "volume_ratio" not in market_data:
+                self._warn_missing_key("volume_ratio")
+        if condition.volume_ratio_above is not None:
+            checks.append(volume_ratio is not None and volume_ratio > condition.volume_ratio_above)
+        if condition.volume_ratio_below is not None:
+            checks.append(volume_ratio is not None and volume_ratio < condition.volume_ratio_below)
+
+        price = self._safe_float(market_data.get("current_price"))
+        if condition.price_above is not None or condition.price_below is not None:
+            if "current_price" not in market_data:
+                self._warn_missing_key("current_price")
+        if condition.price_above is not None:
+            checks.append(price is not None and price > condition.price_above)
+        if condition.price_below is not None:
+            checks.append(price is not None and price < condition.price_below)
+
+        price_change_pct = self._safe_float(market_data.get("price_change_pct"))
+        if condition.price_change_pct_above is not None or condition.price_change_pct_below is not None:
+            if "price_change_pct" not in market_data:
+                self._warn_missing_key("price_change_pct")
+        if condition.price_change_pct_above is not None:
+            checks.append(price_change_pct is not None and price_change_pct > condition.price_change_pct_above)
+        if condition.price_change_pct_below is not None:
+            checks.append(price_change_pct is not None and price_change_pct < condition.price_change_pct_below)
+
+        # Position-aware conditions
+        unrealized_pnl_pct = self._safe_float(market_data.get("unrealized_pnl_pct"))
+        if condition.unrealized_pnl_pct_above is not None or condition.unrealized_pnl_pct_below is not None:
+            if "unrealized_pnl_pct" not in market_data:
+                self._warn_missing_key("unrealized_pnl_pct")
+        if condition.unrealized_pnl_pct_above is not None:
+            checks.append(
+                unrealized_pnl_pct is not None
+                and unrealized_pnl_pct > condition.unrealized_pnl_pct_above
+            )
+        if condition.unrealized_pnl_pct_below is not None:
+            checks.append(
+                unrealized_pnl_pct is not None
+                and unrealized_pnl_pct < condition.unrealized_pnl_pct_below
+            )
+
+        holding_days = self._safe_float(market_data.get("holding_days"))
+        if condition.holding_days_above is not None or condition.holding_days_below is not None:
+            if "holding_days" not in market_data:
+                self._warn_missing_key("holding_days")
+        if condition.holding_days_above is not None:
+            checks.append(
+                holding_days is not None
+                and holding_days > condition.holding_days_above
+            )
+        if condition.holding_days_below is not None:
+            checks.append(
+                holding_days is not None
+                and holding_days < condition.holding_days_below
+            )
+
+        return len(checks) > 0 and all(checks)
+
+    def _evaluate_global_condition(
+        self,
+        condition_str: str,
+        portfolio_data: dict[str, Any],
+    ) -> bool:
+        """Evaluate a simple global condition string against portfolio data.
+
+        Supports: "field < value", "field > value", "field <= value", "field >= value"
+        """
+        parts = condition_str.strip().split()
+        if len(parts) != 3:
+            logger.warning("Invalid global condition format: %s", condition_str)
+            return False
+
+        field_name, operator, value_str = parts
+        try:
+            threshold = float(value_str)
+        except ValueError:
+            logger.warning("Invalid threshold in condition: %s", condition_str)
+            return False
+
+        actual = portfolio_data.get(field_name)
+        if actual is None:
+            return False
+
+        try:
+            actual_val = float(actual)
+        except (ValueError, TypeError):
+            return False
+
+        if operator == "<":
+            return actual_val < threshold
+        elif operator == ">":
+            return actual_val > threshold
+        elif operator == "<=":
+            return actual_val <= threshold
+        elif operator == ">=":
+            return actual_val >= threshold
+        else:
+            logger.warning("Unknown operator in condition: %s", operator)
+            return False
+
+    def _build_match_details(
+        self,
+        condition: StockCondition,
+        market_data: dict[str, Any],
+    ) -> dict[str, Any]:
+        """Build a summary of which conditions matched and their normalized values."""
+        details: dict[str, Any] = {}
+
+        if condition.rsi_below is not None or condition.rsi_above is not None:
+            details["rsi"] = self._safe_float(market_data.get("rsi"))
+        if condition.volume_ratio_above is not None or condition.volume_ratio_below is not None:
+            details["volume_ratio"] = self._safe_float(market_data.get("volume_ratio"))
+        if condition.price_above is not None or condition.price_below is not None:
+            details["current_price"] = self._safe_float(market_data.get("current_price"))
+        if condition.price_change_pct_above is not None or condition.price_change_pct_below is not None:
+            details["price_change_pct"] = self._safe_float(market_data.get("price_change_pct"))
+        if condition.unrealized_pnl_pct_above is not None or condition.unrealized_pnl_pct_below is not None:
+            details["unrealized_pnl_pct"] = self._safe_float(market_data.get("unrealized_pnl_pct"))
+        if condition.holding_days_above is not None or condition.holding_days_below is not None:
+            details["holding_days"] = self._safe_float(market_data.get("holding_days"))
+
+        return details

tests/test_backup.py

@@ -0,0 +1,799 @@
+"""Tests for backup and disaster recovery system."""
+
+from __future__ import annotations
+
+import sqlite3
+import sys
+import tempfile
+from datetime import UTC, datetime, timedelta
+from pathlib import Path
+from unittest.mock import MagicMock, patch
+
+import pytest
+
+from src.backup.exporter import BackupExporter, ExportFormat
+from src.backup.health_monitor import HealthMonitor, HealthStatus
+from src.backup.scheduler import BackupPolicy, BackupScheduler
+
+
+@pytest.fixture
+def temp_db(tmp_path: Path) -> Path:
+    """Create a temporary test database."""
+    db_path = tmp_path / "test_trades.db"
+
+    conn = sqlite3.connect(str(db_path))
+    cursor = conn.cursor()
+
+    # Create trades table
+    cursor.execute("""
+        CREATE TABLE trades (
+            id INTEGER PRIMARY KEY AUTOINCREMENT,
+            timestamp TEXT NOT NULL,
+            stock_code TEXT NOT NULL,
+            action TEXT NOT NULL,
+            quantity INTEGER NOT NULL,
+            price REAL NOT NULL,
+            confidence INTEGER NOT NULL,
+            rationale TEXT,
+            pnl REAL DEFAULT 0.0
+        )
+    """)
+
+    # Insert test data
+    test_trades = [
+        ("2024-01-01T10:00:00Z", "005930", "BUY", 10, 70000.0, 85, "Test buy", 0.0),
+        ("2024-01-01T11:00:00Z", "005930", "SELL", 10, 71000.0, 90, "Test sell", 10000.0),
+        ("2024-01-02T10:00:00Z", "AAPL", "BUY", 5, 180.0, 88, "Tech buy", 0.0),
+    ]
+
+    cursor.executemany(
+        """
+        INSERT INTO trades (timestamp, stock_code, action, quantity, price, confidence, rationale, pnl)
+        VALUES (?, ?, ?, ?, ?, ?, ?, ?)
+        """,
+        test_trades,
+    )
+
+    conn.commit()
+    conn.close()
+
+    return db_path
+
+
+class TestBackupExporter:
+    """Test BackupExporter functionality."""
+
+    def test_exporter_init(self, temp_db: Path) -> None:
+        """Test exporter initialization."""
+        exporter = BackupExporter(str(temp_db))
+        assert exporter.db_path == str(temp_db)
+
+    def test_export_json(self, temp_db: Path, tmp_path: Path) -> None:
+        """Test JSON export."""
+        exporter = BackupExporter(str(temp_db))
+        output_dir = tmp_path / "exports"
+
+        results = exporter.export_all(
+            output_dir, formats=[ExportFormat.JSON], compress=False
+        )
+
+        assert ExportFormat.JSON in results
+        assert results[ExportFormat.JSON].exists()
+        assert results[ExportFormat.JSON].suffix == ".json"
+
+    def test_export_json_compressed(self, temp_db: Path, tmp_path: Path) -> None:
+        """Test compressed JSON export."""
+        exporter = BackupExporter(str(temp_db))
+        output_dir = tmp_path / "exports"
+
+        results = exporter.export_all(
+            output_dir, formats=[ExportFormat.JSON], compress=True
+        )
+
+        assert ExportFormat.JSON in results
+        assert results[ExportFormat.JSON].suffix == ".gz"
+
+    def test_export_csv(self, temp_db: Path, tmp_path: Path) -> None:
+        """Test CSV export."""
+        exporter = BackupExporter(str(temp_db))
+        output_dir = tmp_path / "exports"
+
+        results = exporter.export_all(
+            output_dir, formats=[ExportFormat.CSV], compress=False
+        )
+
+        assert ExportFormat.CSV in results
+        assert results[ExportFormat.CSV].exists()
+
+        # Verify CSV content
+        with open(results[ExportFormat.CSV], "r") as f:
+            lines = f.readlines()
+            assert len(lines) == 4  # Header + 3 rows
+
+    def test_export_all_formats(self, temp_db: Path, tmp_path: Path) -> None:
+        """Test exporting all formats."""
+        exporter = BackupExporter(str(temp_db))
+        output_dir = tmp_path / "exports"
+
+        # Skip Parquet if pyarrow not available
+        try:
+            import pyarrow  # noqa: F401
+
+            formats = [ExportFormat.JSON, ExportFormat.CSV, ExportFormat.PARQUET]
+        except ImportError:
+            formats = [ExportFormat.JSON, ExportFormat.CSV]
+
+        results = exporter.export_all(output_dir, formats=formats, compress=False)
+
+        for fmt in formats:
+            assert fmt in results
+            assert results[fmt].exists()
+
+    def test_incremental_export(self, temp_db: Path, tmp_path: Path) -> None:
+        """Test incremental export."""
+        exporter = BackupExporter(str(temp_db))
+        output_dir = tmp_path / "exports"
+
+        # Export only trades after Jan 2
+        cutoff = datetime(2024, 1, 2, tzinfo=UTC)
+        results = exporter.export_all(
+            output_dir,
+            formats=[ExportFormat.JSON],
+            compress=False,
+            incremental_since=cutoff,
+        )
+
+        # Should only have 1 trade (AAPL on Jan 2)
+        import json
+
+        with open(results[ExportFormat.JSON], "r") as f:
+            data = json.load(f)
+            assert data["record_count"] == 1
+            assert data["trades"][0]["stock_code"] == "AAPL"
+
+    def test_get_export_stats(self, temp_db: Path) -> None:
+        """Test export statistics."""
+        exporter = BackupExporter(str(temp_db))
+        stats = exporter.get_export_stats()
+
+        assert stats["total_trades"] == 3
+        assert "date_range" in stats
+        assert "db_size_bytes" in stats
+
+
+class TestBackupScheduler:
+    """Test BackupScheduler functionality."""
+
+    def test_scheduler_init(self, temp_db: Path, tmp_path: Path) -> None:
+        """Test scheduler initialization."""
+        backup_dir = tmp_path / "backups"
+        scheduler = BackupScheduler(str(temp_db), backup_dir)
+
+        assert scheduler.db_path == temp_db
+        assert (backup_dir / "daily").exists()
+        assert (backup_dir / "weekly").exists()
+        assert (backup_dir / "monthly").exists()
+
+    def test_create_daily_backup(self, temp_db: Path, tmp_path: Path) -> None:
+        """Test daily backup creation."""
+        backup_dir = tmp_path / "backups"
+        scheduler = BackupScheduler(str(temp_db), backup_dir)
+
+        metadata = scheduler.create_backup(BackupPolicy.DAILY, verify=True)
+
+        assert metadata.policy == BackupPolicy.DAILY
+        assert metadata.file_path.exists()
+        assert metadata.size_bytes > 0
+        assert metadata.checksum is not None
+
+    def test_create_weekly_backup(self, temp_db: Path, tmp_path: Path) -> None:
+        """Test weekly backup creation."""
+        backup_dir = tmp_path / "backups"
+        scheduler = BackupScheduler(str(temp_db), backup_dir)
+
+        metadata = scheduler.create_backup(BackupPolicy.WEEKLY, verify=False)
+
+        assert metadata.policy == BackupPolicy.WEEKLY
+        assert metadata.file_path.exists()
+        assert metadata.checksum is None  # verify=False
+
+    def test_list_backups(self, temp_db: Path, tmp_path: Path) -> None:
+        """Test listing backups."""
+        backup_dir = tmp_path / "backups"
+        scheduler = BackupScheduler(str(temp_db), backup_dir)
+
+        scheduler.create_backup(BackupPolicy.DAILY)
+        scheduler.create_backup(BackupPolicy.WEEKLY)
+
+        backups = scheduler.list_backups()
+        assert len(backups) == 2
+
+        daily_backups = scheduler.list_backups(BackupPolicy.DAILY)
+        assert len(daily_backups) == 1
+        assert daily_backups[0].policy == BackupPolicy.DAILY
+
+    def test_cleanup_old_backups(self, temp_db: Path, tmp_path: Path) -> None:
+        """Test cleanup of old backups."""
+        backup_dir = tmp_path / "backups"
+        scheduler = BackupScheduler(str(temp_db), backup_dir, daily_retention_days=0)
+
+        # Create a backup
+        scheduler.create_backup(BackupPolicy.DAILY)
+
+        # Cleanup should remove it (0 day retention)
+        removed = scheduler.cleanup_old_backups()
+        assert removed[BackupPolicy.DAILY] >= 1
+
+    def test_backup_stats(self, temp_db: Path, tmp_path: Path) -> None:
+        """Test backup statistics."""
+        backup_dir = tmp_path / "backups"
+        scheduler = BackupScheduler(str(temp_db), backup_dir)
+
+        scheduler.create_backup(BackupPolicy.DAILY)
+        scheduler.create_backup(BackupPolicy.MONTHLY)
+
+        stats = scheduler.get_backup_stats()
+
+        assert stats["daily"]["count"] == 1
+        assert stats["monthly"]["count"] == 1
+        assert stats["daily"]["total_size_bytes"] > 0
+
+    def test_restore_backup(self, temp_db: Path, tmp_path: Path) -> None:
+        """Test backup restoration."""
+        backup_dir = tmp_path / "backups"
+        scheduler = BackupScheduler(str(temp_db), backup_dir)
+
+        # Create backup
+        metadata = scheduler.create_backup(BackupPolicy.DAILY)
+
+        # Modify database
+        conn = sqlite3.connect(str(temp_db))
+        conn.execute("DELETE FROM trades")
+        conn.commit()
+        conn.close()
+
+        # Restore
+        scheduler.restore_backup(metadata, verify=True)
+
+        # Verify restoration
+        conn = sqlite3.connect(str(temp_db))
+        cursor = conn.execute("SELECT COUNT(*) FROM trades")
+        count = cursor.fetchone()[0]
+        conn.close()
+
+        assert count == 3  # Original 3 trades restored
+
+
+class TestHealthMonitor:
+    """Test HealthMonitor functionality."""
+
+    def test_monitor_init(self, temp_db: Path, tmp_path: Path) -> None:
+        """Test monitor initialization."""
+        backup_dir = tmp_path / "backups"
+        monitor = HealthMonitor(str(temp_db), backup_dir)
+
+        assert monitor.db_path == temp_db
+
+    def test_check_database_health_ok(self, temp_db: Path, tmp_path: Path) -> None:
+        """Test database health check (healthy)."""
+        monitor = HealthMonitor(str(temp_db), tmp_path / "backups")
+        result = monitor.check_database_health()
+
+        assert result.status == HealthStatus.HEALTHY
+        assert "healthy" in result.message.lower()
+        assert result.details is not None
+        assert result.details["trade_count"] == 3
+
+    def test_check_database_health_missing(self, tmp_path: Path) -> None:
+        """Test database health check (missing file)."""
+        non_existent = tmp_path / "missing.db"
+        monitor = HealthMonitor(str(non_existent), tmp_path / "backups")
+        result = monitor.check_database_health()
+
+        assert result.status == HealthStatus.UNHEALTHY
+        assert "not found" in result.message.lower()
+
+    def test_check_disk_space(self, temp_db: Path, tmp_path: Path) -> None:
+        """Test disk space check."""
+        monitor = HealthMonitor(str(temp_db), tmp_path, min_disk_space_gb=0.001)
+        result = monitor.check_disk_space()
+
+        # Should be healthy with minimal requirement
+        assert result.status in [HealthStatus.HEALTHY, HealthStatus.DEGRADED]
+        assert result.details is not None
+        assert "free_gb" in result.details
+
+    def test_check_backup_recency_no_backups(self, temp_db: Path, tmp_path: Path) -> None:
+        """Test backup recency check (no backups)."""
+        backup_dir = tmp_path / "backups"
+        backup_dir.mkdir()
+        (backup_dir / "daily").mkdir()
+
+        monitor = HealthMonitor(str(temp_db), backup_dir)
+        result = monitor.check_backup_recency()
+
+        assert result.status == HealthStatus.UNHEALTHY
+        assert "no" in result.message.lower()
+
+    def test_check_backup_recency_recent(self, temp_db: Path, tmp_path: Path) -> None:
+        """Test backup recency check (recent backup)."""
+        backup_dir = tmp_path / "backups"
+        scheduler = BackupScheduler(str(temp_db), backup_dir)
+        scheduler.create_backup(BackupPolicy.DAILY)
+
+        monitor = HealthMonitor(str(temp_db), backup_dir)
+        result = monitor.check_backup_recency()
+
+        assert result.status == HealthStatus.HEALTHY
+        assert "recent" in result.message.lower()
+
+    def test_run_all_checks(self, temp_db: Path, tmp_path: Path) -> None:
+        """Test running all health checks."""
+        backup_dir = tmp_path / "backups"
+        scheduler = BackupScheduler(str(temp_db), backup_dir)
+        scheduler.create_backup(BackupPolicy.DAILY)
+
+        monitor = HealthMonitor(str(temp_db), backup_dir, min_disk_space_gb=0.001)
+        checks = monitor.run_all_checks()
+
+        assert "database" in checks
+        assert "disk_space" in checks
+        assert "backup_recency" in checks
+        assert checks["database"].status == HealthStatus.HEALTHY
+
+    def test_get_overall_status(self, temp_db: Path, tmp_path: Path) -> None:
+        """Test overall health status."""
+        backup_dir = tmp_path / "backups"
+        scheduler = BackupScheduler(str(temp_db), backup_dir)
+        scheduler.create_backup(BackupPolicy.DAILY)
+
+        monitor = HealthMonitor(str(temp_db), backup_dir, min_disk_space_gb=0.001)
+        status = monitor.get_overall_status()
+
+        assert status in [HealthStatus.HEALTHY, HealthStatus.DEGRADED]
+
+    def test_get_health_report(self, temp_db: Path, tmp_path: Path) -> None:
+        """Test health report generation."""
+        backup_dir = tmp_path / "backups"
+        scheduler = BackupScheduler(str(temp_db), backup_dir)
+        scheduler.create_backup(BackupPolicy.DAILY)
+
+        monitor = HealthMonitor(str(temp_db), backup_dir, min_disk_space_gb=0.001)
+        report = monitor.get_health_report()
+
+        assert "overall_status" in report
+        assert "timestamp" in report
+        assert "checks" in report
+        assert len(report["checks"]) == 3
+
+
+# ---------------------------------------------------------------------------
+# BackupExporter — additional coverage for previously uncovered branches
+# ---------------------------------------------------------------------------
+
+
+@pytest.fixture
+def empty_db(tmp_path: Path) -> Path:
+    """Create a temporary database with NO trade records."""
+    db_path = tmp_path / "empty_trades.db"
+    conn = sqlite3.connect(str(db_path))
+    conn.execute(
+        """CREATE TABLE trades (
+            id INTEGER PRIMARY KEY AUTOINCREMENT,
+            timestamp TEXT NOT NULL,
+            stock_code TEXT NOT NULL,
+            action TEXT NOT NULL,
+            quantity INTEGER NOT NULL,
+            price REAL NOT NULL,
+            confidence INTEGER NOT NULL,
+            rationale TEXT,
+            pnl REAL DEFAULT 0.0
+        )"""
+    )
+    conn.commit()
+    conn.close()
+    return db_path
+
+
+class TestBackupExporterAdditional:
+    """Cover branches missed in the original TestBackupExporter suite."""
+
+    def test_export_all_default_formats(self, temp_db: Path, tmp_path: Path) -> None:
+        """export_all with formats=None must default to JSON+CSV+Parquet path."""
+        exporter = BackupExporter(str(temp_db))
+        # formats=None triggers the default list assignment (line 62)
+        results = exporter.export_all(tmp_path / "out", formats=None, compress=False)
+        # JSON and CSV must always succeed; Parquet needs pyarrow
+        assert ExportFormat.JSON in results
+        assert ExportFormat.CSV in results
+
+    def test_export_all_logs_error_on_failure(
+        self, temp_db: Path, tmp_path: Path
+    ) -> None:
+        """export_all must log an error and continue when one format fails."""
+        exporter = BackupExporter(str(temp_db))
+        # Patch _export_format to raise on JSON, succeed on CSV
+        original = exporter._export_format
+
+        def failing_export(fmt, *args, **kwargs):  # type: ignore[no-untyped-def]
+            if fmt == ExportFormat.JSON:
+                raise RuntimeError("simulated failure")
+            return original(fmt, *args, **kwargs)
+
+        exporter._export_format = failing_export  # type: ignore[method-assign]
+        results = exporter.export_all(
+            tmp_path / "out",
+            formats=[ExportFormat.JSON, ExportFormat.CSV],
+            compress=False,
+        )
+        # JSON failed → not in results; CSV succeeded → in results
+        assert ExportFormat.JSON not in results
+        assert ExportFormat.CSV in results
+
+    def test_export_csv_empty_trades_no_compress(
+        self, empty_db: Path, tmp_path: Path
+    ) -> None:
+        """CSV export with no trades and compress=False must write header row only."""
+        exporter = BackupExporter(str(empty_db))
+        results = exporter.export_all(
+            tmp_path / "out",
+            formats=[ExportFormat.CSV],
+            compress=False,
+        )
+        assert ExportFormat.CSV in results
+        out = results[ExportFormat.CSV]
+        assert out.exists()
+        content = out.read_text()
+        assert "timestamp" in content
+
+    def test_export_csv_empty_trades_compressed(
+        self, empty_db: Path, tmp_path: Path
+    ) -> None:
+        """CSV export with no trades and compress=True must write gzipped header."""
+        import gzip
+
+        exporter = BackupExporter(str(empty_db))
+        results = exporter.export_all(
+            tmp_path / "out",
+            formats=[ExportFormat.CSV],
+            compress=True,
+        )
+        assert ExportFormat.CSV in results
+        out = results[ExportFormat.CSV]
+        assert out.suffix == ".gz"
+        with gzip.open(out, "rt", encoding="utf-8") as f:
+            content = f.read()
+        assert "timestamp" in content
+
+    def test_export_csv_with_data_compressed(
+        self, temp_db: Path, tmp_path: Path
+    ) -> None:
+        """CSV export with data and compress=True must write gzipped rows."""
+        import gzip
+
+        exporter = BackupExporter(str(temp_db))
+        results = exporter.export_all(
+            tmp_path / "out",
+            formats=[ExportFormat.CSV],
+            compress=True,
+        )
+        assert ExportFormat.CSV in results
+        out = results[ExportFormat.CSV]
+        with gzip.open(out, "rt", encoding="utf-8") as f:
+            lines = f.readlines()
+        # Header + 3 data rows
+        assert len(lines) == 4
+
+    def test_export_parquet_raises_import_error_without_pyarrow(
+        self, temp_db: Path, tmp_path: Path
+    ) -> None:
+        """Parquet export must raise ImportError when pyarrow is not installed."""
+        exporter = BackupExporter(str(temp_db))
+        with patch.dict(sys.modules, {"pyarrow": None, "pyarrow.parquet": None}):
+            try:
+                import pyarrow  # noqa: F401
+                pytest.skip("pyarrow is installed; cannot test ImportError path")
+            except ImportError:
+                pass
+            results = exporter.export_all(
+                tmp_path / "out",
+                formats=[ExportFormat.PARQUET],
+                compress=False,
+            )
+            # Parquet export fails gracefully; result dict should not contain it
+            assert ExportFormat.PARQUET not in results
+
+
+# ---------------------------------------------------------------------------
+# CloudStorage — mocked boto3 tests
+# ---------------------------------------------------------------------------
+
+
+@pytest.fixture
+def mock_boto3_module():
+    """Inject a fake boto3 into sys.modules for the duration of the test."""
+    mock = MagicMock()
+    with patch.dict(sys.modules, {"boto3": mock}):
+        yield mock
+
+
+@pytest.fixture
+def s3_config():
+    """Minimal S3Config for tests."""
+    from src.backup.cloud_storage import S3Config
+
+    return S3Config(
+        endpoint_url="http://localhost:9000",
+        access_key="minioadmin",
+        secret_key="minioadmin",
+        bucket_name="test-bucket",
+        region="us-east-1",
+    )
+
+
+class TestCloudStorage:
+    """Test CloudStorage using mocked boto3."""
+
+    def test_init_creates_s3_client(self, mock_boto3_module, s3_config) -> None:
+        """CloudStorage.__init__ must call boto3.client with the correct args."""
+        from src.backup.cloud_storage import CloudStorage
+
+        storage = CloudStorage(s3_config)
+        mock_boto3_module.client.assert_called_once()
+        call_kwargs = mock_boto3_module.client.call_args[1]
+        assert call_kwargs["aws_access_key_id"] == "minioadmin"
+        assert call_kwargs["aws_secret_access_key"] == "minioadmin"
+        assert storage.config == s3_config
+
+    def test_init_raises_if_boto3_missing(self, s3_config) -> None:
+        """CloudStorage.__init__ must raise ImportError when boto3 is absent."""
+        with patch.dict(sys.modules, {"boto3": None}):  # type: ignore[dict-item]
+            with pytest.raises((ImportError, TypeError)):
+                # Re-import to trigger the try/except inside __init__
+                import importlib
+
+                import src.backup.cloud_storage as m
+
+                importlib.reload(m)
+                m.CloudStorage(s3_config)
+
+    def test_upload_file_success(
+        self, mock_boto3_module, s3_config, tmp_path: Path
+    ) -> None:
+        """upload_file must call client.upload_file and return the object key."""
+        from src.backup.cloud_storage import CloudStorage
+
+        test_file = tmp_path / "backup.json.gz"
+        test_file.write_bytes(b"data")
+
+        storage = CloudStorage(s3_config)
+        key = storage.upload_file(test_file, object_key="backups/backup.json.gz")
+
+        assert key == "backups/backup.json.gz"
+        storage.client.upload_file.assert_called_once()
+
+    def test_upload_file_default_key(
+        self, mock_boto3_module, s3_config, tmp_path: Path
+    ) -> None:
+        """upload_file without object_key must use the filename as key."""
+        from src.backup.cloud_storage import CloudStorage
+
+        test_file = tmp_path / "myfile.gz"
+        test_file.write_bytes(b"data")
+
+        storage = CloudStorage(s3_config)
+        key = storage.upload_file(test_file)
+
+        assert key == "myfile.gz"
+
+    def test_upload_file_not_found(
+        self, mock_boto3_module, s3_config, tmp_path: Path
+    ) -> None:
+        """upload_file must raise FileNotFoundError for missing files."""
+        from src.backup.cloud_storage import CloudStorage
+
+        storage = CloudStorage(s3_config)
+        with pytest.raises(FileNotFoundError):
+            storage.upload_file(tmp_path / "nonexistent.gz")
+
+    def test_upload_file_propagates_client_error(
+        self, mock_boto3_module, s3_config, tmp_path: Path
+    ) -> None:
+        """upload_file must re-raise exceptions from the boto3 client."""
+        from src.backup.cloud_storage import CloudStorage
+
+        test_file = tmp_path / "backup.gz"
+        test_file.write_bytes(b"data")
+
+        storage = CloudStorage(s3_config)
+        storage.client.upload_file.side_effect = RuntimeError("network error")
+
+        with pytest.raises(RuntimeError, match="network error"):
+            storage.upload_file(test_file)
+
+    def test_download_file_success(
+        self, mock_boto3_module, s3_config, tmp_path: Path
+    ) -> None:
+        """download_file must call client.download_file and return local path."""
+        from src.backup.cloud_storage import CloudStorage
+
+        storage = CloudStorage(s3_config)
+        dest = tmp_path / "downloads" / "backup.gz"
+
+        result = storage.download_file("backups/backup.gz", dest)
+
+        assert result == dest
+        storage.client.download_file.assert_called_once()
+
+    def test_download_file_propagates_error(
+        self, mock_boto3_module, s3_config, tmp_path: Path
+    ) -> None:
+        """download_file must re-raise exceptions from the boto3 client."""
+        from src.backup.cloud_storage import CloudStorage
+
+        storage = CloudStorage(s3_config)
+        storage.client.download_file.side_effect = RuntimeError("timeout")
+
+        with pytest.raises(RuntimeError, match="timeout"):
+            storage.download_file("key", tmp_path / "dest.gz")
+
+    def test_list_files_returns_objects(
+        self, mock_boto3_module, s3_config
+    ) -> None:
+        """list_files must return parsed file metadata from S3 response."""
+        from datetime import timezone
+
+        from src.backup.cloud_storage import CloudStorage
+
+        storage = CloudStorage(s3_config)
+        storage.client.list_objects_v2.return_value = {
+            "Contents": [
+                {
+                    "Key": "backups/a.gz",
+                    "Size": 1024,
+                    "LastModified": datetime(2026, 1, 1, tzinfo=timezone.utc),
+                    "ETag": '"abc123"',
+                }
+            ]
+        }
+
+        files = storage.list_files(prefix="backups/")
+        assert len(files) == 1
+        assert files[0]["key"] == "backups/a.gz"
+        assert files[0]["size_bytes"] == 1024
+
+    def test_list_files_empty_bucket(
+        self, mock_boto3_module, s3_config
+    ) -> None:
+        """list_files must return empty list when bucket has no objects."""
+        from src.backup.cloud_storage import CloudStorage
+
+        storage = CloudStorage(s3_config)
+        storage.client.list_objects_v2.return_value = {}
+
+        files = storage.list_files()
+        assert files == []
+
+    def test_list_files_propagates_error(
+        self, mock_boto3_module, s3_config
+    ) -> None:
+        """list_files must re-raise exceptions from the boto3 client."""
+        from src.backup.cloud_storage import CloudStorage
+
+        storage = CloudStorage(s3_config)
+        storage.client.list_objects_v2.side_effect = RuntimeError("auth error")
+
+        with pytest.raises(RuntimeError):
+            storage.list_files()
+
+    def test_delete_file_success(
+        self, mock_boto3_module, s3_config
+    ) -> None:
+        """delete_file must call client.delete_object with the correct key."""
+        from src.backup.cloud_storage import CloudStorage
+
+        storage = CloudStorage(s3_config)
+        storage.delete_file("backups/old.gz")
+        storage.client.delete_object.assert_called_once_with(
+            Bucket="test-bucket", Key="backups/old.gz"
+        )
+
+    def test_delete_file_propagates_error(
+        self, mock_boto3_module, s3_config
+    ) -> None:
+        """delete_file must re-raise exceptions from the boto3 client."""
+        from src.backup.cloud_storage import CloudStorage
+
+        storage = CloudStorage(s3_config)
+        storage.client.delete_object.side_effect = RuntimeError("permission denied")
+
+        with pytest.raises(RuntimeError):
+            storage.delete_file("backups/old.gz")
+
+    def test_get_storage_stats_success(
+        self, mock_boto3_module, s3_config
+    ) -> None:
+        """get_storage_stats must aggregate file sizes correctly."""
+        from datetime import timezone
+
+        from src.backup.cloud_storage import CloudStorage
+
+        storage = CloudStorage(s3_config)
+        storage.client.list_objects_v2.return_value = {
+            "Contents": [
+                {
+                    "Key": "a.gz",
+                    "Size": 1024 * 1024,
+                    "LastModified": datetime(2026, 1, 1, tzinfo=timezone.utc),
+                    "ETag": '"x"',
+                },
+                {
+                    "Key": "b.gz",
+                    "Size": 1024 * 1024,
+                    "LastModified": datetime(2026, 1, 2, tzinfo=timezone.utc),
+                    "ETag": '"y"',
+                },
+            ]
+        }
+
+        stats = storage.get_storage_stats()
+        assert stats["total_files"] == 2
+        assert stats["total_size_bytes"] == 2 * 1024 * 1024
+        assert stats["total_size_mb"] == pytest.approx(2.0)
+
+    def test_get_storage_stats_on_error(
+        self, mock_boto3_module, s3_config
+    ) -> None:
+        """get_storage_stats must return error dict without raising on failure."""
+        from src.backup.cloud_storage import CloudStorage
+
+        storage = CloudStorage(s3_config)
+        storage.client.list_objects_v2.side_effect = RuntimeError("no connection")
+
+        stats = storage.get_storage_stats()
+        assert "error" in stats
+        assert stats["total_files"] == 0
+
+    def test_verify_connection_success(
+        self, mock_boto3_module, s3_config
+    ) -> None:
+        """verify_connection must return True when head_bucket succeeds."""
+        from src.backup.cloud_storage import CloudStorage
+
+        storage = CloudStorage(s3_config)
+        result = storage.verify_connection()
+        assert result is True
+
+    def test_verify_connection_failure(
+        self, mock_boto3_module, s3_config
+    ) -> None:
+        """verify_connection must return False when head_bucket raises."""
+        from src.backup.cloud_storage import CloudStorage
+
+        storage = CloudStorage(s3_config)
+        storage.client.head_bucket.side_effect = RuntimeError("no such bucket")
+
+        result = storage.verify_connection()
+        assert result is False
+
+    def test_enable_versioning(
+        self, mock_boto3_module, s3_config
+    ) -> None:
+        """enable_versioning must call put_bucket_versioning."""
+        from src.backup.cloud_storage import CloudStorage
+
+        storage = CloudStorage(s3_config)
+        storage.enable_versioning()
+        storage.client.put_bucket_versioning.assert_called_once()
+
+    def test_enable_versioning_propagates_error(
+        self, mock_boto3_module, s3_config
+    ) -> None:
+        """enable_versioning must re-raise exceptions from the boto3 client."""
+        from src.backup.cloud_storage import CloudStorage
+
+        storage = CloudStorage(s3_config)
+        storage.client.put_bucket_versioning.side_effect = RuntimeError("denied")
+
+        with pytest.raises(RuntimeError):
+            storage.enable_versioning()

tests/test_blackout_manager.py

@@ -0,0 +1,81 @@
+from __future__ import annotations
+
+from datetime import UTC, datetime
+
+from src.core.blackout_manager import (
+    BlackoutOrderManager,
+    QueuedOrderIntent,
+    parse_blackout_windows_kst,
+)
+
+
+def test_parse_blackout_windows_kst() -> None:
+    windows = parse_blackout_windows_kst("23:30-00:10,11:20-11:30,invalid")
+    assert len(windows) == 2
+
+
+def test_blackout_manager_handles_cross_midnight_window() -> None:
+    manager = BlackoutOrderManager(
+        enabled=True,
+        windows=parse_blackout_windows_kst("23:30-00:10"),
+        max_queue_size=10,
+    )
+    # 2026-01-01 23:40 KST = 2026-01-01 14:40 UTC
+    assert manager.in_blackout(datetime(2026, 1, 1, 14, 40, tzinfo=UTC))
+    # 2026-01-02 00:20 KST = 2026-01-01 15:20 UTC
+    assert not manager.in_blackout(datetime(2026, 1, 1, 15, 20, tzinfo=UTC))
+
+
+def test_recovery_batch_only_after_blackout_exit() -> None:
+    manager = BlackoutOrderManager(
+        enabled=True,
+        windows=parse_blackout_windows_kst("23:30-00:10"),
+        max_queue_size=10,
+    )
+    intent = QueuedOrderIntent(
+        market_code="KR",
+        exchange_code="KRX",
+        stock_code="005930",
+        order_type="BUY",
+        quantity=1,
+        price=100.0,
+        source="test",
+        queued_at=datetime.now(UTC),
+    )
+    assert manager.enqueue(intent)
+
+    # Inside blackout: no pop yet
+    inside_blackout = datetime(2026, 1, 1, 14, 40, tzinfo=UTC)
+    assert manager.pop_recovery_batch(inside_blackout) == []
+
+    # Outside blackout: pop full batch once
+    outside_blackout = datetime(2026, 1, 1, 15, 20, tzinfo=UTC)
+    batch = manager.pop_recovery_batch(outside_blackout)
+    assert len(batch) == 1
+    assert manager.pending_count == 0
+
+
+def test_requeued_intent_is_processed_next_non_blackout_cycle() -> None:
+    manager = BlackoutOrderManager(
+        enabled=True,
+        windows=parse_blackout_windows_kst("23:30-00:10"),
+        max_queue_size=10,
+    )
+    intent = QueuedOrderIntent(
+        market_code="KR",
+        exchange_code="KRX",
+        stock_code="005930",
+        order_type="BUY",
+        quantity=1,
+        price=100.0,
+        source="test",
+        queued_at=datetime.now(UTC),
+    )
+    manager.enqueue(intent)
+    outside_blackout = datetime(2026, 1, 1, 15, 20, tzinfo=UTC)
+    first_batch = manager.pop_recovery_batch(outside_blackout)
+    assert len(first_batch) == 1
+
+    manager.requeue(first_batch[0])
+    second_batch = manager.pop_recovery_batch(outside_blackout)
+    assert len(second_batch) == 1

tests/test_brain.py

@@ -2,6 +2,10 @@
 
 from __future__ import annotations
 
+from unittest.mock import AsyncMock, MagicMock, patch
+
+import pytest
+
 from src.brain.gemini_client import GeminiClient
 
 # ---------------------------------------------------------------------------
@@ -89,9 +93,21 @@ class TestMalformedJsonHandling:
 
     def test_json_with_missing_fields_returns_hold(self, settings):
         client = GeminiClient(settings)
-        decision = client.parse_response('{"action": "BUY"}')
+        raw = '{"action": "BUY"}'
+        decision = client.parse_response(raw)
         assert decision.action == "HOLD"
         assert decision.confidence == 0
+        # rationale preserves raw so prompt_override callers (e.g. pre_market_planner)
+        # can extract non-TradeDecision JSON from decision.rationale (#245)
+        assert decision.rationale == raw
+
+    def test_non_trade_decision_json_preserves_raw_in_rationale(self, settings):
+        """Playbook JSON (no action/confidence/rationale) must be preserved for planner."""
+        client = GeminiClient(settings)
+        playbook_json = '{"market_outlook": "neutral", "stocks": []}'
+        decision = client.parse_response(playbook_json)
+        assert decision.action == "HOLD"
+        assert decision.rationale == playbook_json
 
     def test_json_with_invalid_action_returns_hold(self, settings):
         client = GeminiClient(settings)
@@ -126,7 +142,7 @@ class TestPromptConstruction:
             "orderbook": {"asks": [], "bids": []},
             "foreigner_net": -50000,
         }
-        prompt = client.build_prompt(market_data)
+        prompt = client.build_prompt_sync(market_data)
         assert "005930" in prompt
 
     def test_prompt_contains_price(self, settings):
@@ -137,7 +153,7 @@ class TestPromptConstruction:
             "orderbook": {"asks": [], "bids": []},
             "foreigner_net": -50000,
         }
-        prompt = client.build_prompt(market_data)
+        prompt = client.build_prompt_sync(market_data)
         assert "72000" in prompt
 
     def test_prompt_enforces_json_output_format(self, settings):
@@ -148,7 +164,254 @@ class TestPromptConstruction:
             "orderbook": {"asks": [], "bids": []},
             "foreigner_net": 0,
         }
-        prompt = client.build_prompt(market_data)
+        prompt = client.build_prompt_sync(market_data)
         assert "JSON" in prompt
         assert "action" in prompt
         assert "confidence" in prompt
+
+
+# ---------------------------------------------------------------------------
+# Batch Decision Making
+# ---------------------------------------------------------------------------
+
+
+class TestBatchDecisionParsing:
+    """Batch response parser must handle JSON arrays correctly."""
+
+    def test_parse_valid_batch_response(self, settings):
+        client = GeminiClient(settings)
+        stocks_data = [
+            {"stock_code": "AAPL", "current_price": 185.5},
+            {"stock_code": "MSFT", "current_price": 420.0},
+        ]
+        raw = """[
+            {"code": "AAPL", "action": "BUY", "confidence": 85, "rationale": "Strong momentum"},
+            {"code": "MSFT", "action": "HOLD", "confidence": 50, "rationale": "Wait for earnings"}
+        ]"""
+
+        decisions = client._parse_batch_response(raw, stocks_data, token_count=100)
+
+        assert len(decisions) == 2
+        assert decisions["AAPL"].action == "BUY"
+        assert decisions["AAPL"].confidence == 85
+        assert decisions["MSFT"].action == "HOLD"
+        assert decisions["MSFT"].confidence == 50
+
+    def test_parse_batch_with_markdown_wrapper(self, settings):
+        client = GeminiClient(settings)
+        stocks_data = [{"stock_code": "AAPL", "current_price": 185.5}]
+        raw = """```json
+[{"code": "AAPL", "action": "BUY", "confidence": 90, "rationale": "Good"}]
+```"""
+
+        decisions = client._parse_batch_response(raw, stocks_data, token_count=100)
+
+        assert decisions["AAPL"].action == "BUY"
+        assert decisions["AAPL"].confidence == 90
+
+    def test_parse_batch_empty_response_returns_hold_for_all(self, settings):
+        client = GeminiClient(settings)
+        stocks_data = [
+            {"stock_code": "AAPL", "current_price": 185.5},
+            {"stock_code": "MSFT", "current_price": 420.0},
+        ]
+
+        decisions = client._parse_batch_response("", stocks_data, token_count=100)
+
+        assert len(decisions) == 2
+        assert decisions["AAPL"].action == "HOLD"
+        assert decisions["AAPL"].confidence == 0
+        assert decisions["MSFT"].action == "HOLD"
+
+    def test_parse_batch_malformed_json_returns_hold_for_all(self, settings):
+        client = GeminiClient(settings)
+        stocks_data = [{"stock_code": "AAPL", "current_price": 185.5}]
+        raw = "This is not JSON"
+
+        decisions = client._parse_batch_response(raw, stocks_data, token_count=100)
+
+        assert decisions["AAPL"].action == "HOLD"
+        assert decisions["AAPL"].confidence == 0
+
+    def test_parse_batch_not_array_returns_hold_for_all(self, settings):
+        client = GeminiClient(settings)
+        stocks_data = [{"stock_code": "AAPL", "current_price": 185.5}]
+        raw = '{"code": "AAPL", "action": "BUY", "confidence": 90, "rationale": "Good"}'
+
+        decisions = client._parse_batch_response(raw, stocks_data, token_count=100)
+
+        assert decisions["AAPL"].action == "HOLD"
+        assert decisions["AAPL"].confidence == 0
+
+    def test_parse_batch_missing_stock_gets_hold(self, settings):
+        client = GeminiClient(settings)
+        stocks_data = [
+            {"stock_code": "AAPL", "current_price": 185.5},
+            {"stock_code": "MSFT", "current_price": 420.0},
+        ]
+        # Response only has AAPL, MSFT is missing
+        raw = '[{"code": "AAPL", "action": "BUY", "confidence": 85, "rationale": "Good"}]'
+
+        decisions = client._parse_batch_response(raw, stocks_data, token_count=100)
+
+        assert decisions["AAPL"].action == "BUY"
+        assert decisions["MSFT"].action == "HOLD"
+        assert decisions["MSFT"].confidence == 0
+
+    def test_parse_batch_invalid_action_becomes_hold(self, settings):
+        client = GeminiClient(settings)
+        stocks_data = [{"stock_code": "AAPL", "current_price": 185.5}]
+        raw = '[{"code": "AAPL", "action": "YOLO", "confidence": 90, "rationale": "Moon"}]'
+
+        decisions = client._parse_batch_response(raw, stocks_data, token_count=100)
+
+        assert decisions["AAPL"].action == "HOLD"
+
+    def test_parse_batch_low_confidence_becomes_hold(self, settings):
+        client = GeminiClient(settings)
+        stocks_data = [{"stock_code": "AAPL", "current_price": 185.5}]
+        raw = '[{"code": "AAPL", "action": "BUY", "confidence": 65, "rationale": "Weak"}]'
+
+        decisions = client._parse_batch_response(raw, stocks_data, token_count=100)
+
+        assert decisions["AAPL"].action == "HOLD"
+        assert decisions["AAPL"].confidence == 65
+
+    def test_parse_batch_missing_fields_gets_hold(self, settings):
+        client = GeminiClient(settings)
+        stocks_data = [{"stock_code": "AAPL", "current_price": 185.5}]
+        raw = '[{"code": "AAPL", "action": "BUY"}]'  # Missing confidence and rationale
+
+        decisions = client._parse_batch_response(raw, stocks_data, token_count=100)
+
+        assert decisions["AAPL"].action == "HOLD"
+        assert decisions["AAPL"].confidence == 0
+
+
+# ---------------------------------------------------------------------------
+# Prompt Override (used by pre_market_planner)
+# ---------------------------------------------------------------------------
+
+
+class TestPromptOverride:
+    """decide() must use prompt_override when present in market_data."""
+
+    @pytest.mark.asyncio
+    async def test_prompt_override_is_sent_to_gemini(self, settings):
+        """When prompt_override is in market_data, it should be used as the prompt."""
+        client = GeminiClient(settings)
+
+        custom_prompt = "You are a playbook generator. Return JSON with scenarios."
+        playbook_json = '{"market_outlook": "neutral", "stocks": []}'
+
+        mock_response = MagicMock()
+        mock_response.text = playbook_json
+
+        with patch.object(
+            client._client.aio.models,
+            "generate_content",
+            new_callable=AsyncMock,
+            return_value=mock_response,
+        ) as mock_generate:
+            market_data = {
+                "stock_code": "PLANNER",
+                "current_price": 0,
+                "prompt_override": custom_prompt,
+            }
+            decision = await client.decide(market_data)
+
+            # Verify the custom prompt was sent, not a built prompt
+            mock_generate.assert_called_once()
+            actual_prompt = mock_generate.call_args[1].get(
+                "contents", mock_generate.call_args[0][1] if len(mock_generate.call_args[0]) > 1 else None
+            )
+            assert actual_prompt == custom_prompt
+            # Raw response preserved in rationale without parse_response (#247)
+            assert decision.rationale == playbook_json
+
+    @pytest.mark.asyncio
+    async def test_prompt_override_skips_parse_response(self, settings):
+        """prompt_override bypasses parse_response — no Missing fields warning, raw preserved."""
+        client = GeminiClient(settings)
+        client._enable_optimization = True
+
+        custom_prompt = "Custom playbook prompt"
+        playbook_json = '{"market_outlook": "bullish", "stocks": [{"stock_code": "AAPL"}]}'
+
+        mock_response = MagicMock()
+        mock_response.text = playbook_json
+
+        with patch.object(
+            client._client.aio.models,
+            "generate_content",
+            new_callable=AsyncMock,
+            return_value=mock_response,
+        ):
+            with patch.object(client, "parse_response") as mock_parse:
+                market_data = {
+                    "stock_code": "PLANNER",
+                    "current_price": 0,
+                    "prompt_override": custom_prompt,
+                }
+                decision = await client.decide(market_data)
+
+                # parse_response must NOT be called for prompt_override
+                mock_parse.assert_not_called()
+                # Raw playbook JSON preserved in rationale
+                assert decision.rationale == playbook_json
+
+    @pytest.mark.asyncio
+    async def test_prompt_override_takes_priority_over_optimization(self, settings):
+        """prompt_override must win over enable_optimization=True."""
+        client = GeminiClient(settings)
+        client._enable_optimization = True
+
+        custom_prompt = "Explicit playbook prompt"
+
+        mock_response = MagicMock()
+        mock_response.text = '{"market_outlook": "neutral", "stocks": []}'
+
+        with patch.object(
+            client._client.aio.models,
+            "generate_content",
+            new_callable=AsyncMock,
+            return_value=mock_response,
+        ) as mock_generate:
+            market_data = {
+                "stock_code": "PLANNER",
+                "current_price": 0,
+                "prompt_override": custom_prompt,
+            }
+            await client.decide(market_data)
+
+            actual_prompt = mock_generate.call_args[1].get(
+                "contents", mock_generate.call_args[0][1] if len(mock_generate.call_args[0]) > 1 else None
+            )
+            # The custom prompt must be used, not the compressed prompt
+            assert actual_prompt == custom_prompt
+
+    @pytest.mark.asyncio
+    async def test_without_prompt_override_uses_build_prompt(self, settings):
+        """Without prompt_override, decide() should use build_prompt as before."""
+        client = GeminiClient(settings)
+
+        mock_response = MagicMock()
+        mock_response.text = '{"action": "HOLD", "confidence": 50, "rationale": "ok"}'
+
+        with patch.object(
+            client._client.aio.models,
+            "generate_content",
+            new_callable=AsyncMock,
+            return_value=mock_response,
+        ) as mock_generate:
+            market_data = {
+                "stock_code": "005930",
+                "current_price": 72000,
+            }
+            await client.decide(market_data)
+
+            actual_prompt = mock_generate.call_args[1].get(
+                "contents", mock_generate.call_args[0][1] if len(mock_generate.call_args[0]) > 1 else None
+            )
+            # Should contain stock code from build_prompt, not be a custom override
+            assert "005930" in actual_prompt

tests/test_broker.py

@@ -3,7 +3,7 @@
 from __future__ import annotations
 
 import asyncio
-from unittest.mock import AsyncMock, patch
+from unittest.mock import AsyncMock, MagicMock, patch
 
 import pytest
 
@@ -49,6 +49,110 @@ class TestTokenManagement:
 
         await broker.close()
 
+    @pytest.mark.asyncio
+    async def test_concurrent_token_refresh_calls_api_once(self, settings):
+        """Multiple concurrent token requests should only call API once."""
+        broker = KISBroker(settings)
+
+        # Track how many times the mock API is called
+        call_count = [0]
+
+        def create_mock_resp():
+            call_count[0] += 1
+            mock_resp = AsyncMock()
+            mock_resp.status = 200
+            mock_resp.json = AsyncMock(
+                return_value={
+                    "access_token": "tok_concurrent",
+                    "token_type": "Bearer",
+                    "expires_in": 86400,
+                }
+            )
+            mock_resp.__aenter__ = AsyncMock(return_value=mock_resp)
+            mock_resp.__aexit__ = AsyncMock(return_value=False)
+            return mock_resp
+
+        with patch("aiohttp.ClientSession.post", return_value=create_mock_resp()):
+            # Launch 5 concurrent token requests
+            tokens = await asyncio.gather(
+                broker._ensure_token(),
+                broker._ensure_token(),
+                broker._ensure_token(),
+                broker._ensure_token(),
+                broker._ensure_token(),
+            )
+
+            # All should get the same token
+            assert all(t == "tok_concurrent" for t in tokens)
+            # API should be called only once (due to lock)
+            assert call_count[0] == 1
+
+        await broker.close()
+
+    @pytest.mark.asyncio
+    async def test_token_refresh_cooldown_waits_then_retries(self, settings):
+        """Token refresh should wait out cooldown then retry (issue #54)."""
+        broker = KISBroker(settings)
+        broker._refresh_cooldown = 0.1  # Short cooldown for testing
+
+        # All attempts fail with 403 (EGW00133)
+        mock_resp_403 = AsyncMock()
+        mock_resp_403.status = 403
+        mock_resp_403.text = AsyncMock(
+            return_value='{"error_code":"EGW00133","error_description":"접근토큰 발급 잠시 후 다시 시도하세요(1분당 1회)"}'
+        )
+        mock_resp_403.__aenter__ = AsyncMock(return_value=mock_resp_403)
+        mock_resp_403.__aexit__ = AsyncMock(return_value=False)
+
+        with patch("aiohttp.ClientSession.post", return_value=mock_resp_403):
+            # First attempt should fail with 403
+            with pytest.raises(ConnectionError, match="Token refresh failed"):
+                await broker._ensure_token()
+
+            # Second attempt within cooldown should wait then retry (and still get 403)
+            with pytest.raises(ConnectionError, match="Token refresh failed"):
+                await broker._ensure_token()
+
+        await broker.close()
+
+    @pytest.mark.asyncio
+    async def test_token_refresh_allowed_after_cooldown(self, settings):
+        """Token refresh should be allowed after cooldown period expires."""
+        broker = KISBroker(settings)
+        broker._refresh_cooldown = 0.1  # Very short cooldown for testing
+
+        # First attempt fails
+        mock_resp_403 = AsyncMock()
+        mock_resp_403.status = 403
+        mock_resp_403.text = AsyncMock(return_value='{"error_code":"EGW00133"}')
+        mock_resp_403.__aenter__ = AsyncMock(return_value=mock_resp_403)
+        mock_resp_403.__aexit__ = AsyncMock(return_value=False)
+
+        # Second attempt succeeds
+        mock_resp_200 = AsyncMock()
+        mock_resp_200.status = 200
+        mock_resp_200.json = AsyncMock(
+            return_value={
+                "access_token": "tok_after_cooldown",
+                "expires_in": 86400,
+            }
+        )
+        mock_resp_200.__aenter__ = AsyncMock(return_value=mock_resp_200)
+        mock_resp_200.__aexit__ = AsyncMock(return_value=False)
+
+        with patch("aiohttp.ClientSession.post", return_value=mock_resp_403):
+            with pytest.raises(ConnectionError, match="Token refresh failed"):
+                await broker._ensure_token()
+
+        # Wait for cooldown to expire
+        await asyncio.sleep(0.15)
+
+        with patch("aiohttp.ClientSession.post", return_value=mock_resp_200):
+            token = await broker._ensure_token()
+            assert token == "tok_after_cooldown"
+
+        await broker.close()
+
 
 # ---------------------------------------------------------------------------
 # Network Error Handling
@@ -107,6 +211,38 @@ class TestRateLimiter:
         await broker._rate_limiter.acquire()
         await broker.close()
 
+    @pytest.mark.asyncio
+    async def test_send_order_acquires_rate_limiter_twice(self, settings):
+        """send_order must acquire rate limiter for both hash key and order call."""
+        broker = KISBroker(settings)
+        broker._access_token = "tok"
+        broker._token_expires_at = asyncio.get_event_loop().time() + 3600
+
+        # Mock hash key response
+        mock_hash_resp = AsyncMock()
+        mock_hash_resp.status = 200
+        mock_hash_resp.json = AsyncMock(return_value={"HASH": "abc123"})
+        mock_hash_resp.__aenter__ = AsyncMock(return_value=mock_hash_resp)
+        mock_hash_resp.__aexit__ = AsyncMock(return_value=False)
+
+        # Mock order response
+        mock_order_resp = AsyncMock()
+        mock_order_resp.status = 200
+        mock_order_resp.json = AsyncMock(return_value={"rt_cd": "0"})
+        mock_order_resp.__aenter__ = AsyncMock(return_value=mock_order_resp)
+        mock_order_resp.__aexit__ = AsyncMock(return_value=False)
+
+        with patch(
+            "aiohttp.ClientSession.post", side_effect=[mock_hash_resp, mock_order_resp]
+        ):
+            with patch.object(
+                broker._rate_limiter, "acquire", new_callable=AsyncMock
+            ) as mock_acquire:
+                await broker.send_order("005930", "BUY", 1, 50000)
+                assert mock_acquire.call_count == 2
+
+        await broker.close()
+
 
 # ---------------------------------------------------------------------------
 # Hash Key Generation
@@ -136,3 +272,671 @@ class TestHashKey:
             assert len(hash_key) > 0
 
         await broker.close()
+
+    @pytest.mark.asyncio
+    async def test_hash_key_acquires_rate_limiter(self, settings):
+        """_get_hash_key must go through the rate limiter to prevent burst."""
+        broker = KISBroker(settings)
+        broker._access_token = "tok"
+        broker._token_expires_at = asyncio.get_event_loop().time() + 3600
+
+        body = {"CANO": "12345678", "ACNT_PRDT_CD": "01"}
+
+        mock_resp = AsyncMock()
+        mock_resp.status = 200
+        mock_resp.json = AsyncMock(return_value={"HASH": "abc123hash"})
+        mock_resp.__aenter__ = AsyncMock(return_value=mock_resp)
+        mock_resp.__aexit__ = AsyncMock(return_value=False)
+
+        with patch("aiohttp.ClientSession.post", return_value=mock_resp):
+            with patch.object(
+                broker._rate_limiter, "acquire", new_callable=AsyncMock
+            ) as mock_acquire:
+                await broker._get_hash_key(body)
+                mock_acquire.assert_called_once()
+
+        await broker.close()
+
+
+# ---------------------------------------------------------------------------
+# fetch_market_rankings — TR_ID, path, params (issue #155)
+# ---------------------------------------------------------------------------
+
+
+def _make_ranking_mock(items: list[dict]) -> AsyncMock:
+    """Build a mock HTTP response returning ranking items."""
+    mock_resp = AsyncMock()
+    mock_resp.status = 200
+    mock_resp.json = AsyncMock(return_value={"output": items})
+    mock_resp.__aenter__ = AsyncMock(return_value=mock_resp)
+    mock_resp.__aexit__ = AsyncMock(return_value=False)
+    return mock_resp
+
+
+class TestFetchMarketRankings:
+    """Verify correct TR_ID, API path, and params per ranking_type (issue #155)."""
+
+    @pytest.fixture
+    def broker(self, settings) -> KISBroker:
+        b = KISBroker(settings)
+        b._access_token = "tok"
+        b._token_expires_at = float("inf")
+        b._rate_limiter.acquire = AsyncMock()
+        return b
+
+    @pytest.mark.asyncio
+    async def test_volume_uses_correct_tr_id_and_path(self, broker: KISBroker) -> None:
+        mock_resp = _make_ranking_mock([])
+        with patch("aiohttp.ClientSession.get", return_value=mock_resp) as mock_get:
+            await broker.fetch_market_rankings(ranking_type="volume")
+
+        call_kwargs = mock_get.call_args
+        url = call_kwargs[0][0] if call_kwargs[0] else call_kwargs[1].get("url", "")
+        headers = call_kwargs[1].get("headers", {})
+        params = call_kwargs[1].get("params", {})
+
+        assert "volume-rank" in url
+        assert headers.get("tr_id") == "FHPST01710000"
+        assert params.get("FID_COND_SCR_DIV_CODE") == "20171"
+        assert params.get("FID_TRGT_EXLS_CLS_CODE") == "0000000000"
+
+    @pytest.mark.asyncio
+    async def test_fluctuation_uses_correct_tr_id_and_path(self, broker: KISBroker) -> None:
+        mock_resp = _make_ranking_mock([])
+        with patch("aiohttp.ClientSession.get", return_value=mock_resp) as mock_get:
+            await broker.fetch_market_rankings(ranking_type="fluctuation")
+
+        call_kwargs = mock_get.call_args
+        url = call_kwargs[0][0] if call_kwargs[0] else call_kwargs[1].get("url", "")
+        headers = call_kwargs[1].get("headers", {})
+        params = call_kwargs[1].get("params", {})
+
+        assert "ranking/fluctuation" in url
+        assert headers.get("tr_id") == "FHPST01700000"
+        assert params.get("fid_cond_scr_div_code") == "20170"
+        # 실전 API는 4자리("0000") 거부 — 1자리("0")여야 한다 (#240)
+        assert params.get("fid_rank_sort_cls_code") == "0"
+
+    @pytest.mark.asyncio
+    async def test_volume_returns_parsed_rows(self, broker: KISBroker) -> None:
+        items = [
+            {
+                "mksc_shrn_iscd": "005930",
+                "hts_kor_isnm": "삼성전자",
+                "stck_prpr": "75000",
+                "acml_vol": "10000000",
+                "prdy_ctrt": "2.5",
+                "vol_inrt": "150",
+            }
+        ]
+        mock_resp = _make_ranking_mock(items)
+        with patch("aiohttp.ClientSession.get", return_value=mock_resp):
+            result = await broker.fetch_market_rankings(ranking_type="volume")
+
+        assert len(result) == 1
+        assert result[0]["stock_code"] == "005930"
+        assert result[0]["price"] == 75000.0
+        assert result[0]["change_rate"] == 2.5
+
+    @pytest.mark.asyncio
+    async def test_fluctuation_parses_stck_shrn_iscd(self, broker: KISBroker) -> None:
+        """실전 API는 mksc_shrn_iscd 대신 stck_shrn_iscd를 반환한다 (#240)."""
+        items = [
+            {
+                "stck_shrn_iscd": "015260",
+                "hts_kor_isnm": "에이엔피",
+                "stck_prpr": "794",
+                "acml_vol": "4896196",
+                "prdy_ctrt": "29.74",
+                "vol_inrt": "0",
+            }
+        ]
+        mock_resp = _make_ranking_mock(items)
+        with patch("aiohttp.ClientSession.get", return_value=mock_resp):
+            result = await broker.fetch_market_rankings(ranking_type="fluctuation")
+
+        assert len(result) == 1
+        assert result[0]["stock_code"] == "015260"
+        assert result[0]["change_rate"] == 29.74
+
+
+# ---------------------------------------------------------------------------
+# KRX tick unit / round-down helpers (issue #157)
+# ---------------------------------------------------------------------------
+
+
+from src.broker.kis_api import kr_tick_unit, kr_round_down  # noqa: E402
+
+
+class TestKrTickUnit:
+    """kr_tick_unit and kr_round_down must implement KRX price tick rules."""
+
+    @pytest.mark.parametrize(
+        "price, expected_tick",
+        [
+            (1999, 1),
+            (2000, 5),
+            (4999, 5),
+            (5000, 10),
+            (19999, 10),
+            (20000, 50),
+            (49999, 50),
+            (50000, 100),
+            (199999, 100),
+            (200000, 500),
+            (499999, 500),
+            (500000, 1000),
+            (1000000, 1000),
+        ],
+    )
+    def test_tick_unit_boundaries(self, price: int, expected_tick: int) -> None:
+        assert kr_tick_unit(price) == expected_tick
+
+    @pytest.mark.parametrize(
+        "price, expected_rounded",
+        [
+            (188150, 188100),   # 100원 단위, 50원 잔여 → 내림
+            (188100, 188100),   # 이미 정렬됨
+            (75050, 75000),     # 100원 단위, 50원 잔여 → 내림
+            (49950, 49950),     # 50원 단위 정렬됨
+            (49960, 49950),     # 50원 단위, 10원 잔여 → 내림
+            (1999, 1999),       # 1원 단위 → 그대로
+            (5003, 5000),       # 10원 단위, 3원 잔여 → 내림
+        ],
+    )
+    def test_round_down_to_tick(self, price: int, expected_rounded: int) -> None:
+        assert kr_round_down(price) == expected_rounded
+
+
+# ---------------------------------------------------------------------------
+# get_current_price (issue #157)
+# ---------------------------------------------------------------------------
+
+
+class TestGetCurrentPrice:
+    """get_current_price must use inquire-price API and return (price, change, foreigner)."""
+
+    @pytest.fixture
+    def broker(self, settings) -> KISBroker:
+        b = KISBroker(settings)
+        b._access_token = "tok"
+        b._token_expires_at = float("inf")
+        b._rate_limiter.acquire = AsyncMock()
+        return b
+
+    @pytest.mark.asyncio
+    async def test_returns_correct_fields(self, broker: KISBroker) -> None:
+        mock_resp = AsyncMock()
+        mock_resp.status = 200
+        mock_resp.json = AsyncMock(
+            return_value={
+                "rt_cd": "0",
+                "output": {
+                    "stck_prpr": "188600",
+                    "prdy_ctrt": "3.97",
+                    "frgn_ntby_qty": "12345",
+                },
+            }
+        )
+        mock_resp.__aenter__ = AsyncMock(return_value=mock_resp)
+        mock_resp.__aexit__ = AsyncMock(return_value=False)
+
+        with patch("aiohttp.ClientSession.get", return_value=mock_resp) as mock_get:
+            price, change_pct, foreigner = await broker.get_current_price("005930")
+
+        assert price == 188600.0
+        assert change_pct == 3.97
+        assert foreigner == 12345.0
+
+        call_kwargs = mock_get.call_args
+        url = call_kwargs[0][0] if call_kwargs[0] else call_kwargs[1].get("url", "")
+        headers = call_kwargs[1].get("headers", {})
+        assert "inquire-price" in url
+        assert headers.get("tr_id") == "FHKST01010100"
+
+    @pytest.mark.asyncio
+    async def test_http_error_raises_connection_error(self, broker: KISBroker) -> None:
+        mock_resp = AsyncMock()
+        mock_resp.status = 500
+        mock_resp.text = AsyncMock(return_value="Internal Server Error")
+        mock_resp.__aenter__ = AsyncMock(return_value=mock_resp)
+        mock_resp.__aexit__ = AsyncMock(return_value=False)
+
+        with patch("aiohttp.ClientSession.get", return_value=mock_resp):
+            with pytest.raises(ConnectionError, match="get_current_price failed"):
+                await broker.get_current_price("005930")
+
+
+# ---------------------------------------------------------------------------
+# send_order tick rounding and ORD_DVSN (issue #157)
+# ---------------------------------------------------------------------------
+
+
+class TestSendOrderTickRounding:
+    """send_order must apply KRX tick rounding and correct ORD_DVSN codes."""
+
+    @pytest.fixture
+    def broker(self, settings) -> KISBroker:
+        b = KISBroker(settings)
+        b._access_token = "tok"
+        b._token_expires_at = float("inf")
+        b._rate_limiter.acquire = AsyncMock()
+        return b
+
+    @pytest.mark.asyncio
+    async def test_limit_order_rounds_down_to_tick(self, broker: KISBroker) -> None:
+        """Price 188150 (not on 100-won tick) must be rounded to 188100."""
+        mock_hash = AsyncMock()
+        mock_hash.status = 200
+        mock_hash.json = AsyncMock(return_value={"HASH": "h"})
+        mock_hash.__aenter__ = AsyncMock(return_value=mock_hash)
+        mock_hash.__aexit__ = AsyncMock(return_value=False)
+
+        mock_order = AsyncMock()
+        mock_order.status = 200
+        mock_order.json = AsyncMock(return_value={"rt_cd": "0"})
+        mock_order.__aenter__ = AsyncMock(return_value=mock_order)
+        mock_order.__aexit__ = AsyncMock(return_value=False)
+
+        with patch(
+            "aiohttp.ClientSession.post", side_effect=[mock_hash, mock_order]
+        ) as mock_post:
+            await broker.send_order("005930", "BUY", 1, price=188150)
+
+        order_call = mock_post.call_args_list[1]
+        body = order_call[1].get("json", {})
+        assert body["ORD_UNPR"] == "188100"  # rounded down
+        assert body["ORD_DVSN"] == "00"       # 지정가
+
+    @pytest.mark.asyncio
+    async def test_limit_order_ord_dvsn_is_00(self, broker: KISBroker) -> None:
+        """send_order with price>0 must use ORD_DVSN='00' (지정가)."""
+        mock_hash = AsyncMock()
+        mock_hash.status = 200
+        mock_hash.json = AsyncMock(return_value={"HASH": "h"})
+        mock_hash.__aenter__ = AsyncMock(return_value=mock_hash)
+        mock_hash.__aexit__ = AsyncMock(return_value=False)
+
+        mock_order = AsyncMock()
+        mock_order.status = 200
+        mock_order.json = AsyncMock(return_value={"rt_cd": "0"})
+        mock_order.__aenter__ = AsyncMock(return_value=mock_order)
+        mock_order.__aexit__ = AsyncMock(return_value=False)
+
+        with patch(
+            "aiohttp.ClientSession.post", side_effect=[mock_hash, mock_order]
+        ) as mock_post:
+            await broker.send_order("005930", "BUY", 1, price=50000)
+
+        order_call = mock_post.call_args_list[1]
+        body = order_call[1].get("json", {})
+        assert body["ORD_DVSN"] == "00"
+
+    @pytest.mark.asyncio
+    async def test_market_order_ord_dvsn_is_01(self, broker: KISBroker) -> None:
+        """send_order with price=0 must use ORD_DVSN='01' (시장가)."""
+        mock_hash = AsyncMock()
+        mock_hash.status = 200
+        mock_hash.json = AsyncMock(return_value={"HASH": "h"})
+        mock_hash.__aenter__ = AsyncMock(return_value=mock_hash)
+        mock_hash.__aexit__ = AsyncMock(return_value=False)
+
+        mock_order = AsyncMock()
+        mock_order.status = 200
+        mock_order.json = AsyncMock(return_value={"rt_cd": "0"})
+        mock_order.__aenter__ = AsyncMock(return_value=mock_order)
+        mock_order.__aexit__ = AsyncMock(return_value=False)
+
+        with patch(
+            "aiohttp.ClientSession.post", side_effect=[mock_hash, mock_order]
+        ) as mock_post:
+            await broker.send_order("005930", "SELL", 1, price=0)
+
+        order_call = mock_post.call_args_list[1]
+        body = order_call[1].get("json", {})
+        assert body["ORD_DVSN"] == "01"
+
+
+# ---------------------------------------------------------------------------
+# TR_ID live/paper branching (issues #201, #202, #203)
+# ---------------------------------------------------------------------------
+
+
+class TestTRIDBranchingDomestic:
+    """get_balance and send_order must use correct TR_ID for live vs paper mode."""
+
+    def _make_broker(self, settings, mode: str) -> KISBroker:
+        from src.config import Settings
+
+        s = Settings(
+            KIS_APP_KEY=settings.KIS_APP_KEY,
+            KIS_APP_SECRET=settings.KIS_APP_SECRET,
+            KIS_ACCOUNT_NO=settings.KIS_ACCOUNT_NO,
+            GEMINI_API_KEY=settings.GEMINI_API_KEY,
+            DB_PATH=":memory:",
+            ENABLED_MARKETS="KR",
+            MODE=mode,
+        )
+        b = KISBroker(s)
+        b._access_token = "tok"
+        b._token_expires_at = float("inf")
+        b._rate_limiter.acquire = AsyncMock()
+        return b
+
+    @pytest.mark.asyncio
+    async def test_get_balance_paper_uses_vttc8434r(self, settings) -> None:
+        broker = self._make_broker(settings, "paper")
+        mock_resp = AsyncMock()
+        mock_resp.status = 200
+        mock_resp.json = AsyncMock(
+            return_value={"output1": [], "output2": {}}
+        )
+        mock_resp.__aenter__ = AsyncMock(return_value=mock_resp)
+        mock_resp.__aexit__ = AsyncMock(return_value=False)
+
+        with patch("aiohttp.ClientSession.get", return_value=mock_resp) as mock_get:
+            await broker.get_balance()
+
+        headers = mock_get.call_args[1].get("headers", {})
+        assert headers["tr_id"] == "VTTC8434R"
+
+    @pytest.mark.asyncio
+    async def test_get_balance_live_uses_tttc8434r(self, settings) -> None:
+        broker = self._make_broker(settings, "live")
+        mock_resp = AsyncMock()
+        mock_resp.status = 200
+        mock_resp.json = AsyncMock(
+            return_value={"output1": [], "output2": {}}
+        )
+        mock_resp.__aenter__ = AsyncMock(return_value=mock_resp)
+        mock_resp.__aexit__ = AsyncMock(return_value=False)
+
+        with patch("aiohttp.ClientSession.get", return_value=mock_resp) as mock_get:
+            await broker.get_balance()
+
+        headers = mock_get.call_args[1].get("headers", {})
+        assert headers["tr_id"] == "TTTC8434R"
+
+    @pytest.mark.asyncio
+    async def test_send_order_buy_paper_uses_vttc0012u(self, settings) -> None:
+        broker = self._make_broker(settings, "paper")
+        mock_hash = AsyncMock()
+        mock_hash.status = 200
+        mock_hash.json = AsyncMock(return_value={"HASH": "h"})
+        mock_hash.__aenter__ = AsyncMock(return_value=mock_hash)
+        mock_hash.__aexit__ = AsyncMock(return_value=False)
+
+        mock_order = AsyncMock()
+        mock_order.status = 200
+        mock_order.json = AsyncMock(return_value={"rt_cd": "0"})
+        mock_order.__aenter__ = AsyncMock(return_value=mock_order)
+        mock_order.__aexit__ = AsyncMock(return_value=False)
+
+        with patch(
+            "aiohttp.ClientSession.post", side_effect=[mock_hash, mock_order]
+        ) as mock_post:
+            await broker.send_order("005930", "BUY", 1)
+
+        order_headers = mock_post.call_args_list[1][1].get("headers", {})
+        assert order_headers["tr_id"] == "VTTC0012U"
+
+    @pytest.mark.asyncio
+    async def test_send_order_buy_live_uses_tttc0012u(self, settings) -> None:
+        broker = self._make_broker(settings, "live")
+        mock_hash = AsyncMock()
+        mock_hash.status = 200
+        mock_hash.json = AsyncMock(return_value={"HASH": "h"})
+        mock_hash.__aenter__ = AsyncMock(return_value=mock_hash)
+        mock_hash.__aexit__ = AsyncMock(return_value=False)
+
+        mock_order = AsyncMock()
+        mock_order.status = 200
+        mock_order.json = AsyncMock(return_value={"rt_cd": "0"})
+        mock_order.__aenter__ = AsyncMock(return_value=mock_order)
+        mock_order.__aexit__ = AsyncMock(return_value=False)
+
+        with patch(
+            "aiohttp.ClientSession.post", side_effect=[mock_hash, mock_order]
+        ) as mock_post:
+            await broker.send_order("005930", "BUY", 1)
+
+        order_headers = mock_post.call_args_list[1][1].get("headers", {})
+        assert order_headers["tr_id"] == "TTTC0012U"
+
+    @pytest.mark.asyncio
+    async def test_send_order_sell_paper_uses_vttc0011u(self, settings) -> None:
+        broker = self._make_broker(settings, "paper")
+        mock_hash = AsyncMock()
+        mock_hash.status = 200
+        mock_hash.json = AsyncMock(return_value={"HASH": "h"})
+        mock_hash.__aenter__ = AsyncMock(return_value=mock_hash)
+        mock_hash.__aexit__ = AsyncMock(return_value=False)
+
+        mock_order = AsyncMock()
+        mock_order.status = 200
+        mock_order.json = AsyncMock(return_value={"rt_cd": "0"})
+        mock_order.__aenter__ = AsyncMock(return_value=mock_order)
+        mock_order.__aexit__ = AsyncMock(return_value=False)
+
+        with patch(
+            "aiohttp.ClientSession.post", side_effect=[mock_hash, mock_order]
+        ) as mock_post:
+            await broker.send_order("005930", "SELL", 1)
+
+        order_headers = mock_post.call_args_list[1][1].get("headers", {})
+        assert order_headers["tr_id"] == "VTTC0011U"
+
+    @pytest.mark.asyncio
+    async def test_send_order_sell_live_uses_tttc0011u(self, settings) -> None:
+        broker = self._make_broker(settings, "live")
+        mock_hash = AsyncMock()
+        mock_hash.status = 200
+        mock_hash.json = AsyncMock(return_value={"HASH": "h"})
+        mock_hash.__aenter__ = AsyncMock(return_value=mock_hash)
+        mock_hash.__aexit__ = AsyncMock(return_value=False)
+
+        mock_order = AsyncMock()
+        mock_order.status = 200
+        mock_order.json = AsyncMock(return_value={"rt_cd": "0"})
+        mock_order.__aenter__ = AsyncMock(return_value=mock_order)
+        mock_order.__aexit__ = AsyncMock(return_value=False)
+
+        with patch(
+            "aiohttp.ClientSession.post", side_effect=[mock_hash, mock_order]
+        ) as mock_post:
+            await broker.send_order("005930", "SELL", 1)
+
+        order_headers = mock_post.call_args_list[1][1].get("headers", {})
+        assert order_headers["tr_id"] == "TTTC0011U"
+
+
+# ---------------------------------------------------------------------------
+# Domestic Pending Orders (get_domestic_pending_orders)
+# ---------------------------------------------------------------------------
+
+
+class TestGetDomesticPendingOrders:
+    """get_domestic_pending_orders must return [] in paper mode and call TTTC0084R in live."""
+
+    def _make_broker(self, settings, mode: str) -> KISBroker:
+        from src.config import Settings
+
+        s = Settings(
+            KIS_APP_KEY=settings.KIS_APP_KEY,
+            KIS_APP_SECRET=settings.KIS_APP_SECRET,
+            KIS_ACCOUNT_NO=settings.KIS_ACCOUNT_NO,
+            GEMINI_API_KEY=settings.GEMINI_API_KEY,
+            DB_PATH=":memory:",
+            ENABLED_MARKETS="KR",
+            MODE=mode,
+        )
+        b = KISBroker(s)
+        b._access_token = "tok"
+        b._token_expires_at = float("inf")
+        b._rate_limiter.acquire = AsyncMock()
+        return b
+
+    @pytest.mark.asyncio
+    async def test_paper_mode_returns_empty(self, settings) -> None:
+        """Paper mode must return [] immediately without any API call."""
+        broker = self._make_broker(settings, "paper")
+
+        with patch("aiohttp.ClientSession.get") as mock_get:
+            result = await broker.get_domestic_pending_orders()
+
+        assert result == []
+        mock_get.assert_not_called()
+
+    @pytest.mark.asyncio
+    async def test_live_mode_calls_tttc0084r_with_correct_params(
+        self, settings
+    ) -> None:
+        """Live mode must call TTTC0084R with INQR_DVSN_1/2 and paging params."""
+        broker = self._make_broker(settings, "live")
+        pending = [{"odno": "001", "pdno": "005930", "psbl_qty": "10"}]
+        mock_resp = AsyncMock()
+        mock_resp.status = 200
+        mock_resp.json = AsyncMock(return_value={"output": pending})
+        mock_resp.__aenter__ = AsyncMock(return_value=mock_resp)
+        mock_resp.__aexit__ = AsyncMock(return_value=False)
+
+        with patch("aiohttp.ClientSession.get", return_value=mock_resp) as mock_get:
+            result = await broker.get_domestic_pending_orders()
+
+        assert result == pending
+        headers = mock_get.call_args[1].get("headers", {})
+        assert headers["tr_id"] == "TTTC0084R"
+        params = mock_get.call_args[1].get("params", {})
+        assert params["INQR_DVSN_1"] == "0"
+        assert params["INQR_DVSN_2"] == "0"
+
+    @pytest.mark.asyncio
+    async def test_live_mode_connection_error(self, settings) -> None:
+        """Network error must raise ConnectionError."""
+        import aiohttp as _aiohttp
+
+        broker = self._make_broker(settings, "live")
+
+        with patch(
+            "aiohttp.ClientSession.get",
+            side_effect=_aiohttp.ClientError("timeout"),
+        ):
+            with pytest.raises(ConnectionError):
+                await broker.get_domestic_pending_orders()
+
+
+# ---------------------------------------------------------------------------
+# Domestic Order Cancellation (cancel_domestic_order)
+# ---------------------------------------------------------------------------
+
+
+class TestCancelDomesticOrder:
+    """cancel_domestic_order must use correct TR_ID and build body correctly."""
+
+    def _make_broker(self, settings, mode: str) -> KISBroker:
+        from src.config import Settings
+
+        s = Settings(
+            KIS_APP_KEY=settings.KIS_APP_KEY,
+            KIS_APP_SECRET=settings.KIS_APP_SECRET,
+            KIS_ACCOUNT_NO=settings.KIS_ACCOUNT_NO,
+            GEMINI_API_KEY=settings.GEMINI_API_KEY,
+            DB_PATH=":memory:",
+            ENABLED_MARKETS="KR",
+            MODE=mode,
+        )
+        b = KISBroker(s)
+        b._access_token = "tok"
+        b._token_expires_at = float("inf")
+        b._rate_limiter.acquire = AsyncMock()
+        return b
+
+    def _make_post_mocks(self, order_payload: dict) -> tuple:
+        mock_hash = AsyncMock()
+        mock_hash.status = 200
+        mock_hash.json = AsyncMock(return_value={"HASH": "h"})
+        mock_hash.__aenter__ = AsyncMock(return_value=mock_hash)
+        mock_hash.__aexit__ = AsyncMock(return_value=False)
+
+        mock_order = AsyncMock()
+        mock_order.status = 200
+        mock_order.json = AsyncMock(return_value=order_payload)
+        mock_order.__aenter__ = AsyncMock(return_value=mock_order)
+        mock_order.__aexit__ = AsyncMock(return_value=False)
+
+        return mock_hash, mock_order
+
+    @pytest.mark.asyncio
+    async def test_live_uses_tttc0013u(self, settings) -> None:
+        """Live mode must use TR_ID TTTC0013U."""
+        broker = self._make_broker(settings, "live")
+        mock_hash, mock_order = self._make_post_mocks({"rt_cd": "0"})
+
+        with patch(
+            "aiohttp.ClientSession.post", side_effect=[mock_hash, mock_order]
+        ) as mock_post:
+            await broker.cancel_domestic_order("005930", "ORD001", "BRNO01", 5)
+
+        order_headers = mock_post.call_args_list[1][1].get("headers", {})
+        assert order_headers["tr_id"] == "TTTC0013U"
+
+    @pytest.mark.asyncio
+    async def test_paper_uses_vttc0013u(self, settings) -> None:
+        """Paper mode must use TR_ID VTTC0013U."""
+        broker = self._make_broker(settings, "paper")
+        mock_hash, mock_order = self._make_post_mocks({"rt_cd": "0"})
+
+        with patch(
+            "aiohttp.ClientSession.post", side_effect=[mock_hash, mock_order]
+        ) as mock_post:
+            await broker.cancel_domestic_order("005930", "ORD001", "BRNO01", 5)
+
+        order_headers = mock_post.call_args_list[1][1].get("headers", {})
+        assert order_headers["tr_id"] == "VTTC0013U"
+
+    @pytest.mark.asyncio
+    async def test_cancel_sets_rvse_cncl_dvsn_cd_02(self, settings) -> None:
+        """Body must have RVSE_CNCL_DVSN_CD='02' (취소) and QTY_ALL_ORD_YN='Y'."""
+        broker = self._make_broker(settings, "live")
+        mock_hash, mock_order = self._make_post_mocks({"rt_cd": "0"})
+
+        with patch(
+            "aiohttp.ClientSession.post", side_effect=[mock_hash, mock_order]
+        ) as mock_post:
+            await broker.cancel_domestic_order("005930", "ORD001", "BRNO01", 5)
+
+        body = mock_post.call_args_list[1][1].get("json", {})
+        assert body["RVSE_CNCL_DVSN_CD"] == "02"
+        assert body["QTY_ALL_ORD_YN"] == "Y"
+        assert body["ORD_UNPR"] == "0"
+
+    @pytest.mark.asyncio
+    async def test_cancel_sets_krx_fwdg_ord_orgno_in_body(self, settings) -> None:
+        """Body must include KRX_FWDG_ORD_ORGNO and ORGN_ODNO from arguments."""
+        broker = self._make_broker(settings, "live")
+        mock_hash, mock_order = self._make_post_mocks({"rt_cd": "0"})
+
+        with patch(
+            "aiohttp.ClientSession.post", side_effect=[mock_hash, mock_order]
+        ) as mock_post:
+            await broker.cancel_domestic_order("005930", "ORD123", "BRN456", 3)
+
+        body = mock_post.call_args_list[1][1].get("json", {})
+        assert body["KRX_FWDG_ORD_ORGNO"] == "BRN456"
+        assert body["ORGN_ODNO"] == "ORD123"
+        assert body["ORD_QTY"] == "3"
+
+    @pytest.mark.asyncio
+    async def test_cancel_sets_hashkey_header(self, settings) -> None:
+        """Request must include hashkey header (same pattern as send_order)."""
+        broker = self._make_broker(settings, "live")
+        mock_hash, mock_order = self._make_post_mocks({"rt_cd": "0"})
+
+        with patch(
+            "aiohttp.ClientSession.post", side_effect=[mock_hash, mock_order]
+        ) as mock_post:
+            await broker.cancel_domestic_order("005930", "ORD001", "BRNO01", 2)
+
+        order_headers = mock_post.call_args_list[1][1].get("headers", {})
+        assert "hashkey" in order_headers
+        assert order_headers["hashkey"] == "h"

tests/test_context.py

@@ -0,0 +1,629 @@
+"""Tests for the multi-layered context management system."""
+
+from __future__ import annotations
+
+import sqlite3
+from datetime import UTC, datetime, timedelta
+
+import pytest
+
+from src.context.aggregator import ContextAggregator
+from src.context.layer import LAYER_CONFIG, ContextLayer
+from src.context.store import ContextStore
+from src.context.summarizer import ContextSummarizer
+from src.db import init_db, log_trade
+
+
+@pytest.fixture
+def db_conn() -> sqlite3.Connection:
+    """Provide an in-memory database connection."""
+    return init_db(":memory:")
+
+
+@pytest.fixture
+def store(db_conn: sqlite3.Connection) -> ContextStore:
+    """Provide a ContextStore instance."""
+    return ContextStore(db_conn)
+
+
+@pytest.fixture
+def aggregator(db_conn: sqlite3.Connection) -> ContextAggregator:
+    """Provide a ContextAggregator instance."""
+    return ContextAggregator(db_conn)
+
+
+class TestContextStore:
+    """Test suite for ContextStore CRUD operations."""
+
+    def test_set_and_get_context(self, store: ContextStore) -> None:
+        """Test setting and retrieving a context value."""
+        store.set_context(ContextLayer.L6_DAILY, "2026-02-04", "total_pnl", 1234.56)
+
+        value = store.get_context(ContextLayer.L6_DAILY, "2026-02-04", "total_pnl")
+        assert value == 1234.56
+
+    def test_get_nonexistent_context(self, store: ContextStore) -> None:
+        """Test retrieving a non-existent context returns None."""
+        value = store.get_context(ContextLayer.L6_DAILY, "2026-02-04", "nonexistent")
+        assert value is None
+
+    def test_update_existing_context(self, store: ContextStore) -> None:
+        """Test updating an existing context value."""
+        store.set_context(ContextLayer.L6_DAILY, "2026-02-04", "total_pnl", 100.0)
+        store.set_context(ContextLayer.L6_DAILY, "2026-02-04", "total_pnl", 200.0)
+
+        value = store.get_context(ContextLayer.L6_DAILY, "2026-02-04", "total_pnl")
+        assert value == 200.0
+
+    def test_get_all_contexts_for_layer(self, store: ContextStore) -> None:
+        """Test retrieving all contexts for a specific layer."""
+        store.set_context(ContextLayer.L6_DAILY, "2026-02-04", "total_pnl", 100.0)
+        store.set_context(ContextLayer.L6_DAILY, "2026-02-04", "trade_count", 10)
+        store.set_context(ContextLayer.L6_DAILY, "2026-02-04", "win_rate", 60.5)
+
+        contexts = store.get_all_contexts(ContextLayer.L6_DAILY, "2026-02-04")
+        assert len(contexts) == 3
+        assert contexts["total_pnl"] == 100.0
+        assert contexts["trade_count"] == 10
+        assert contexts["win_rate"] == 60.5
+
+    def test_get_latest_timeframe(self, store: ContextStore) -> None:
+        """Test getting the most recent timeframe for a layer."""
+        store.set_context(ContextLayer.L6_DAILY, "2026-02-01", "total_pnl", 100.0)
+        store.set_context(ContextLayer.L6_DAILY, "2026-02-03", "total_pnl", 200.0)
+        store.set_context(ContextLayer.L6_DAILY, "2026-02-02", "total_pnl", 150.0)
+
+        latest = store.get_latest_timeframe(ContextLayer.L6_DAILY)
+        # Latest by updated_at, which should be the last one set
+        assert latest == "2026-02-02"
+
+    def test_delete_old_contexts(
+        self, store: ContextStore, db_conn: sqlite3.Connection
+    ) -> None:
+        """Test deleting contexts older than a cutoff date."""
+        # Insert contexts with specific old timestamps
+        # (bypassing set_context which uses current time)
+        old_date = "2026-01-01T00:00:00+00:00"
+        new_date = "2026-02-01T00:00:00+00:00"
+
+        db_conn.execute(
+            """
+            INSERT INTO contexts (layer, timeframe, key, value, created_at, updated_at)
+            VALUES (?, ?, ?, ?, ?, ?)
+            """,
+            (ContextLayer.L6_DAILY.value, "2026-01-01", "total_pnl", "100.0", old_date, old_date),
+        )
+        db_conn.execute(
+            """
+            INSERT INTO contexts (layer, timeframe, key, value, created_at, updated_at)
+            VALUES (?, ?, ?, ?, ?, ?)
+            """,
+            (ContextLayer.L6_DAILY.value, "2026-02-01", "total_pnl", "200.0", new_date, new_date),
+        )
+        db_conn.commit()
+
+        # Delete contexts before 2026-01-15
+        cutoff = "2026-01-15T00:00:00+00:00"
+        deleted = store.delete_old_contexts(ContextLayer.L6_DAILY, cutoff)
+
+        # Should delete the 2026-01-01 context
+        assert deleted == 1
+        assert store.get_context(ContextLayer.L6_DAILY, "2026-02-01", "total_pnl") == 200.0
+        assert store.get_context(ContextLayer.L6_DAILY, "2026-01-01", "total_pnl") is None
+
+    def test_cleanup_expired_contexts(
+        self, store: ContextStore, db_conn: sqlite3.Connection
+    ) -> None:
+        """Test automatic cleanup based on retention policies."""
+        # Set old contexts for L7 (7 day retention)
+        old_date = (datetime.now(UTC) - timedelta(days=10)).isoformat()
+        db_conn.execute(
+            """
+            INSERT INTO contexts (layer, timeframe, key, value, created_at, updated_at)
+            VALUES (?, ?, ?, ?, ?, ?)
+            """,
+            (ContextLayer.L7_REALTIME.value, "2026-01-01", "price", "100.0", old_date, old_date),
+        )
+        db_conn.commit()
+
+        deleted_counts = store.cleanup_expired_contexts()
+
+        # Should delete the old L7 context (10 days > 7 day retention)
+        assert deleted_counts[ContextLayer.L7_REALTIME] == 1
+
+        # L1 has no retention limit, so nothing should be deleted
+        assert deleted_counts[ContextLayer.L1_LEGACY] == 0
+
+    def test_context_metadata_initialized(
+        self, store: ContextStore, db_conn: sqlite3.Connection
+    ) -> None:
+        """Test that context metadata is properly initialized."""
+        cursor = db_conn.execute("SELECT COUNT(*) FROM context_metadata")
+        count = cursor.fetchone()[0]
+
+        # Should have metadata for all 7 layers
+        assert count == 7
+
+        # Verify L1 metadata
+        cursor = db_conn.execute(
+            "SELECT description, retention_days FROM context_metadata WHERE layer = ?",
+            (ContextLayer.L1_LEGACY.value,),
+        )
+        row = cursor.fetchone()
+        assert row is not None
+        assert "Cumulative trading history" in row[0]
+        assert row[1] is None  # No retention limit for L1
+
+
+class TestContextAggregator:
+    """Test suite for ContextAggregator."""
+
+    def test_aggregate_daily_from_trades(
+        self, aggregator: ContextAggregator, db_conn: sqlite3.Connection
+    ) -> None:
+        """Test aggregating daily metrics from trades."""
+        date = datetime.now(UTC).date().isoformat()
+
+        # Create sample trades
+        log_trade(db_conn, "005930", "BUY", 85, "Good signal", quantity=10, price=70000, pnl=500)
+        log_trade(db_conn, "000660", "SELL", 90, "Take profit", quantity=5, price=50000, pnl=1500)
+        log_trade(db_conn, "035720", "HOLD", 75, "Wait", quantity=0, price=0, pnl=0)
+
+        # Manually set timestamps to the target date
+        db_conn.execute(
+            f"UPDATE trades SET timestamp = '{date}T10:00:00+00:00'"
+        )
+        db_conn.commit()
+
+        # Aggregate
+        aggregator.aggregate_daily_from_trades(date, market="KR")
+
+        # Verify L6 contexts
+        store = aggregator.store
+        assert store.get_context(ContextLayer.L6_DAILY, date, "trade_count_KR") == 3
+        assert store.get_context(ContextLayer.L6_DAILY, date, "buys_KR") == 1
+        assert store.get_context(ContextLayer.L6_DAILY, date, "sells_KR") == 1
+        assert store.get_context(ContextLayer.L6_DAILY, date, "holds_KR") == 1
+        assert store.get_context(ContextLayer.L6_DAILY, date, "total_pnl_KR") == 2000.0
+        assert store.get_context(ContextLayer.L6_DAILY, date, "unique_stocks_KR") == 3
+        # 2 wins, 0 losses
+        assert store.get_context(ContextLayer.L6_DAILY, date, "win_rate_KR") == 100.0
+
+    def test_aggregate_weekly_from_daily(self, aggregator: ContextAggregator) -> None:
+        """Test aggregating weekly metrics from daily."""
+        week = "2026-W06"
+
+        # Set daily contexts
+        aggregator.store.set_context(
+            ContextLayer.L6_DAILY, "2026-02-02", "total_pnl_KR", 100.0
+        )
+        aggregator.store.set_context(
+            ContextLayer.L6_DAILY, "2026-02-03", "total_pnl_KR", 200.0
+        )
+        aggregator.store.set_context(
+            ContextLayer.L6_DAILY, "2026-02-02", "avg_confidence_KR", 80.0
+        )
+        aggregator.store.set_context(
+            ContextLayer.L6_DAILY, "2026-02-03", "avg_confidence_KR", 85.0
+        )
+
+        # Aggregate
+        aggregator.aggregate_weekly_from_daily(week)
+
+        # Verify L5 contexts
+        store = aggregator.store
+        weekly_pnl = store.get_context(ContextLayer.L5_WEEKLY, week, "weekly_pnl_KR")
+        avg_conf = store.get_context(ContextLayer.L5_WEEKLY, week, "avg_confidence_KR")
+
+        assert weekly_pnl == 300.0
+        assert avg_conf == 82.5
+
+    def test_aggregate_monthly_from_weekly(self, aggregator: ContextAggregator) -> None:
+        """Test aggregating monthly metrics from weekly."""
+        month = "2026-02"
+
+        # Set weekly contexts
+        aggregator.store.set_context(
+            ContextLayer.L5_WEEKLY, "2026-W05", "weekly_pnl_KR", 100.0
+        )
+        aggregator.store.set_context(
+            ContextLayer.L5_WEEKLY, "2026-W06", "weekly_pnl_KR", 200.0
+        )
+        aggregator.store.set_context(
+            ContextLayer.L5_WEEKLY, "2026-W07", "weekly_pnl_KR", 150.0
+        )
+
+        # Aggregate
+        aggregator.aggregate_monthly_from_weekly(month)
+
+        # Verify L4 contexts
+        store = aggregator.store
+        monthly_pnl = store.get_context(ContextLayer.L4_MONTHLY, month, "monthly_pnl")
+        assert monthly_pnl == 450.0
+
+    def test_aggregate_quarterly_from_monthly(self, aggregator: ContextAggregator) -> None:
+        """Test aggregating quarterly metrics from monthly."""
+        quarter = "2026-Q1"
+
+        # Set monthly contexts for Q1 (Jan, Feb, Mar)
+        aggregator.store.set_context(ContextLayer.L4_MONTHLY, "2026-01", "monthly_pnl", 1000.0)
+        aggregator.store.set_context(ContextLayer.L4_MONTHLY, "2026-02", "monthly_pnl", 2000.0)
+        aggregator.store.set_context(ContextLayer.L4_MONTHLY, "2026-03", "monthly_pnl", 1500.0)
+
+        # Aggregate
+        aggregator.aggregate_quarterly_from_monthly(quarter)
+
+        # Verify L3 contexts
+        store = aggregator.store
+        quarterly_pnl = store.get_context(ContextLayer.L3_QUARTERLY, quarter, "quarterly_pnl")
+        assert quarterly_pnl == 4500.0
+
+    def test_aggregate_annual_from_quarterly(self, aggregator: ContextAggregator) -> None:
+        """Test aggregating annual metrics from quarterly."""
+        year = "2026"
+
+        # Set quarterly contexts for all 4 quarters
+        aggregator.store.set_context(ContextLayer.L3_QUARTERLY, "2026-Q1", "quarterly_pnl", 4500.0)
+        aggregator.store.set_context(ContextLayer.L3_QUARTERLY, "2026-Q2", "quarterly_pnl", 5000.0)
+        aggregator.store.set_context(ContextLayer.L3_QUARTERLY, "2026-Q3", "quarterly_pnl", 4800.0)
+        aggregator.store.set_context(ContextLayer.L3_QUARTERLY, "2026-Q4", "quarterly_pnl", 5200.0)
+
+        # Aggregate
+        aggregator.aggregate_annual_from_quarterly(year)
+
+        # Verify L2 contexts
+        store = aggregator.store
+        annual_pnl = store.get_context(ContextLayer.L2_ANNUAL, year, "annual_pnl")
+        assert annual_pnl == 19500.0
+
+    def test_aggregate_legacy_from_annual(self, aggregator: ContextAggregator) -> None:
+        """Test aggregating legacy metrics from all annual data."""
+        # Set annual contexts for multiple years
+        aggregator.store.set_context(ContextLayer.L2_ANNUAL, "2024", "annual_pnl", 10000.0)
+        aggregator.store.set_context(ContextLayer.L2_ANNUAL, "2025", "annual_pnl", 15000.0)
+        aggregator.store.set_context(ContextLayer.L2_ANNUAL, "2026", "annual_pnl", 20000.0)
+
+        # Aggregate
+        aggregator.aggregate_legacy_from_annual()
+
+        # Verify L1 contexts
+        store = aggregator.store
+        total_pnl = store.get_context(ContextLayer.L1_LEGACY, "LEGACY", "total_pnl")
+        years_traded = store.get_context(ContextLayer.L1_LEGACY, "LEGACY", "years_traded")
+        avg_annual_pnl = store.get_context(ContextLayer.L1_LEGACY, "LEGACY", "avg_annual_pnl")
+
+        assert total_pnl == 45000.0
+        assert years_traded == 3
+        assert avg_annual_pnl == 15000.0
+
+    def test_run_all_aggregations(
+        self, aggregator: ContextAggregator, db_conn: sqlite3.Connection
+    ) -> None:
+        """Test running all aggregations from L7 to L1."""
+        date = datetime.now(UTC).date().isoformat()
+
+        # Create sample trades
+        log_trade(db_conn, "005930", "BUY", 85, "Good signal", quantity=10, price=70000, pnl=1000)
+
+        # Set timestamp
+        db_conn.execute(f"UPDATE trades SET timestamp = '{date}T10:00:00+00:00'")
+        db_conn.commit()
+
+        # Run all aggregations
+        aggregator.run_all_aggregations()
+
+        # Verify data exists in each layer
+        store = aggregator.store
+        assert store.get_context(ContextLayer.L6_DAILY, date, "total_pnl_KR") == 1000.0
+        from datetime import date as date_cls
+        trade_date = date_cls.fromisoformat(date)
+        iso_year, iso_week, _ = trade_date.isocalendar()
+        trade_week = f"{iso_year}-W{iso_week:02d}"
+        assert store.get_context(ContextLayer.L5_WEEKLY, trade_week, "weekly_pnl_KR") is not None
+        trade_month = f"{trade_date.year}-{trade_date.month:02d}"
+        trade_quarter = f"{trade_date.year}-Q{(trade_date.month - 1) // 3 + 1}"
+        trade_year = str(trade_date.year)
+        assert store.get_context(ContextLayer.L4_MONTHLY, trade_month, "monthly_pnl") == 1000.0
+        assert store.get_context(ContextLayer.L3_QUARTERLY, trade_quarter, "quarterly_pnl") == 1000.0
+        assert store.get_context(ContextLayer.L2_ANNUAL, trade_year, "annual_pnl") == 1000.0
+
+
+class TestLayerMetadata:
+    """Test suite for layer metadata configuration."""
+
+    def test_all_layers_have_metadata(self) -> None:
+        """Test that all 7 layers have metadata defined."""
+        assert len(LAYER_CONFIG) == 7
+
+        for layer in ContextLayer:
+            assert layer in LAYER_CONFIG
+
+    def test_layer_retention_policies(self) -> None:
+        """Test layer retention policies are correctly configured."""
+        # L1 should have no retention limit
+        assert LAYER_CONFIG[ContextLayer.L1_LEGACY].retention_days is None
+
+        # L7 should have the shortest retention (7 days)
+        assert LAYER_CONFIG[ContextLayer.L7_REALTIME].retention_days == 7
+
+        # L2 should have a long retention (10 years)
+        assert LAYER_CONFIG[ContextLayer.L2_ANNUAL].retention_days == 365 * 10
+
+    def test_layer_aggregation_chain(self) -> None:
+        """Test that the aggregation chain is properly configured."""
+        # L7 has no source (leaf layer)
+        assert LAYER_CONFIG[ContextLayer.L7_REALTIME].aggregation_source is None
+
+        # L6 aggregates from L7
+        assert LAYER_CONFIG[ContextLayer.L6_DAILY].aggregation_source == ContextLayer.L7_REALTIME
+
+        # L5 aggregates from L6
+        assert LAYER_CONFIG[ContextLayer.L5_WEEKLY].aggregation_source == ContextLayer.L6_DAILY
+
+        # L4 aggregates from L5
+        assert LAYER_CONFIG[ContextLayer.L4_MONTHLY].aggregation_source == ContextLayer.L5_WEEKLY
+
+        # L3 aggregates from L4
+        assert LAYER_CONFIG[ContextLayer.L3_QUARTERLY].aggregation_source == ContextLayer.L4_MONTHLY
+
+        # L2 aggregates from L3
+        assert LAYER_CONFIG[ContextLayer.L2_ANNUAL].aggregation_source == ContextLayer.L3_QUARTERLY
+
+        # L1 aggregates from L2
+        assert LAYER_CONFIG[ContextLayer.L1_LEGACY].aggregation_source == ContextLayer.L2_ANNUAL
+
+
+# ---------------------------------------------------------------------------
+# ContextSummarizer tests
+# ---------------------------------------------------------------------------
+
+
+@pytest.fixture
+def summarizer(db_conn: sqlite3.Connection) -> ContextSummarizer:
+    """Provide a ContextSummarizer backed by an in-memory store."""
+    return ContextSummarizer(ContextStore(db_conn))
+
+
+class TestContextSummarizer:
+    """Test suite for ContextSummarizer."""
+
+    # ------------------------------------------------------------------
+    # summarize_numeric_values
+    # ------------------------------------------------------------------
+
+    def test_summarize_empty_values(self, summarizer: ContextSummarizer) -> None:
+        """Empty list must return SummaryStats with count=0 and no other fields."""
+        stats = summarizer.summarize_numeric_values([])
+        assert stats.count == 0
+        assert stats.mean is None
+        assert stats.min is None
+        assert stats.max is None
+
+    def test_summarize_single_value(self, summarizer: ContextSummarizer) -> None:
+        """Single-element list must return correct stats with std=0 and trend=flat."""
+        stats = summarizer.summarize_numeric_values([42.0])
+        assert stats.count == 1
+        assert stats.mean == 42.0
+        assert stats.std == 0.0
+        assert stats.trend == "flat"
+
+    def test_summarize_upward_trend(self, summarizer: ContextSummarizer) -> None:
+        """Increasing values must produce trend='up'."""
+        values = [1.0, 2.0, 3.0, 10.0, 20.0, 30.0]
+        stats = summarizer.summarize_numeric_values(values)
+        assert stats.trend == "up"
+
+    def test_summarize_downward_trend(self, summarizer: ContextSummarizer) -> None:
+        """Decreasing values must produce trend='down'."""
+        values = [30.0, 20.0, 10.0, 3.0, 2.0, 1.0]
+        stats = summarizer.summarize_numeric_values(values)
+        assert stats.trend == "down"
+
+    def test_summarize_flat_trend(self, summarizer: ContextSummarizer) -> None:
+        """Stable values must produce trend='flat'."""
+        values = [100.0, 100.1, 99.9, 100.0, 100.2, 99.8]
+        stats = summarizer.summarize_numeric_values(values)
+        assert stats.trend == "flat"
+
+    # ------------------------------------------------------------------
+    # summarize_layer
+    # ------------------------------------------------------------------
+
+    def test_summarize_layer_no_data(
+        self, summarizer: ContextSummarizer
+    ) -> None:
+        """summarize_layer with no data must return the 'No data' sentinel."""
+        result = summarizer.summarize_layer(ContextLayer.L6_DAILY)
+        assert result["count"] == 0
+        assert "No data" in result["summary"]
+
+    def test_summarize_layer_numeric(
+        self, summarizer: ContextSummarizer, db_conn: sqlite3.Connection
+    ) -> None:
+        """summarize_layer must collect numeric values and produce stats."""
+        store = summarizer.store
+        store.set_context(ContextLayer.L6_DAILY, "2026-02-01", "total_pnl", 100.0)
+        store.set_context(ContextLayer.L6_DAILY, "2026-02-02", "total_pnl", 200.0)
+
+        result = summarizer.summarize_layer(ContextLayer.L6_DAILY)
+        assert "total_entries" in result
+
+    def test_summarize_layer_with_dict_values(
+        self, summarizer: ContextSummarizer
+    ) -> None:
+        """summarize_layer must handle dict values by extracting numeric subkeys."""
+        store = summarizer.store
+        # set_context serialises the value as JSON, so passing a dict works
+        store.set_context(
+            ContextLayer.L6_DAILY, "2026-02-01", "metrics",
+            {"win_rate": 65.0, "label": "good"}
+        )
+
+        result = summarizer.summarize_layer(ContextLayer.L6_DAILY)
+        assert "total_entries" in result
+        # numeric subkey "win_rate" should appear as "metrics.win_rate"
+        assert "metrics.win_rate" in result
+
+    def test_summarize_layer_with_string_values(
+        self, summarizer: ContextSummarizer
+    ) -> None:
+        """summarize_layer must count string values separately."""
+        store = summarizer.store
+        # set_context stores string values as JSON-encoded strings
+        store.set_context(ContextLayer.L6_DAILY, "2026-02-01", "outlook", "BULLISH")
+
+        result = summarizer.summarize_layer(ContextLayer.L6_DAILY)
+        # String fields contribute a `<key>_count` entry
+        assert "outlook_count" in result
+
+    # ------------------------------------------------------------------
+    # rolling_window_summary
+    # ------------------------------------------------------------------
+
+    def test_rolling_window_summary_basic(
+        self, summarizer: ContextSummarizer
+    ) -> None:
+        """rolling_window_summary must return the expected structure."""
+        store = summarizer.store
+        store.set_context(ContextLayer.L6_DAILY, "2026-02-01", "pnl", 500.0)
+
+        result = summarizer.rolling_window_summary(ContextLayer.L6_DAILY)
+        assert "window_days" in result
+        assert "recent_data" in result
+        assert "historical_summary" in result
+
+    def test_rolling_window_summary_no_older_data(
+        self, summarizer: ContextSummarizer
+    ) -> None:
+        """rolling_window_summary with summarize_older=False skips history."""
+        result = summarizer.rolling_window_summary(
+            ContextLayer.L6_DAILY, summarize_older=False
+        )
+        assert result["historical_summary"] == {}
+
+    # ------------------------------------------------------------------
+    # aggregate_to_higher_layer
+    # ------------------------------------------------------------------
+
+    def test_aggregate_to_higher_layer_mean(
+        self, summarizer: ContextSummarizer
+    ) -> None:
+        """aggregate_to_higher_layer with 'mean' via dict subkeys returns average."""
+        store = summarizer.store
+        # Use different outer keys but same inner metric key so get_all_contexts
+        # returns multiple rows with the target subkey.
+        store.set_context(ContextLayer.L6_DAILY, "2026-02-01", "day1", {"pnl": 100.0})
+        store.set_context(ContextLayer.L6_DAILY, "2026-02-01", "day2", {"pnl": 200.0})
+
+        result = summarizer.aggregate_to_higher_layer(
+            ContextLayer.L6_DAILY, ContextLayer.L5_WEEKLY, "pnl", "mean"
+        )
+        assert result == pytest.approx(150.0)
+
+    def test_aggregate_to_higher_layer_sum(
+        self, summarizer: ContextSummarizer
+    ) -> None:
+        """aggregate_to_higher_layer with 'sum' must return the total."""
+        store = summarizer.store
+        store.set_context(ContextLayer.L6_DAILY, "2026-02-01", "day1", {"pnl": 100.0})
+        store.set_context(ContextLayer.L6_DAILY, "2026-02-01", "day2", {"pnl": 200.0})
+
+        result = summarizer.aggregate_to_higher_layer(
+            ContextLayer.L6_DAILY, ContextLayer.L5_WEEKLY, "pnl", "sum"
+        )
+        assert result == pytest.approx(300.0)
+
+    def test_aggregate_to_higher_layer_max(
+        self, summarizer: ContextSummarizer
+    ) -> None:
+        """aggregate_to_higher_layer with 'max' must return the maximum."""
+        store = summarizer.store
+        store.set_context(ContextLayer.L6_DAILY, "2026-02-01", "day1", {"pnl": 100.0})
+        store.set_context(ContextLayer.L6_DAILY, "2026-02-01", "day2", {"pnl": 200.0})
+
+        result = summarizer.aggregate_to_higher_layer(
+            ContextLayer.L6_DAILY, ContextLayer.L5_WEEKLY, "pnl", "max"
+        )
+        assert result == pytest.approx(200.0)
+
+    def test_aggregate_to_higher_layer_min(
+        self, summarizer: ContextSummarizer
+    ) -> None:
+        """aggregate_to_higher_layer with 'min' must return the minimum."""
+        store = summarizer.store
+        store.set_context(ContextLayer.L6_DAILY, "2026-02-01", "day1", {"pnl": 100.0})
+        store.set_context(ContextLayer.L6_DAILY, "2026-02-01", "day2", {"pnl": 200.0})
+
+        result = summarizer.aggregate_to_higher_layer(
+            ContextLayer.L6_DAILY, ContextLayer.L5_WEEKLY, "pnl", "min"
+        )
+        assert result == pytest.approx(100.0)
+
+    def test_aggregate_to_higher_layer_no_data(
+        self, summarizer: ContextSummarizer
+    ) -> None:
+        """aggregate_to_higher_layer with no matching key must return None."""
+        result = summarizer.aggregate_to_higher_layer(
+            ContextLayer.L6_DAILY, ContextLayer.L5_WEEKLY, "nonexistent", "mean"
+        )
+        assert result is None
+
+    def test_aggregate_to_higher_layer_unknown_func_defaults_to_mean(
+        self, summarizer: ContextSummarizer
+    ) -> None:
+        """Unknown aggregation function must fall back to mean."""
+        store = summarizer.store
+        store.set_context(ContextLayer.L6_DAILY, "2026-02-01", "day1", {"pnl": 100.0})
+        store.set_context(ContextLayer.L6_DAILY, "2026-02-01", "day2", {"pnl": 200.0})
+
+        result = summarizer.aggregate_to_higher_layer(
+            ContextLayer.L6_DAILY, ContextLayer.L5_WEEKLY, "pnl", "unknown_func"
+        )
+        assert result == pytest.approx(150.0)
+
+    # ------------------------------------------------------------------
+    # create_compact_summary + format_summary_for_prompt
+    # ------------------------------------------------------------------
+
+    def test_create_compact_summary(
+        self, summarizer: ContextSummarizer
+    ) -> None:
+        """create_compact_summary must produce a dict keyed by layer value."""
+        store = summarizer.store
+        store.set_context(ContextLayer.L6_DAILY, "2026-02-01", "pnl", 100.0)
+
+        result = summarizer.create_compact_summary([ContextLayer.L6_DAILY])
+        assert ContextLayer.L6_DAILY.value in result
+
+    def test_format_summary_for_prompt_with_numeric_metrics(
+        self, summarizer: ContextSummarizer
+    ) -> None:
+        """format_summary_for_prompt must render avg/trend fields."""
+        store = summarizer.store
+        store.set_context(ContextLayer.L6_DAILY, "2026-02-01", "pnl", 100.0)
+        store.set_context(ContextLayer.L6_DAILY, "2026-02-02", "pnl", 200.0)
+
+        compact = summarizer.create_compact_summary([ContextLayer.L6_DAILY])
+        text = summarizer.format_summary_for_prompt(compact)
+        assert isinstance(text, str)
+
+    def test_format_summary_for_prompt_skips_empty_layers(
+        self, summarizer: ContextSummarizer
+    ) -> None:
+        """format_summary_for_prompt must skip layers with no metrics."""
+        summary = {ContextLayer.L6_DAILY.value: {}}
+        text = summarizer.format_summary_for_prompt(summary)
+        assert text == ""
+
+    def test_format_summary_non_dict_value(
+        self, summarizer: ContextSummarizer
+    ) -> None:
+        """format_summary_for_prompt must render non-dict values as plain text."""
+        summary = {
+            "daily": {
+                "plain_count": 42,
+            }
+        }
+        text = summarizer.format_summary_for_prompt(summary)
+        assert "plain_count" in text
+        assert "42" in text

tests/test_context_scheduler.py

@@ -0,0 +1,104 @@
+"""Tests for ContextScheduler."""
+
+from __future__ import annotations
+
+from dataclasses import dataclass
+from datetime import UTC, datetime
+
+from src.context.scheduler import ContextScheduler
+
+
+@dataclass
+class StubAggregator:
+    """Stub aggregator that records calls."""
+
+    weekly_calls: list[str]
+    monthly_calls: list[str]
+    quarterly_calls: list[str]
+    annual_calls: list[str]
+    legacy_calls: int
+
+    def aggregate_weekly_from_daily(self, week: str) -> None:
+        self.weekly_calls.append(week)
+
+    def aggregate_monthly_from_weekly(self, month: str) -> None:
+        self.monthly_calls.append(month)
+
+    def aggregate_quarterly_from_monthly(self, quarter: str) -> None:
+        self.quarterly_calls.append(quarter)
+
+    def aggregate_annual_from_quarterly(self, year: str) -> None:
+        self.annual_calls.append(year)
+
+    def aggregate_legacy_from_annual(self) -> None:
+        self.legacy_calls += 1
+
+
+@dataclass
+class StubStore:
+    """Stub store that records cleanup calls."""
+
+    cleanup_calls: int = 0
+
+    def cleanup_expired_contexts(self) -> None:
+        self.cleanup_calls += 1
+
+
+def make_scheduler() -> tuple[ContextScheduler, StubAggregator, StubStore]:
+    aggregator = StubAggregator([], [], [], [], 0)
+    store = StubStore()
+    scheduler = ContextScheduler(aggregator=aggregator, store=store)
+    return scheduler, aggregator, store
+
+
+def test_run_if_due_weekly() -> None:
+    scheduler, aggregator, store = make_scheduler()
+    now = datetime(2026, 2, 8, 10, 0, tzinfo=UTC)  # Sunday
+
+    result = scheduler.run_if_due(now)
+
+    assert result.weekly is True
+    assert aggregator.weekly_calls == ["2026-W06"]
+    assert store.cleanup_calls == 1
+
+
+def test_run_if_due_monthly() -> None:
+    scheduler, aggregator, _store = make_scheduler()
+    now = datetime(2026, 2, 28, 12, 0, tzinfo=UTC)  # Last day of month
+
+    result = scheduler.run_if_due(now)
+
+    assert result.monthly is True
+    assert aggregator.monthly_calls == ["2026-02"]
+
+
+def test_run_if_due_quarterly() -> None:
+    scheduler, aggregator, _store = make_scheduler()
+    now = datetime(2026, 3, 31, 12, 0, tzinfo=UTC)  # Last day of Q1
+
+    result = scheduler.run_if_due(now)
+
+    assert result.quarterly is True
+    assert aggregator.quarterly_calls == ["2026-Q1"]
+
+
+def test_run_if_due_annual_and_legacy() -> None:
+    scheduler, aggregator, _store = make_scheduler()
+    now = datetime(2026, 12, 31, 12, 0, tzinfo=UTC)
+
+    result = scheduler.run_if_due(now)
+
+    assert result.annual is True
+    assert result.legacy is True
+    assert aggregator.annual_calls == ["2026"]
+    assert aggregator.legacy_calls == 1
+
+
+def test_cleanup_runs_once_per_day() -> None:
+    scheduler, _aggregator, store = make_scheduler()
+    now = datetime(2026, 2, 9, 9, 0, tzinfo=UTC)
+
+    scheduler.run_if_due(now)
+    scheduler.run_if_due(now)
+
+    assert store.cleanup_calls == 1

tests/test_daily_review.py

@@ -0,0 +1,387 @@
+"""Tests for DailyReviewer."""
+
+from __future__ import annotations
+
+import json
+import sqlite3
+from types import SimpleNamespace
+from unittest.mock import AsyncMock, MagicMock
+
+import pytest
+
+from src.context.layer import ContextLayer
+from src.context.store import ContextStore
+from src.db import init_db, log_trade
+from src.evolution.daily_review import DailyReviewer
+from src.evolution.scorecard import DailyScorecard
+from src.logging.decision_logger import DecisionLogger
+
+from datetime import UTC, datetime
+
+TODAY = datetime.now(UTC).strftime("%Y-%m-%d")
+
+
+@pytest.fixture
+def db_conn() -> sqlite3.Connection:
+    return init_db(":memory:")
+
+
+@pytest.fixture
+def context_store(db_conn: sqlite3.Connection) -> ContextStore:
+    return ContextStore(db_conn)
+
+
+def _log_decision(
+    logger: DecisionLogger,
+    *,
+    stock_code: str,
+    market: str,
+    action: str,
+    confidence: int,
+    scenario_match: dict[str, float] | None = None,
+) -> str:
+    return logger.log_decision(
+        stock_code=stock_code,
+        market=market,
+        exchange_code="KRX" if market == "KR" else "NASDAQ",
+        action=action,
+        confidence=confidence,
+        rationale="test",
+        context_snapshot={"scenario_match": scenario_match or {}},
+        input_data={"stock_code": stock_code},
+    )
+
+
+def test_generate_scorecard_market_scoped(
+    db_conn: sqlite3.Connection, context_store: ContextStore,
+) -> None:
+    reviewer = DailyReviewer(db_conn, context_store)
+    logger = DecisionLogger(db_conn)
+
+    buy_id = _log_decision(
+        logger,
+        stock_code="005930",
+        market="KR",
+        action="BUY",
+        confidence=90,
+        scenario_match={"rsi": 29.0},
+    )
+    _log_decision(
+        logger,
+        stock_code="000660",
+        market="KR",
+        action="HOLD",
+        confidence=60,
+    )
+    _log_decision(
+        logger,
+        stock_code="AAPL",
+        market="US",
+        action="SELL",
+        confidence=80,
+        scenario_match={"volume_ratio": 2.1},
+    )
+
+    log_trade(
+        db_conn,
+        "005930",
+        "BUY",
+        90,
+        "buy",
+        quantity=1,
+        price=100.0,
+        pnl=10.0,
+        market="KR",
+        exchange_code="KRX",
+        decision_id=buy_id,
+    )
+    log_trade(
+        db_conn,
+        "000660",
+        "HOLD",
+        60,
+        "hold",
+        quantity=0,
+        price=0.0,
+        pnl=0.0,
+        market="KR",
+        exchange_code="KRX",
+    )
+    log_trade(
+        db_conn,
+        "AAPL",
+        "SELL",
+        80,
+        "sell",
+        quantity=1,
+        price=200.0,
+        pnl=-5.0,
+        market="US",
+        exchange_code="NASDAQ",
+    )
+
+    scorecard = reviewer.generate_scorecard(TODAY, "KR")
+
+    assert scorecard.market == "KR"
+    assert scorecard.total_decisions == 2
+    assert scorecard.buys == 1
+    assert scorecard.sells == 0
+    assert scorecard.holds == 1
+    assert scorecard.total_pnl == 10.0
+    assert scorecard.win_rate == 100.0
+    assert scorecard.avg_confidence == 75.0
+    assert scorecard.scenario_match_rate == 50.0
+
+
+def test_generate_scorecard_top_winners_and_losers(
+    db_conn: sqlite3.Connection, context_store: ContextStore,
+) -> None:
+    reviewer = DailyReviewer(db_conn, context_store)
+    logger = DecisionLogger(db_conn)
+
+    for code, pnl in [("005930", 30.0), ("000660", 10.0), ("035420", -15.0), ("051910", -5.0)]:
+        decision_id = _log_decision(
+            logger,
+            stock_code=code,
+            market="KR",
+            action="BUY" if pnl >= 0 else "SELL",
+            confidence=80,
+            scenario_match={"rsi": 30.0},
+        )
+        log_trade(
+            db_conn,
+            code,
+            "BUY" if pnl >= 0 else "SELL",
+            80,
+            "test",
+            quantity=1,
+            price=100.0,
+            pnl=pnl,
+            market="KR",
+            exchange_code="KRX",
+            decision_id=decision_id,
+        )
+
+    scorecard = reviewer.generate_scorecard(TODAY, "KR")
+    assert scorecard.top_winners == ["005930", "000660"]
+    assert scorecard.top_losers == ["035420", "051910"]
+
+
+def test_generate_scorecard_empty_day(
+    db_conn: sqlite3.Connection, context_store: ContextStore,
+) -> None:
+    reviewer = DailyReviewer(db_conn, context_store)
+    scorecard = reviewer.generate_scorecard(TODAY, "KR")
+
+    assert scorecard.total_decisions == 0
+    assert scorecard.total_pnl == 0.0
+    assert scorecard.win_rate == 0.0
+    assert scorecard.avg_confidence == 0.0
+    assert scorecard.scenario_match_rate == 0.0
+    assert scorecard.top_winners == []
+    assert scorecard.top_losers == []
+
+
+@pytest.mark.asyncio
+async def test_generate_lessons_without_gemini_returns_empty(
+    db_conn: sqlite3.Connection, context_store: ContextStore,
+) -> None:
+    reviewer = DailyReviewer(db_conn, context_store, gemini_client=None)
+    lessons = await reviewer.generate_lessons(
+        DailyScorecard(
+            date="2026-02-14",
+            market="KR",
+            total_decisions=1,
+            buys=1,
+            sells=0,
+            holds=0,
+            total_pnl=5.0,
+            win_rate=100.0,
+            avg_confidence=90.0,
+            scenario_match_rate=100.0,
+        )
+    )
+    assert lessons == []
+
+
+@pytest.mark.asyncio
+async def test_generate_lessons_parses_json_array(
+    db_conn: sqlite3.Connection, context_store: ContextStore,
+) -> None:
+    mock_gemini = MagicMock()
+    mock_gemini.decide = AsyncMock(
+        return_value=SimpleNamespace(rationale='["Cut losers earlier", "Reduce midday churn"]')
+    )
+    reviewer = DailyReviewer(db_conn, context_store, gemini_client=mock_gemini)
+
+    lessons = await reviewer.generate_lessons(
+        DailyScorecard(
+            date="2026-02-14",
+            market="KR",
+            total_decisions=3,
+            buys=1,
+            sells=1,
+            holds=1,
+            total_pnl=-2.5,
+            win_rate=50.0,
+            avg_confidence=70.0,
+            scenario_match_rate=66.7,
+        )
+    )
+    assert lessons == ["Cut losers earlier", "Reduce midday churn"]
+
+
+@pytest.mark.asyncio
+async def test_generate_lessons_fallback_to_lines(
+    db_conn: sqlite3.Connection, context_store: ContextStore,
+) -> None:
+    mock_gemini = MagicMock()
+    mock_gemini.decide = AsyncMock(
+        return_value=SimpleNamespace(rationale="- Keep risk tighter\n- Increase selectivity")
+    )
+    reviewer = DailyReviewer(db_conn, context_store, gemini_client=mock_gemini)
+
+    lessons = await reviewer.generate_lessons(
+        DailyScorecard(
+            date="2026-02-14",
+            market="US",
+            total_decisions=2,
+            buys=1,
+            sells=1,
+            holds=0,
+            total_pnl=1.0,
+            win_rate=50.0,
+            avg_confidence=75.0,
+            scenario_match_rate=100.0,
+        )
+    )
+    assert lessons == ["Keep risk tighter", "Increase selectivity"]
+
+
+@pytest.mark.asyncio
+async def test_generate_lessons_handles_gemini_error(
+    db_conn: sqlite3.Connection, context_store: ContextStore,
+) -> None:
+    mock_gemini = MagicMock()
+    mock_gemini.decide = AsyncMock(side_effect=RuntimeError("boom"))
+    reviewer = DailyReviewer(db_conn, context_store, gemini_client=mock_gemini)
+
+    lessons = await reviewer.generate_lessons(
+        DailyScorecard(
+            date="2026-02-14",
+            market="US",
+            total_decisions=0,
+            buys=0,
+            sells=0,
+            holds=0,
+            total_pnl=0.0,
+            win_rate=0.0,
+            avg_confidence=0.0,
+            scenario_match_rate=0.0,
+        )
+    )
+    assert lessons == []
+
+
+def test_store_scorecard_in_context(
+    db_conn: sqlite3.Connection, context_store: ContextStore,
+) -> None:
+    reviewer = DailyReviewer(db_conn, context_store)
+    scorecard = DailyScorecard(
+        date="2026-02-14",
+        market="KR",
+        total_decisions=5,
+        buys=2,
+        sells=1,
+        holds=2,
+        total_pnl=15.0,
+        win_rate=66.67,
+        avg_confidence=82.0,
+        scenario_match_rate=80.0,
+        lessons=["Keep position sizing stable"],
+        cross_market_note="US risk-off",
+    )
+
+    reviewer.store_scorecard_in_context(scorecard)
+
+    stored = context_store.get_context(
+        ContextLayer.L6_DAILY,
+        "2026-02-14",
+        "scorecard_KR",
+    )
+    assert stored is not None
+    assert stored["market"] == "KR"
+    assert stored["total_pnl"] == 15.0
+    assert stored["lessons"] == ["Keep position sizing stable"]
+
+
+def test_store_scorecard_key_is_market_scoped(
+    db_conn: sqlite3.Connection, context_store: ContextStore,
+) -> None:
+    reviewer = DailyReviewer(db_conn, context_store)
+    kr = DailyScorecard(
+        date="2026-02-14",
+        market="KR",
+        total_decisions=1,
+        buys=1,
+        sells=0,
+        holds=0,
+        total_pnl=1.0,
+        win_rate=100.0,
+        avg_confidence=90.0,
+        scenario_match_rate=100.0,
+    )
+    us = DailyScorecard(
+        date="2026-02-14",
+        market="US",
+        total_decisions=1,
+        buys=0,
+        sells=1,
+        holds=0,
+        total_pnl=-1.0,
+        win_rate=0.0,
+        avg_confidence=70.0,
+        scenario_match_rate=100.0,
+    )
+
+    reviewer.store_scorecard_in_context(kr)
+    reviewer.store_scorecard_in_context(us)
+
+    kr_ctx = context_store.get_context(ContextLayer.L6_DAILY, "2026-02-14", "scorecard_KR")
+    us_ctx = context_store.get_context(ContextLayer.L6_DAILY, "2026-02-14", "scorecard_US")
+
+    assert kr_ctx["market"] == "KR"
+    assert us_ctx["market"] == "US"
+    assert kr_ctx["total_pnl"] == 1.0
+    assert us_ctx["total_pnl"] == -1.0
+
+
+def test_generate_scorecard_handles_invalid_context_snapshot(
+    db_conn: sqlite3.Connection, context_store: ContextStore,
+) -> None:
+    reviewer = DailyReviewer(db_conn, context_store)
+    db_conn.execute(
+        """
+        INSERT INTO decision_logs (
+            decision_id, timestamp, stock_code, market, exchange_code,
+            action, confidence, rationale, context_snapshot, input_data
+        ) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?)
+        """,
+        (
+            "d1",
+            "2026-02-14T09:00:00+00:00",
+            "005930",
+            "KR",
+            "KRX",
+            "HOLD",
+            50,
+            "test",
+            "{invalid_json",
+            json.dumps({}),
+        ),
+    )
+    db_conn.commit()
+
+    scorecard = reviewer.generate_scorecard("2026-02-14", "KR")
+    assert scorecard.total_decisions == 1
+    assert scorecard.scenario_match_rate == 0.0