feat: 해외주식 지정가 버퍼 최적화 BUY +0.2% / SELL -0.2% (#211)

기존 정책(BUY +0.5%, SELL 현재가)의 두 가지 문제를 해결:
- BUY 0.5% 버퍼는 대형주에서 불필요한 과다 지불 유발 ($50K 규모에서 연간 수십 달러 손실)
- SELL 현재가 지정가는 가격이 소폭 하락 시 미체결 위험 (bid < last_price 구간)

변경:
- BUY: current_price * 1.005 → current_price * 1.002 (+0.2%)
  대형주 기준 90%+ 체결률 유지하면서 과다 지불 최소화
- SELL: current_price → current_price * 0.998 (-0.2%)
  bid가 last_price 아래일 때도 체결 보장
- VTS(paper)와 live 동일 정책 적용 — 더 현실적인 시뮬레이션
- KIS 시장가 주문은 상한가 기준 수량 계산 버그로 사용 안 함(유지)

테스트:
- test_overseas_buy_order_uses_limit_price: 1.005 → 1.002 업데이트
- test_overseas_sell_order_uses_limit_price_below_current: 신규 추가

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

.env.example

@@ -1,36 +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 (requests per second for KIS API)
-# Reduced to 5.0 to avoid "초당 거래건수 초과" errors (EGW00201)
-RATE_LIMIT_RPS=5.0
+# ============================================================
+# Rate Limiting
+# ============================================================
+# 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 — for enhanced decision-making)
+# ============================================================
+# 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

CLAUDE.md

@@ -94,6 +94,7 @@ Smart Scanner runs in `TRADE_MODE=realtime` only. Daily mode uses static watchli
 - **[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
 
@@ -170,7 +171,7 @@ Markets auto-detected based on timezone and enabled in `ENABLED_MARKETS` env var
 - `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 < 80 → force HOLD — cannot be weakened
+- Confidence 임계값 (market_outlook별, 낮출 수 없음): BEARISH ≥ 90, NEUTRAL/기본 ≥ 80, BULLISH ≥ 75
 - All code changes → corresponding tests → coverage ≥ 80%
 
 ## Contributing

docs/commands.md

@@ -192,6 +192,27 @@ When `TELEGRAM_COMMANDS_ENABLED=true` (default), the bot accepts these interacti
 
 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

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/requirements-log.md

@@ -7,6 +7,32 @@
 
 ---
 
+## 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

src/analysis/smart_scanner.py

@@ -175,7 +175,7 @@ class SmartVolatilityScanner:
                 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 * 4.0)))
+                implied_rsi = max(0.0, min(100.0, 50.0 + (change_rate * 2.0)))
 
                 candidates.append(
                     ScanCandidate(
@@ -282,7 +282,7 @@ class SmartVolatilityScanner:
             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 * 4.0)))
+            implied_rsi = max(0.0, min(100.0, 50.0 + (change_rate * 2.0)))
             candidates.append(
                 ScanCandidate(
                     stock_code=stock_code,
@@ -338,7 +338,7 @@ class SmartVolatilityScanner:
 
                 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 * 4.0)))
+                implied_rsi = max(0.0, min(100.0, 50.0 + (change_rate * 2.0)))
                 candidates.append(
                     ScanCandidate(
                         stock_code=stock_code,

src/broker/kis_api.py

@@ -285,7 +285,10 @@ class KISBroker:
         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,
@@ -330,7 +333,13 @@ 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

src/broker/overseas.py

@@ -175,8 +175,12 @@ 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,
@@ -229,8 +233,12 @@ 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,

src/config.py

@@ -13,11 +13,11 @@ 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

src/dashboard/app.py

@@ -3,8 +3,9 @@
 from __future__ import annotations
 
 import json
+import os
 import sqlite3
-from datetime import UTC, datetime
+from datetime import UTC, datetime, timezone
 from pathlib import Path
 from typing import Any
 
@@ -79,6 +80,35 @@ def create_dashboard_app(db_path: str) -> FastAPI:
                 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,
                 "markets": market_status,
@@ -87,6 +117,11 @@ def create_dashboard_app(db_path: str) -> FastAPI:
                     "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}")
@@ -341,12 +376,68 @@ def create_dashboard_app(db_path: str) -> FastAPI:
                 )
             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
 
 

src/dashboard/static/index.html

@@ -13,6 +13,7 @@
         --muted: #9fb3c8;
         --accent: #3cb371;
         --red: #e05555;
+        --warn: #e8a040;
         --border: #28455f;
       }
       * { box-sizing: border-box; margin: 0; padding: 0; }
@@ -43,6 +44,25 @@
       }
       .refresh-btn:hover { border-color: var(--accent); color: var(--accent); }
 
+      /* 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); } }
@@ -123,9 +143,80 @@
       .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>
@@ -134,6 +225,13 @@
       <header>
         <h1>&#x1F40D; The Ouroboros</h1>
         <div class="header-right">
+          <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>
@@ -163,6 +261,30 @@
         </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">
@@ -206,6 +328,72 @@
           </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>
@@ -242,6 +430,71 @@
         </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');
@@ -258,6 +511,7 @@
             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);
         } catch {}
       }
 
@@ -378,13 +632,129 @@
         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 });

src/db.py

@@ -14,6 +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 (
@@ -28,12 +33,13 @@ def init_db(db_path: str) -> sqlite3.Connection:
             pnl REAL DEFAULT 0.0,
             market TEXT DEFAULT 'KR',
             exchange_code TEXT DEFAULT 'KRX',
-            decision_id TEXT
+            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()}
 
@@ -45,6 +51,8 @@ def init_db(db_path: str) -> sqlite3.Connection:
         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(
@@ -131,6 +139,25 @@ def init_db(db_path: str) -> sqlite3.Connection:
     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
 
@@ -148,6 +175,7 @@ def log_trade(
     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.
 
@@ -163,6 +191,8 @@ def log_trade(
         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
@@ -171,9 +201,10 @@ def log_trade(
         """
         INSERT INTO trades (
             timestamp, stock_code, action, confidence, rationale,
-            quantity, price, pnl, market, exchange_code, selection_context, decision_id
+            quantity, price, pnl, market, exchange_code, selection_context, decision_id,
+            mode
         )
-        VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)
+        VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)
         """,
         (
             datetime.now(UTC).isoformat(),
@@ -188,6 +219,7 @@ def log_trade(
             exchange_code,
             context_json,
             decision_id,
+            mode,
         ),
     )
     conn.commit()

src/main.py

@@ -40,9 +40,9 @@ from src.evolution.daily_review import DailyReviewer
 from src.evolution.optimizer import EvolutionOptimizer
 from src.logging.decision_logger import DecisionLogger
 from src.logging_config import setup_logging
-from src.markets.schedule import MarketInfo, get_next_market_open, get_open_markets
+from src.markets.schedule import MARKETS, MarketInfo, get_next_market_open, get_open_markets
 from src.notifications.telegram_client import NotificationFilter, TelegramClient, TelegramCommandHandler
-from src.strategy.models import DayPlaybook
+from src.strategy.models import DayPlaybook, MarketOutlook
 from src.strategy.playbook_store import PlaybookStore
 from src.strategy.pre_market_planner import PreMarketPlanner
 from src.strategy.scenario_engine import ScenarioEngine
@@ -81,12 +81,136 @@ def safe_float(value: str | float | None, default: float = 0.0) -> float:
 TRADE_INTERVAL_SECONDS = 60
 SCAN_INTERVAL_SECONDS = 60  # Scan markets every 60 seconds
 MAX_CONNECTION_RETRIES = 3
+_BUY_COOLDOWN_SECONDS = 600  # 10-minute cooldown after insufficient-balance rejection
 
 # Daily trading mode constants (for Free tier API efficiency)
 DAILY_TRADE_SESSIONS = 4  # Number of trading sessions per day
 TRADE_SESSION_INTERVAL_HOURS = 6  # Hours between sessions
 
 
+async def _retry_connection(coro_factory: Any, *args: Any, label: str = "", **kwargs: Any) -> Any:
+    """Call an async function retrying on ConnectionError with exponential backoff.
+
+    Retries up to MAX_CONNECTION_RETRIES times (exclusive of the first attempt),
+    sleeping 2^attempt seconds between attempts.  Use only for idempotent read
+    operations — never for order submission.
+
+    Args:
+        coro_factory: Async callable (method or function) to invoke.
+        *args: Positional arguments forwarded to coro_factory.
+        label: Human-readable label for log messages.
+        **kwargs: Keyword arguments forwarded to coro_factory.
+
+    Raises:
+        ConnectionError: If all retries are exhausted.
+    """
+    for attempt in range(1, MAX_CONNECTION_RETRIES + 1):
+        try:
+            return await coro_factory(*args, **kwargs)
+        except ConnectionError as exc:
+            if attempt < MAX_CONNECTION_RETRIES:
+                wait_secs = 2 ** attempt
+                logger.warning(
+                    "Connection error %s (attempt %d/%d), retrying in %ds: %s",
+                    label,
+                    attempt,
+                    MAX_CONNECTION_RETRIES,
+                    wait_secs,
+                    exc,
+                )
+                await asyncio.sleep(wait_secs)
+            else:
+                logger.error(
+                    "Connection error %s — all %d retries exhausted: %s",
+                    label,
+                    MAX_CONNECTION_RETRIES,
+                    exc,
+                )
+                raise
+
+
+async def sync_positions_from_broker(
+    broker: Any,
+    overseas_broker: Any,
+    db_conn: Any,
+    settings: "Settings",
+) -> int:
+    """Sync open positions from the live broker into the local DB at startup.
+
+    Fetches current holdings from the broker for all configured markets and
+    inserts a synthetic BUY record for any position that the DB does not
+    already know about.  This prevents double-buy when positions were opened
+    in a previous session or entered manually outside the system.
+
+    Returns:
+        Number of new positions synced.
+    """
+    synced = 0
+    seen_exchange_codes: set[str] = set()
+
+    for market_code in settings.enabled_market_list:
+        market = MARKETS.get(market_code)
+        if market is None:
+            continue
+
+        try:
+            if market.is_domestic:
+                balance_data = await broker.get_balance()
+                log_market = market_code  # "KR"
+            else:
+                if market.exchange_code in seen_exchange_codes:
+                    continue
+                seen_exchange_codes.add(market.exchange_code)
+                balance_data = await overseas_broker.get_overseas_balance(
+                    market.exchange_code
+                )
+                log_market = market_code  # e.g. "US_NASDAQ"
+        except ConnectionError as exc:
+            logger.warning(
+                "Startup sync: balance fetch failed for %s — skipping: %s",
+                market_code,
+                exc,
+            )
+            continue
+
+        held_codes = _extract_held_codes_from_balance(
+            balance_data, is_domestic=market.is_domestic
+        )
+        for stock_code in held_codes:
+            if get_open_position(db_conn, stock_code, log_market):
+                continue  # already tracked
+            qty = _extract_held_qty_from_balance(
+                balance_data, stock_code, is_domestic=market.is_domestic
+            )
+            log_trade(
+                conn=db_conn,
+                stock_code=stock_code,
+                action="BUY",
+                confidence=0,
+                rationale="[startup-sync] Position detected from broker at startup",
+                quantity=qty,
+                price=0.0,
+                market=log_market,
+                exchange_code=market.exchange_code,
+                mode=settings.MODE,
+            )
+            logger.info(
+                "Startup sync: %s/%s recorded as open position (qty=%d)",
+                log_market,
+                stock_code,
+                qty,
+            )
+            synced += 1
+
+    if synced:
+        logger.info(
+            "Startup sync complete: %d position(s) synced from broker", synced
+        )
+    else:
+        logger.info("Startup sync: no new positions to sync from broker")
+    return synced
+
+
 def _extract_symbol_from_holding(item: dict[str, Any]) -> str:
     """Extract symbol from overseas holding payload variants."""
     for key in (
@@ -106,6 +230,82 @@ def _extract_symbol_from_holding(item: dict[str, Any]) -> str:
     return ""
 
 
+def _extract_held_codes_from_balance(
+    balance_data: dict[str, Any],
+    *,
+    is_domestic: bool,
+) -> list[str]:
+    """Return stock codes with a positive orderable quantity from a balance response.
+
+    Uses the broker's live output1 as the source of truth so that partial fills
+    and manual external trades are always reflected correctly.
+    """
+    output1 = balance_data.get("output1", [])
+    if isinstance(output1, dict):
+        output1 = [output1]
+    if not isinstance(output1, list):
+        return []
+
+    codes: list[str] = []
+    for holding in output1:
+        if not isinstance(holding, dict):
+            continue
+        code_key = "pdno" if is_domestic else "ovrs_pdno"
+        code = str(holding.get(code_key, "")).strip().upper()
+        if not code:
+            continue
+        if is_domestic:
+            qty = int(holding.get("ord_psbl_qty") or holding.get("hldg_qty") or 0)
+        else:
+            qty = int(holding.get("ovrs_cblc_qty") or holding.get("hldg_qty") or 0)
+        if qty > 0:
+            codes.append(code)
+    return codes
+
+
+def _extract_held_qty_from_balance(
+    balance_data: dict[str, Any],
+    stock_code: str,
+    *,
+    is_domestic: bool,
+) -> int:
+    """Extract the broker-confirmed orderable quantity for a stock.
+
+    Uses the broker's live balance response (output1) as the source of truth
+    rather than the local DB, because DB records reflect order quantity which
+    may differ from actual fill quantity due to partial fills.
+
+    Domestic fields (VTTC8434R output1):
+        pdno          — 종목코드
+        ord_psbl_qty  — 주문가능수량 (preferred: excludes unsettled)
+        hldg_qty      — 보유수량 (fallback)
+
+    Overseas fields (output1):
+        ovrs_pdno     — 종목코드
+        ovrs_cblc_qty — 해외잔고수량 (preferred)
+        hldg_qty      — 보유수량 (fallback)
+    """
+    output1 = balance_data.get("output1", [])
+    if isinstance(output1, dict):
+        output1 = [output1]
+    if not isinstance(output1, list):
+        return 0
+
+    for holding in output1:
+        if not isinstance(holding, dict):
+            continue
+        code_key = "pdno" if is_domestic else "ovrs_pdno"
+        held_code = str(holding.get(code_key, "")).strip().upper()
+        if held_code != stock_code.strip().upper():
+            continue
+        if is_domestic:
+            qty = int(holding.get("ord_psbl_qty") or holding.get("hldg_qty") or 0)
+        else:
+            qty = int(holding.get("ovrs_cblc_qty") or holding.get("hldg_qty") or 0)
+        return qty
+    return 0
+
+
 def _determine_order_quantity(
     *,
     action: str,
@@ -113,16 +313,40 @@ def _determine_order_quantity(
     total_cash: float,
     candidate: ScanCandidate | None,
     settings: Settings | None,
+    broker_held_qty: int = 0,
+    playbook_allocation_pct: float | None = None,
+    scenario_confidence: int = 80,
 ) -> int:
-    """Determine order quantity using volatility-aware position sizing."""
-    if action != "BUY":
-        return 1
+    """Determine order quantity using volatility-aware position sizing.
+
+    Priority:
+    1. playbook_allocation_pct (AI-specified) scaled by scenario_confidence
+    2. Fallback: volatility-score-based allocation from scanner candidate
+    """
+    if action == "SELL":
+        return broker_held_qty
     if current_price <= 0 or total_cash <= 0:
         return 0
 
     if settings is None or not settings.POSITION_SIZING_ENABLED:
         return 1
 
+    # Use AI-specified allocation_pct if available
+    if playbook_allocation_pct is not None:
+        # Confidence scaling: confidence 80 → 1.0x, confidence 95 → 1.19x
+        confidence_scale = scenario_confidence / 80.0
+        effective_pct = min(
+            settings.POSITION_MAX_ALLOCATION_PCT,
+            max(
+                settings.POSITION_MIN_ALLOCATION_PCT,
+                playbook_allocation_pct * confidence_scale,
+            ),
+        )
+        budget = total_cash * (effective_pct / 100.0)
+        quantity = int(budget // current_price)
+        return max(0, quantity)
+
+    # Fallback: volatility-score-based allocation
     target_score = max(1.0, settings.POSITION_VOLATILITY_TARGET_SCORE)
     observed_score = candidate.score if candidate else target_score
     observed_score = max(1.0, min(100.0, observed_score))
@@ -198,6 +422,7 @@ async def trading_cycle(
     stock_code: str,
     scan_candidates: dict[str, dict[str, ScanCandidate]],
     settings: Settings | None = None,
+    buy_cooldown: dict[str, float] | None = None,
 ) -> None:
     """Execute one trading cycle for a single stock."""
     cycle_start_time = asyncio.get_event_loop().time()
@@ -238,7 +463,13 @@ async def trading_cycle(
         purchase_total = safe_float(balance_info.get("frcr_buy_amt_smtl", "0") or "0")
 
         # Paper mode fallback: VTS overseas balance API often fails for many accounts.
-        if total_cash <= 0 and settings and settings.PAPER_OVERSEAS_CASH > 0:
+        # Only activate in paper mode — live mode must use real balance from KIS.
+        if (
+            total_cash <= 0
+            and settings
+            and settings.MODE == "paper"
+            and settings.PAPER_OVERSEAS_CASH > 0
+        ):
             logger.debug(
                 "Overseas cash balance is 0 for %s; using paper fallback %.2f USD",
                 market.exchange_code,
@@ -328,6 +559,17 @@ async def trading_cycle(
             {"volume_ratio": candidate.volume_ratio},
         )
 
+    # Write pnl_pct to system_metrics (dashboard-only table, separate from AI context tree)
+    db_conn.execute(
+        "INSERT OR REPLACE INTO system_metrics (key, value, updated_at) VALUES (?, ?, ?)",
+        (
+            f"portfolio_pnl_pct_{market.code}",
+            json.dumps({"pnl_pct": round(pnl_pct, 4)}),
+            datetime.now(UTC).isoformat(),
+        ),
+    )
+    db_conn.commit()
+
     # Build portfolio data for global rule evaluation
     portfolio_data = {
         "portfolio_pnl_pct": pnl_pct,
@@ -380,6 +622,61 @@ async def trading_cycle(
     )
     stock_playbook = playbook.get_stock_playbook(stock_code)
 
+    # 2.1. Apply market_outlook-based BUY confidence threshold
+    if decision.action == "BUY":
+        base_threshold = (settings.CONFIDENCE_THRESHOLD if settings else 80)
+        outlook = playbook.market_outlook
+        if outlook == MarketOutlook.BEARISH:
+            min_confidence = 90
+        elif outlook == MarketOutlook.BULLISH:
+            min_confidence = 75
+        else:
+            min_confidence = base_threshold
+        if match.confidence < min_confidence:
+            logger.info(
+                "BUY suppressed for %s (%s): confidence %d < %d (market_outlook=%s)",
+                stock_code,
+                market.name,
+                match.confidence,
+                min_confidence,
+                outlook.value,
+            )
+            decision = TradeDecision(
+                action="HOLD",
+                confidence=match.confidence,
+                rationale=(
+                    f"BUY confidence {match.confidence} < {min_confidence} "
+                    f"(market_outlook={outlook.value})"
+                ),
+            )
+
+    # BUY 결정 전 기존 포지션 체크 (중복 매수 방지)
+    if decision.action == "BUY":
+        existing_position = get_open_position(db_conn, stock_code, market.code)
+        if not existing_position:
+            # SELL 지정가 접수 후 미체결 시 DB는 종료로 기록되나 브로커는 여전히 보유 중.
+            # 국내/해외 모두 라이브 브로커 잔고를 authoritative source로 사용.
+            broker_qty = _extract_held_qty_from_balance(
+                balance_data, stock_code, is_domestic=market.is_domestic
+            )
+            if broker_qty > 0:
+                existing_position = {"price": 0.0, "quantity": broker_qty}
+        if existing_position:
+            decision = TradeDecision(
+                action="HOLD",
+                confidence=decision.confidence,
+                rationale=(
+                    f"Already holding {stock_code} "
+                    f"(entry={existing_position['price']:.4f}, "
+                    f"qty={existing_position['quantity']})"
+                ),
+            )
+            logger.info(
+                "BUY suppressed for %s (%s): already holding open position",
+                stock_code,
+                market.name,
+            )
+
     if decision.action == "HOLD":
         open_position = get_open_position(db_conn, stock_code, market.code)
         if open_position:
@@ -484,12 +781,23 @@ async def trading_cycle(
     trade_price = current_price
     trade_pnl = 0.0
     if decision.action in ("BUY", "SELL"):
+        broker_held_qty = (
+            _extract_held_qty_from_balance(
+                balance_data, stock_code, is_domestic=market.is_domestic
+            )
+            if decision.action == "SELL"
+            else 0
+        )
+        matched_scenario = match.matched_scenario
         quantity = _determine_order_quantity(
             action=decision.action,
             current_price=current_price,
             total_cash=total_cash,
             candidate=candidate,
             settings=settings,
+            broker_held_qty=broker_held_qty,
+            playbook_allocation_pct=matched_scenario.allocation_pct if matched_scenario else None,
+            scenario_confidence=match.confidence,
         )
         if quantity <= 0:
             logger.info(
@@ -503,13 +811,33 @@ async def trading_cycle(
             return
         order_amount = current_price * quantity
 
-        # 4. Risk check BEFORE order
+        # 4. Check BUY cooldown (set when a prior BUY failed due to insufficient balance)
+        if decision.action == "BUY" and buy_cooldown is not None:
+            cooldown_key = f"{market.code}:{stock_code}"
+            cooldown_until = buy_cooldown.get(cooldown_key, 0.0)
+            now = asyncio.get_event_loop().time()
+            if now < cooldown_until:
+                remaining = int(cooldown_until - now)
+                logger.info(
+                    "Skip BUY %s (%s): insufficient-balance cooldown active (%ds remaining)",
+                    stock_code,
+                    market.name,
+                    remaining,
+                )
+                return
+
+        # 5a. Risk check BEFORE order
+        # SELL orders do not consume cash (they receive it), so fat-finger check
+        # is skipped for SELLs — only circuit breaker applies.
         try:
-            risk.validate_order(
-                current_pnl_pct=pnl_pct,
-                order_amount=order_amount,
-                total_cash=total_cash,
-            )
+            if decision.action == "SELL":
+                risk.check_circuit_breaker(pnl_pct)
+            else:
+                risk.validate_order(
+                    current_pnl_pct=pnl_pct,
+                    order_amount=order_amount,
+                    total_cash=total_cash,
+                )
         except FatFingerRejected as exc:
             try:
                 await telegram.notify_fat_finger(
@@ -532,31 +860,45 @@ async def trading_cycle(
                 price=0,  # market order
             )
         else:
-            # For overseas orders:
-            # - KIS VTS only accepts limit orders (지정가만 가능)
-            # - BUY: use 0.5% premium over last price to improve fill probability
-            #   (ask price is typically slightly above last, and VTS won't fill below ask)
-            # - SELL: use last price as the limit
+            # For overseas orders, always use limit orders (지정가):
+            # - KIS market orders (ORD_DVSN=01) calculate quantity based on upper limit
+            #   price (상한가 기준), resulting in only 60-80% of intended cash being used.
+            # - BUY: +0.2% above last price — tight enough to minimise overpayment while
+            #   achieving >90% fill rate on large-cap US stocks.
+            # - SELL: -0.2% below last price — ensures fill even when price dips slightly
+            #   (placing at exact last price risks no-fill if the bid is just below).
             if decision.action == "BUY":
-                order_price = round(current_price * 1.005, 4)
+                order_price = round(current_price * 1.002, 4)
             else:
-                order_price = current_price
+                order_price = round(current_price * 0.998, 4)
             result = await overseas_broker.send_overseas_order(
                 exchange_code=market.exchange_code,
                 stock_code=stock_code,
                 order_type=decision.action,
                 quantity=quantity,
-                price=order_price,  # limit order — KIS VTS rejects market orders
+                price=order_price,  # limit order
             )
             # Check if KIS rejected the order (rt_cd != "0")
             if result.get("rt_cd", "") != "0":
                 order_succeeded = False
+                msg1 = result.get("msg1") or ""
                 logger.warning(
                     "Overseas order not accepted for %s: rt_cd=%s msg=%s",
                     stock_code,
                     result.get("rt_cd"),
-                    result.get("msg1"),
+                    msg1,
                 )
+                # Set BUY cooldown when the rejection is due to insufficient balance
+                if decision.action == "BUY" and buy_cooldown is not None and "주문가능금액" in msg1:
+                    cooldown_key = f"{market.code}:{stock_code}"
+                    buy_cooldown[cooldown_key] = (
+                        asyncio.get_event_loop().time() + _BUY_COOLDOWN_SECONDS
+                    )
+                    logger.info(
+                        "BUY cooldown set for %s: %.0fs (insufficient balance)",
+                        stock_code,
+                        _BUY_COOLDOWN_SECONDS,
+                    )
         logger.info("Order result: %s", result.get("msg1", "OK"))
 
         # 5.5. Notify trade execution (only on success)
@@ -611,6 +953,7 @@ async def trading_cycle(
         exchange_code=market.exchange_code,
         selection_context=selection_context,
         decision_id=decision_id,
+        mode=settings.MODE if settings else "paper",
     )
 
     # 7. Latency monitoring
@@ -649,21 +992,36 @@ async def run_daily_session(
     telegram: TelegramClient,
     settings: Settings,
     smart_scanner: SmartVolatilityScanner | None = None,
-) -> None:
+    daily_start_eval: float = 0.0,
+) -> float:
     """Execute one daily trading session.
 
     V2 proactive strategy: 1 Gemini call for playbook generation,
     then local scenario evaluation per stock (0 API calls).
+
+    Args:
+        daily_start_eval: Portfolio evaluation at the start of the trading day.
+            Used to compute intra-day P&L for the Circuit Breaker.
+            Pass 0.0 on the first session of each day; the function will set
+            it from the first balance query and return it for subsequent
+            sessions.
+
+    Returns:
+        The daily_start_eval value that should be forwarded to the next
+        session of the same trading day.
     """
     # Get currently open markets
     open_markets = get_open_markets(settings.enabled_market_list)
 
     if not open_markets:
         logger.info("No markets open for this session")
-        return
+        return daily_start_eval
 
     logger.info("Starting daily trading session for %d markets", len(open_markets))
 
+    # BUY cooldown: prevents retrying stocks rejected for insufficient balance
+    daily_buy_cooldown: dict[str, float] = {}  # "{market_code}:{stock_code}" -> expiry timestamp
+
     # Process each open market
     for market in open_markets:
         # Use market-local date for playbook keying
@@ -743,11 +1101,18 @@ async def run_daily_session(
             try:
                 if market.is_domestic:
                     current_price, price_change_pct, foreigner_net = (
-                        await broker.get_current_price(stock_code)
+                        await _retry_connection(
+                            broker.get_current_price,
+                            stock_code,
+                            label=stock_code,
+                        )
                     )
                 else:
-                    price_data = await overseas_broker.get_overseas_price(
-                        market.exchange_code, stock_code
+                    price_data = await _retry_connection(
+                        overseas_broker.get_overseas_price,
+                        market.exchange_code,
+                        stock_code,
+                        label=f"{stock_code}@{market.exchange_code}",
                     )
                     current_price = safe_float(
                         price_data.get("output", {}).get("last", "0")
@@ -798,9 +1163,27 @@ async def run_daily_session(
             logger.warning("No valid stock data for market %s", market.code)
             continue
 
-        # Get balance data once for the market
+        # Get balance data once for the market (read-only — safe to retry)
+        try:
+            if market.is_domestic:
+                balance_data = await _retry_connection(
+                    broker.get_balance, label=f"balance:{market.code}"
+                )
+            else:
+                balance_data = await _retry_connection(
+                    overseas_broker.get_overseas_balance,
+                    market.exchange_code,
+                    label=f"overseas_balance:{market.exchange_code}",
+                )
+        except ConnectionError as exc:
+            logger.error(
+                "Balance fetch failed for market %s after all retries — skipping market: %s",
+                market.code,
+                exc,
+            )
+            continue
+
         if market.is_domestic:
-            balance_data = await broker.get_balance()
             output2 = balance_data.get("output2", [{}])
             total_eval = safe_float(
                 output2[0].get("tot_evlu_amt", "0")
@@ -812,7 +1195,6 @@ async def run_daily_session(
                 output2[0].get("pchs_amt_smtl_amt", "0")
             ) if output2 else 0
         else:
-            balance_data = await overseas_broker.get_overseas_balance(market.exchange_code)
             output2 = balance_data.get("output2", [{}])
             if isinstance(output2, list) and output2:
                 balance_info = output2[0]
@@ -827,19 +1209,35 @@ async def run_daily_session(
                 balance_info.get("frcr_buy_amt_smtl", "0") or "0"
             )
             # Paper mode fallback: VTS overseas balance API often fails for many accounts.
-            if total_cash <= 0 and settings.PAPER_OVERSEAS_CASH > 0:
+            # Only activate in paper mode — live mode must use real balance from KIS.
+            if (
+                total_cash <= 0
+                and settings.MODE == "paper"
+                and settings.PAPER_OVERSEAS_CASH > 0
+            ):
                 total_cash = settings.PAPER_OVERSEAS_CASH
 
-            # VTS overseas balance API often returns 0; use paper fallback.
-            if total_cash <= 0 and settings.PAPER_OVERSEAS_CASH > 0:
-                total_cash = settings.PAPER_OVERSEAS_CASH
+        # Capture the day's opening portfolio value on the first market processed
+        # in this session.  Used to compute intra-day P&L for the CB instead of
+        # the cumulative purchase_total which spans the entire account history.
+        if daily_start_eval <= 0 and total_eval > 0:
+            daily_start_eval = total_eval
+            logger.info(
+                "Daily CB baseline set: total_eval=%.2f (first balance of the day)",
+                daily_start_eval,
+            )
 
-        # Calculate daily P&L %
-        pnl_pct = (
-            ((total_eval - purchase_total) / purchase_total * 100)
-            if purchase_total > 0
-            else 0.0
-        )
+        # Daily P&L: compare current eval vs start-of-day eval.
+        # Falls back to purchase_total if daily_start_eval is unavailable (e.g. paper
+        # mode where balance API returns 0 for all values).
+        if daily_start_eval > 0:
+            pnl_pct = (total_eval - daily_start_eval) / daily_start_eval * 100
+        else:
+            pnl_pct = (
+                ((total_eval - purchase_total) / purchase_total * 100)
+                if purchase_total > 0
+                else 0.0
+            )
         portfolio_data = {
             "portfolio_pnl_pct": pnl_pct,
             "total_cash": total_cash,
@@ -870,6 +1268,33 @@ async def run_daily_session(
                 decision.confidence,
             )
 
+            # BUY 중복 방지: 브로커 잔고 기반 (미체결 SELL 리밋 주문 보호)
+            if decision.action == "BUY":
+                daily_existing = get_open_position(db_conn, stock_code, market.code)
+                if not daily_existing:
+                    # SELL 지정가 접수 후 미체결 시 DB는 종료로 기록되나 브로커는 여전히 보유 중.
+                    # 국내/해외 모두 라이브 브로커 잔고를 authoritative source로 사용.
+                    broker_qty = _extract_held_qty_from_balance(
+                        balance_data, stock_code, is_domestic=market.is_domestic
+                    )
+                    if broker_qty > 0:
+                        daily_existing = {"price": 0.0, "quantity": broker_qty}
+                if daily_existing:
+                    decision = TradeDecision(
+                        action="HOLD",
+                        confidence=decision.confidence,
+                        rationale=(
+                            f"Already holding {stock_code} "
+                            f"(entry={daily_existing['price']:.4f}, "
+                            f"qty={daily_existing['quantity']})"
+                        ),
+                    )
+                    logger.info(
+                        "BUY suppressed for %s (%s): already holding open position",
+                        stock_code,
+                        market.name,
+                    )
+
             # Log decision
             context_snapshot = {
                 "L1": {
@@ -909,12 +1334,20 @@ async def run_daily_session(
             trade_pnl = 0.0
             order_succeeded = True
             if decision.action in ("BUY", "SELL"):
+                daily_broker_held_qty = (
+                    _extract_held_qty_from_balance(
+                        balance_data, stock_code, is_domestic=market.is_domestic
+                    )
+                    if decision.action == "SELL"
+                    else 0
+                )
                 quantity = _determine_order_quantity(
                     action=decision.action,
                     current_price=stock_data["current_price"],
                     total_cash=total_cash,
                     candidate=candidate_map.get(stock_code),
                     settings=settings,
+                    broker_held_qty=daily_broker_held_qty,
                 )
                 if quantity <= 0:
                     logger.info(
@@ -928,13 +1361,33 @@ async def run_daily_session(
                     continue
                 order_amount = stock_data["current_price"] * quantity
 
+                # Check BUY cooldown (insufficient balance)
+                if decision.action == "BUY":
+                    daily_cooldown_key = f"{market.code}:{stock_code}"
+                    daily_cooldown_until = daily_buy_cooldown.get(daily_cooldown_key, 0.0)
+                    now = asyncio.get_event_loop().time()
+                    if now < daily_cooldown_until:
+                        remaining = int(daily_cooldown_until - now)
+                        logger.info(
+                            "Skip BUY %s (%s): insufficient-balance cooldown active (%ds remaining)",
+                            stock_code,
+                            market.name,
+                            remaining,
+                        )
+                        continue
+
                 # Risk check
+                # SELL orders do not consume cash (they receive it), so fat-finger
+                # check is skipped for SELLs — only circuit breaker applies.
                 try:
-                    risk.validate_order(
-                        current_pnl_pct=pnl_pct,
-                        order_amount=order_amount,
-                        total_cash=total_cash,
-                    )
+                    if decision.action == "SELL":
+                        risk.check_circuit_breaker(pnl_pct)
+                    else:
+                        risk.validate_order(
+                            current_pnl_pct=pnl_pct,
+                            order_amount=order_amount,
+                            total_cash=total_cash,
+                        )
                 except FatFingerRejected as exc:
                     try:
                         await telegram.notify_fat_finger(
@@ -984,12 +1437,23 @@ async def run_daily_session(
                         )
                         if result.get("rt_cd", "") != "0":
                             order_succeeded = False
+                            daily_msg1 = result.get("msg1") or ""
                             logger.warning(
                                 "Overseas order not accepted for %s: rt_cd=%s msg=%s",
                                 stock_code,
                                 result.get("rt_cd"),
-                                result.get("msg1"),
+                                daily_msg1,
                             )
+                            if decision.action == "BUY" and "주문가능금액" in daily_msg1:
+                                daily_cooldown_key = f"{market.code}:{stock_code}"
+                                daily_buy_cooldown[daily_cooldown_key] = (
+                                    asyncio.get_event_loop().time() + _BUY_COOLDOWN_SECONDS
+                                )
+                                logger.info(
+                                    "BUY cooldown set for %s: %.0fs (insufficient balance)",
+                                    stock_code,
+                                    _BUY_COOLDOWN_SECONDS,
+                                )
                     logger.info("Order result: %s", result.get("msg1", "OK"))
 
                     # Notify trade execution (only on success)
@@ -1038,9 +1502,11 @@ async def run_daily_session(
                 market=market.code,
                 exchange_code=market.exchange_code,
                 decision_id=decision_id,
+                mode=settings.MODE,
             )
 
     logger.info("Daily trading session completed")
+    return daily_start_eval
 
 
 async def _handle_market_close(
@@ -1153,10 +1619,18 @@ def _start_dashboard_server(settings: Settings) -> threading.Thread | None:
     if not settings.DASHBOARD_ENABLED:
         return None
 
+    # Validate dependencies before spawning the thread so startup failures are
+    # reported synchronously (avoids the misleading "started" → "failed" log pair).
+    try:
+        import uvicorn  # noqa: F401
+        from src.dashboard import create_dashboard_app  # noqa: F401
+    except ImportError as exc:
+        logger.warning("Dashboard server unavailable (missing dependency): %s", exc)
+        return None
+
     def _serve() -> None:
         try:
             import uvicorn
-
             from src.dashboard import create_dashboard_app
 
             app = create_dashboard_app(settings.DB_PATH)
@@ -1167,7 +1641,7 @@ def _start_dashboard_server(settings: Settings) -> threading.Thread | None:
                 log_level="info",
             )
         except Exception as exc:
-            logger.warning("Dashboard server failed to start: %s", exc)
+            logger.warning("Dashboard server stopped unexpectedly: %s", exc)
 
     thread = threading.Thread(
         target=_serve,
@@ -1608,6 +2082,9 @@ async def run(settings: Settings) -> None:
     # Active stocks per market (dynamically discovered by scanner)
     active_stocks: dict[str, list[str]] = {}  # market_code -> [stock_codes]
 
+    # BUY cooldown: prevents retrying a stock rejected for insufficient balance
+    buy_cooldown: dict[str, float] = {}  # "{market_code}:{stock_code}" -> expiry timestamp
+
     # Initialize latency control system
     criticality_assessor = CriticalityAssessor(
         critical_pnl_threshold=-2.5,  # Near circuit breaker at -3.0%
@@ -1647,6 +2124,12 @@ async def run(settings: Settings) -> None:
     except Exception as exc:
         logger.warning("System startup notification failed: %s", exc)
 
+    # Sync broker positions → DB to prevent double-buy on restart
+    try:
+        await sync_positions_from_broker(broker, overseas_broker, db_conn, settings)
+    except Exception as exc:
+        logger.warning("Startup position sync failed (non-fatal): %s", exc)
+
     # Start command handler
     try:
         await command_handler.start_polling()
@@ -1665,13 +2148,26 @@ async def run(settings: Settings) -> None:
 
             session_interval = settings.SESSION_INTERVAL_HOURS * 3600  # Convert to seconds
 
+            # daily_start_eval: portfolio eval captured at the first session of each
+            # trading day.  Reset on calendar-date change so the CB measures only
+            # today's drawdown, not cumulative account history.
+            _cb_daily_start_eval: float = 0.0
+            _cb_last_date: str = ""
+
             while not shutdown.is_set():
                 # Wait for trading to be unpaused
                 await pause_trading.wait()
                 _run_context_scheduler(context_scheduler, now=datetime.now(UTC))
 
+                # Reset intra-day CB baseline on a new calendar date
+                today_str = datetime.now(UTC).date().isoformat()
+                if today_str != _cb_last_date:
+                    _cb_last_date = today_str
+                    _cb_daily_start_eval = 0.0
+                    logger.info("New trading day %s — daily CB baseline reset", today_str)
+
                 try:
-                    await run_daily_session(
+                    _cb_daily_start_eval = await run_daily_session(
                         broker,
                         overseas_broker,
                         scenario_engine,
@@ -1685,9 +2181,14 @@ async def run(settings: Settings) -> None:
                         telegram,
                         settings,
                         smart_scanner=smart_scanner,
+                        daily_start_eval=_cb_daily_start_eval,
                     )
                 except CircuitBreakerTripped:
                     logger.critical("Circuit breaker tripped — shutting down")
+                    await telegram.notify_circuit_breaker(
+                        pnl_pct=settings.CIRCUIT_BREAKER_PCT,
+                        threshold=settings.CIRCUIT_BREAKER_PCT,
+                    )
                     shutdown.set()
                     break
                 except Exception as exc:
@@ -1881,8 +2382,38 @@ async def run(settings: Settings) -> None:
                         except Exception as exc:
                             logger.error("Smart Scanner failed for %s: %s", market.name, exc)
 
-                    # Get active stocks from scanner (dynamic, no static fallback)
-                    stock_codes = active_stocks.get(market.code, [])
+                    # Get active stocks from scanner (dynamic, no static fallback).
+                    # Also include currently-held positions so stop-loss /
+                    # take-profit can fire even when a holding drops off the
+                    # scanner.  Broker balance is the source of truth here —
+                    # unlike the local DB it reflects actual fills and any
+                    # manual trades done outside the bot.
+                    scanner_codes = active_stocks.get(market.code, [])
+                    try:
+                        if market.is_domestic:
+                            held_balance = await broker.get_balance()
+                        else:
+                            held_balance = await overseas_broker.get_overseas_balance(
+                                market.exchange_code
+                            )
+                        held_codes = _extract_held_codes_from_balance(
+                            held_balance, is_domestic=market.is_domestic
+                        )
+                    except Exception as exc:
+                        logger.warning(
+                            "Failed to fetch holdings for %s: %s — skipping holdings merge",
+                            market.name, exc,
+                        )
+                        held_codes = []
+
+                    stock_codes = list(dict.fromkeys(scanner_codes + held_codes))
+                    extra_held = [c for c in held_codes if c not in set(scanner_codes)]
+                    if extra_held:
+                        logger.info(
+                            "Holdings added to loop for %s (not in scanner): %s",
+                            market.name, extra_held,
+                        )
+
                     if not stock_codes:
                         logger.debug("No active stocks for market %s", market.code)
                         continue
@@ -1920,6 +2451,7 @@ async def run(settings: Settings) -> None:
                                     stock_code,
                                     scan_candidates,
                                     settings,
+                                    buy_cooldown,
                                 )
                                 break  # Success — exit retry loop
                             except CircuitBreakerTripped as exc:
@@ -1974,6 +2506,8 @@ async def run(settings: Settings) -> None:
                 except TimeoutError:
                     pass  # Normal — timeout means it's time for next cycle
     finally:
+        # Notify shutdown before closing resources
+        await telegram.notify_system_shutdown("Normal shutdown")
         # Clean up resources
         await command_handler.stop_polling()
         await broker.close()

src/notifications/telegram_client.py

@@ -604,9 +604,19 @@ class TelegramCommandHandler:
             async with session.post(url, json=payload) as resp:
                 if resp.status != 200:
                     error_text = await resp.text()
-                    logger.error(
-                        "getUpdates API error (status=%d): %s", resp.status, error_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()

src/strategies/v20260220_210124_evolved.py

@@ -0,0 +1,114 @@
+"""Auto-generated strategy: v20260220_210124
+
+Generated at: 2026-02-20T21:01:24.706847+00:00
+Rationale: Auto-evolved from 6 failures. Primary failure markets: ['US_AMEX', 'US_NYSE', 'US_NASDAQ']. Average loss: -194.69
+"""
+
+from __future__ import annotations
+from typing import Any
+from src.strategies.base import BaseStrategy
+
+
+class Strategy_v20260220_210124(BaseStrategy):
+    """Strategy: v20260220_210124"""
+
+    def evaluate(self, market_data: dict[str, Any]) -> dict[str, Any]:
+        import datetime
+
+        # --- Strategy Constants ---
+        # Minimum price for a stock to be considered for trading (avoids penny stocks)
+        MIN_PRICE = 5.0
+
+        # Momentum signal thresholds (stricter than previous failures)
+        MOMENTUM_PRICE_CHANGE_THRESHOLD = 7.0  # % price change
+        MOMENTUM_VOLUME_RATIO_THRESHOLD = 4.0  # X times average volume
+
+        # Oversold signal thresholds (more conservative)
+        OVERSOLD_RSI_THRESHOLD = 25.0  # RSI value (lower means more oversold)
+
+        # Confidence levels
+        CONFIDENCE_HOLD = 30
+        CONFIDENCE_BUY_OVERSOLD = 65
+        CONFIDENCE_BUY_MOMENTUM = 85
+        CONFIDENCE_BUY_STRONG_MOMENTUM = 90  # For higher-priced stocks with strong momentum
+
+        # Market hours in UTC (9:30 AM ET to 4:00 PM ET)
+        MARKET_OPEN_UTC = datetime.time(14, 30)
+        MARKET_CLOSE_UTC = datetime.time(21, 0)
+
+        # Volatile periods within market hours (UTC) to avoid
+        # First hour after open (14:30 UTC - 15:30 UTC)
+        VOLATILE_OPEN_END_UTC = datetime.time(15, 30)
+        # Last 30 minutes before close (20:30 UTC - 21:00 UTC)
+        VOLATILE_CLOSE_START_UTC = datetime.time(20, 30)
+
+        current_price = market_data.get('current_price')
+        price_change_pct = market_data.get('price_change_pct')
+        volume_ratio = market_data.get('volume_ratio')  # Assumed pre-computed indicator
+        rsi = market_data.get('rsi')                    # Assumed pre-computed indicator
+        timestamp_str = market_data.get('timestamp')
+
+        action = "HOLD"
+        confidence = CONFIDENCE_HOLD
+        rationale = "Initial HOLD: No clear signal or conditions not met."
+
+        # --- 1. Basic Data Validation ---
+        if current_price is None or price_change_pct is None:
+            return {"action": "HOLD", "confidence": CONFIDENCE_HOLD,
+                    "rationale": "Insufficient core data (price or price change) to evaluate."}
+
+        # --- 2. Price Filter: Avoid low-priced/penny stocks ---
+        if current_price < MIN_PRICE:
+            return {"action": "HOLD", "confidence": CONFIDENCE_HOLD,
+                    "rationale": f"Avoiding low-priced stock (${current_price:.2f} < ${MIN_PRICE:.2f})."}
+
+        # --- 3. Time Filter: Only trade during core market hours ---
+        if timestamp_str:
+            try:
+                dt_object = datetime.datetime.fromisoformat(timestamp_str)
+                current_time_utc = dt_object.time()
+
+                if not (MARKET_OPEN_UTC <= current_time_utc < MARKET_CLOSE_UTC):
+                    return {"action": "HOLD", "confidence": CONFIDENCE_HOLD,
+                            "rationale": f"Avoiding trade outside core market hours ({current_time_utc} UTC)."}
+
+                if (MARKET_OPEN_UTC <= current_time_utc < VOLATILE_OPEN_END_UTC) or \
+                   (VOLATILE_CLOSE_START_UTC <= current_time_utc < MARKET_CLOSE_UTC):
+                    return {"action": "HOLD", "confidence": CONFIDENCE_HOLD,
+                            "rationale": f"Avoiding trade during volatile market open/close periods ({current_time_utc} UTC)."}
+
+            except ValueError:
+                rationale += " (Warning: Malformed timestamp, time filters skipped)"
+
+        # --- Initialize signal states ---
+        has_momentum_buy_signal = False
+        has_oversold_buy_signal = False
+
+        # --- 4. Evaluate Enhanced Buy Signals ---
+
+        # Momentum Buy Signal
+        if volume_ratio is not None and \
+           price_change_pct > MOMENTUM_PRICE_CHANGE_THRESHOLD and \
+           volume_ratio > MOMENTUM_VOLUME_RATIO_THRESHOLD:
+            has_momentum_buy_signal = True
+            rationale = f"Momentum BUY: Price change {price_change_pct:.2f}%, Volume {volume_ratio:.2f}x."
+            confidence = CONFIDENCE_BUY_MOMENTUM
+            if current_price >= 10.0:
+                confidence = CONFIDENCE_BUY_STRONG_MOMENTUM
+
+        # Oversold Buy Signal
+        if rsi is not None and rsi < OVERSOLD_RSI_THRESHOLD:
+            has_oversold_buy_signal = True
+            if not has_momentum_buy_signal:
+                rationale = f"Oversold BUY: RSI {rsi:.2f}."
+                confidence = CONFIDENCE_BUY_OVERSOLD
+                if current_price >= 10.0:
+                    confidence = min(CONFIDENCE_BUY_OVERSOLD + 5, 80)
+
+        # --- 5. Decision Logic ---
+        if has_momentum_buy_signal:
+            action = "BUY"
+        elif has_oversold_buy_signal:
+            action = "BUY"
+
+        return {"action": action, "confidence": confidence, "rationale": rationale}

src/strategies/v20260220_210159_evolved.py

@@ -0,0 +1,97 @@
+"""Auto-generated strategy: v20260220_210159
+
+Generated at: 2026-02-20T21:01:59.391523+00:00
+Rationale: Auto-evolved from 6 failures. Primary failure markets: ['US_AMEX', 'US_NYSE', 'US_NASDAQ']. Average loss: -194.69
+"""
+
+from __future__ import annotations
+from typing import Any
+from src.strategies.base import BaseStrategy
+
+
+class Strategy_v20260220_210159(BaseStrategy):
+    """Strategy: v20260220_210159"""
+
+    def evaluate(self, market_data: dict[str, Any]) -> dict[str, Any]:
+        import datetime
+
+        current_price = market_data.get('current_price')
+        price_change_pct = market_data.get('price_change_pct')
+        volume_ratio = market_data.get('volume_ratio')
+        rsi = market_data.get('rsi')
+        timestamp_str = market_data.get('timestamp')
+        market_name = market_data.get('market')
+
+        # Default action
+        action = "HOLD"
+        confidence = 0
+        rationale = "No strong signal or conditions not met."
+
+        # --- FAILURE PATTERN AVOIDANCE ---
+
+        # 1. Avoid low-priced/penny stocks
+        MIN_PRICE_THRESHOLD = 5.0  # USD
+        if current_price is not None and current_price < MIN_PRICE_THRESHOLD:
+            rationale = (
+                f"HOLD: Stock price (${current_price:.2f}) is below minimum threshold "
+                f"(${MIN_PRICE_THRESHOLD:.2f}). Past failures consistently involved low-priced stocks."
+            )
+            return {"action": action, "confidence": confidence, "rationale": rationale}
+
+        # 2. Avoid early market hour volatility
+        if timestamp_str:
+            try:
+                dt_obj = datetime.datetime.fromisoformat(timestamp_str)
+                utc_hour = dt_obj.hour
+                utc_minute = dt_obj.minute
+
+                if (utc_hour == 14 and utc_minute < 45) or (utc_hour == 13 and utc_minute >= 30):
+                    rationale = (
+                        f"HOLD: Trading during early market hours (UTC {utc_hour}:{utc_minute}), "
+                        f"a period identified with past failures due to high volatility."
+                    )
+                    return {"action": action, "confidence": confidence, "rationale": rationale}
+            except ValueError:
+                pass
+
+        # --- IMPROVED BUY STRATEGY ---
+
+        # Momentum BUY signal
+        if volume_ratio is not None and price_change_pct is not None:
+            if price_change_pct > 7.0 and volume_ratio > 3.0:
+                action = "BUY"
+                confidence = 70
+                rationale = "Improved BUY: Momentum signal with high volume and above price threshold."
+
+                if market_name == 'US_AMEX':
+                    confidence = max(55, confidence - 5)
+                    rationale += " (Adjusted lower for AMEX market's higher risk profile)."
+                elif market_name == 'US_NASDAQ' and price_change_pct > 20:
+                    confidence = max(50, confidence - 10)
+                    rationale += " (Adjusted lower for aggressive NASDAQ momentum volatility)."
+
+                if price_change_pct > 15.0:
+                    confidence = max(50, confidence - 5)
+                    rationale += " (Caution: Very high daily price change, potential for reversal)."
+
+                return {"action": action, "confidence": confidence, "rationale": rationale}
+
+        # Oversold BUY signal
+        if rsi is not None and price_change_pct is not None:
+            if rsi < 30 and price_change_pct < -3.0:
+                action = "BUY"
+                confidence = 65
+                rationale = "Improved BUY: Oversold signal with recent decline and above price threshold."
+
+                if market_name == 'US_AMEX':
+                    confidence = max(50, confidence - 5)
+                    rationale += " (Adjusted lower for AMEX market's higher risk on oversold assets)."
+
+                if price_change_pct < -10.0:
+                    confidence = max(45, confidence - 10)
+                    rationale += " (Caution: Very steep decline, potential falling knife)."
+
+                return {"action": action, "confidence": confidence, "rationale": rationale}
+
+        # If no specific BUY signal, default to HOLD
+        return {"action": action, "confidence": confidence, "rationale": rationale}

src/strategies/v20260220_210244_evolved.py

@@ -0,0 +1,88 @@
+"""Auto-generated strategy: v20260220_210244
+
+Generated at: 2026-02-20T21:02:44.387355+00:00
+Rationale: Auto-evolved from 6 failures. Primary failure markets: ['US_AMEX', 'US_NYSE', 'US_NASDAQ']. Average loss: -194.69
+"""
+
+from __future__ import annotations
+from typing import Any
+from src.strategies.base import BaseStrategy
+
+
+class Strategy_v20260220_210244(BaseStrategy):
+    """Strategy: v20260220_210244"""
+
+    def evaluate(self, market_data: dict[str, Any]) -> dict[str, Any]:
+        from datetime import datetime
+
+        # Extract required data points safely
+        current_price = market_data.get("current_price")
+        price_change_pct = market_data.get("price_change_pct")
+        volume_ratio = market_data.get("volume_ratio")
+        rsi = market_data.get("rsi")
+        timestamp_str = market_data.get("timestamp")
+        market_name = market_data.get("market")
+        stock_code = market_data.get("stock_code", "UNKNOWN")
+
+        # Default action is HOLD with conservative confidence and rationale
+        action = "HOLD"
+        confidence = 50
+        rationale = f"No strong BUY signal for {stock_code} or awaiting more favorable conditions after avoiding known failure patterns."
+
+        # --- 1. Failure Pattern Avoidance Filters ---
+
+        # A. Avoid low-priced (penny) stocks
+        if current_price is not None and current_price < 5.0:
+            return {
+                "action": "HOLD",
+                "confidence": 50,
+                "rationale": f"AVOID {stock_code}: Stock price (${current_price:.2f}) is below minimum threshold ($5.00) for BUY action. Identified past failures on highly volatile, low-priced stocks."
+            }
+
+        # B. Avoid initiating BUY trades during identified high-volatility hours
+        if timestamp_str:
+            try:
+                trade_hour = datetime.fromisoformat(timestamp_str).hour
+                if trade_hour in [14, 20]:
+                    return {
+                        "action": "HOLD",
+                        "confidence": 50,
+                        "rationale": f"AVOID {stock_code}: Trading during historically volatile hour ({trade_hour} UTC) where previous BUYs resulted in losses. Prefer to observe market stability."
+                    }
+            except ValueError:
+                pass
+
+        # C. Be cautious with extreme momentum spikes
+        if volume_ratio is not None and price_change_pct is not None:
+            if volume_ratio >= 9.0 and price_change_pct >= 15.0:
+                return {
+                    "action": "HOLD",
+                    "confidence": 50,
+                    "rationale": f"AVOID {stock_code}: Extreme short-term momentum detected (price change: +{price_change_pct:.2f}%, volume ratio: {volume_ratio:.1f}x). Historical failures indicate buying into such rapid spikes often leads to reversals."
+                }
+
+        # D. Be cautious with "oversold" signals without further confirmation
+        if rsi is not None and rsi < 30:
+            return {
+                "action": "HOLD",
+                "confidence": 50,
+                "rationale": f"AVOID {stock_code}: Oversold signal (RSI={rsi:.1f}) detected. While often a BUY signal, historical failures on similar 'oversold' trades suggest waiting for stronger confirmation."
+            }
+
+        # --- 2. Improved BUY Signal Generation ---
+        if volume_ratio is not None and 2.0 <= volume_ratio < 9.0 and \
+           price_change_pct is not None and 2.0 <= price_change_pct < 15.0:
+
+            action = "BUY"
+            confidence = 70
+            rationale = f"BUY {stock_code}: Moderate momentum detected (price change: +{price_change_pct:.2f}%, volume ratio: {volume_ratio:.1f}x). Passed filters for price and extreme momentum, avoiding past failure patterns."
+
+            if market_name in ["US_AMEX", "US_NASDAQ"]:
+                confidence = max(60, confidence - 5)
+                rationale += f" Adjusted confidence for {market_name} market characteristics."
+            elif market_name == "US_NYSE":
+                confidence = max(65, confidence)
+
+            confidence = max(50, min(85, confidence))
+
+        return {"action": action, "confidence": confidence, "rationale": rationale}

src/strategy/models.py

@@ -46,6 +46,18 @@ class StockCondition(BaseModel):
 
     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
@@ -56,6 +68,10 @@ class StockCondition(BaseModel):
     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."""
@@ -70,6 +86,10 @@ class StockCondition(BaseModel):
                 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,
             )
         )
 

src/strategy/pre_market_planner.py

@@ -75,6 +75,7 @@ class PreMarketPlanner:
         market: str,
         candidates: list[ScanCandidate],
         today: date | None = None,
+        current_holdings: list[dict] | None = None,
     ) -> DayPlaybook:
         """Generate a DayPlaybook for a market using Gemini.
 
@@ -82,6 +83,10 @@ class PreMarketPlanner:
             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.
@@ -106,6 +111,7 @@ class PreMarketPlanner:
                 context_data,
                 self_market_scorecard,
                 cross_market,
+                current_holdings=current_holdings,
             )
 
             # 3. Call Gemini
@@ -118,7 +124,8 @@ class PreMarketPlanner:
 
             # 4. Parse response
             playbook = self._parse_response(
-                decision.rationale, today, market, candidates, cross_market
+                decision.rationale, today, market, candidates, cross_market,
+                current_holdings=current_holdings,
             )
             playbook_with_tokens = playbook.model_copy(
                 update={"token_count": decision.token_count}
@@ -230,6 +237,7 @@ class PreMarketPlanner:
         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
@@ -241,6 +249,26 @@ class PreMarketPlanner:
             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 = (
@@ -273,10 +301,20 @@ class PreMarketPlanner:
                     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"
@@ -294,7 +332,8 @@ class PreMarketPlanner:
             f'      "stock_code": "...",\n'
             f'      "scenarios": [\n'
             f'        {{\n'
-            f'          "condition": {{"rsi_below": 30, "volume_ratio_above": 2.0}},\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'
@@ -308,7 +347,8 @@ class PreMarketPlanner:
             f'}}\n\n'
             f"Rules:\n"
             f"- Max {max_scenarios} scenarios per stock\n"
-            f"- Only use stocks from the candidates list\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"
@@ -321,12 +361,19 @@ class PreMarketPlanner:
         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")
@@ -390,6 +437,10 @@ class PreMarketPlanner:
                 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():

src/strategy/scenario_engine.py

@@ -206,6 +206,37 @@ class ScenarioEngine:
         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(
@@ -266,5 +297,9 @@ class ScenarioEngine:
             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

@@ -3,9 +3,11 @@
 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
 
@@ -363,3 +365,435 @@ class TestHealthMonitor:
         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_broker.py

@@ -572,4 +572,156 @@ class TestSendOrderTickRounding:
         order_call = mock_post.call_args_list[1]
         body = order_call[1].get("json", {})
         assert body["ORD_DVSN"] == "01"
-        assert body["ORD_UNPR"] == "0"
+
+
+# ---------------------------------------------------------------------------
+# 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"

tests/test_context.py

@@ -10,6 +10,7 @@ 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
 
 
@@ -370,3 +371,259 @@ class TestLayerMetadata:
 
         # 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_dashboard.py

@@ -316,3 +316,100 @@ def test_pnl_history_market_filter(tmp_path: Path) -> None:
     # KR has 1 trade with pnl=2.0
     assert len(body["labels"]) >= 1
     assert body["pnl"][0] == 2.0
+
+
+def test_positions_returns_open_buy(tmp_path: Path) -> None:
+    """BUY가 마지막 거래인 종목은 포지션으로 반환되어야 한다."""
+    app = _app(tmp_path)
+    get_positions = _endpoint(app, "/api/positions")
+    body = get_positions()
+    # seed_db: 005930은 BUY (오픈), AAPL은 SELL (마지막)
+    assert body["count"] == 1
+    pos = body["positions"][0]
+    assert pos["stock_code"] == "005930"
+    assert pos["market"] == "KR"
+    assert pos["quantity"] == 1
+    assert pos["entry_price"] == 70000
+
+
+def test_positions_excludes_closed_sell(tmp_path: Path) -> None:
+    """마지막 거래가 SELL인 종목은 포지션에 나타나지 않아야 한다."""
+    app = _app(tmp_path)
+    get_positions = _endpoint(app, "/api/positions")
+    body = get_positions()
+    codes = [p["stock_code"] for p in body["positions"]]
+    assert "AAPL" not in codes
+
+
+def test_positions_empty_when_no_trades(tmp_path: Path) -> None:
+    """거래 내역이 없으면 빈 포지션 목록을 반환해야 한다."""
+    db_path = tmp_path / "empty.db"
+    conn = init_db(str(db_path))
+    conn.close()
+    app = create_dashboard_app(str(db_path))
+    get_positions = _endpoint(app, "/api/positions")
+    body = get_positions()
+    assert body["count"] == 0
+    assert body["positions"] == []
+
+
+def _seed_cb_context(conn: sqlite3.Connection, pnl_pct: float, market: str = "KR") -> None:
+    import json as _json
+    conn.execute(
+        "INSERT OR REPLACE INTO system_metrics (key, value, updated_at) VALUES (?, ?, ?)",
+        (
+            f"portfolio_pnl_pct_{market}",
+            _json.dumps({"pnl_pct": pnl_pct}),
+            "2026-02-22T10:00:00+00:00",
+        ),
+    )
+    conn.commit()
+
+
+def test_status_circuit_breaker_ok(tmp_path: Path) -> None:
+    """pnl_pct가 -2.0%보다 높으면 status=ok를 반환해야 한다."""
+    db_path = tmp_path / "cb_ok.db"
+    conn = init_db(str(db_path))
+    _seed_cb_context(conn, -1.0)
+    conn.close()
+    app = create_dashboard_app(str(db_path))
+    get_status = _endpoint(app, "/api/status")
+    body = get_status()
+    cb = body["circuit_breaker"]
+    assert cb["status"] == "ok"
+    assert cb["current_pnl_pct"] == -1.0
+    assert cb["threshold_pct"] == -3.0
+
+
+def test_status_circuit_breaker_warning(tmp_path: Path) -> None:
+    """pnl_pct가 -2.0% 이하이면 status=warning을 반환해야 한다."""
+    db_path = tmp_path / "cb_warn.db"
+    conn = init_db(str(db_path))
+    _seed_cb_context(conn, -2.5)
+    conn.close()
+    app = create_dashboard_app(str(db_path))
+    get_status = _endpoint(app, "/api/status")
+    body = get_status()
+    assert body["circuit_breaker"]["status"] == "warning"
+
+
+def test_status_circuit_breaker_tripped(tmp_path: Path) -> None:
+    """pnl_pct가 임계값(-3.0%) 이하이면 status=tripped를 반환해야 한다."""
+    db_path = tmp_path / "cb_tripped.db"
+    conn = init_db(str(db_path))
+    _seed_cb_context(conn, -3.5)
+    conn.close()
+    app = create_dashboard_app(str(db_path))
+    get_status = _endpoint(app, "/api/status")
+    body = get_status()
+    assert body["circuit_breaker"]["status"] == "tripped"
+
+
+def test_status_circuit_breaker_unknown_when_no_data(tmp_path: Path) -> None:
+    """L7 context에 pnl_pct 데이터가 없으면 status=unknown을 반환해야 한다."""
+    app = _app(tmp_path)  # seed_db에는 portfolio_pnl_pct 없음
+    get_status = _endpoint(app, "/api/status")
+    body = get_status()
+    cb = body["circuit_breaker"]
+    assert cb["status"] == "unknown"
+    assert cb["current_pnl_pct"] is None

tests/test_db.py

@@ -1,5 +1,8 @@
 """Tests for database helper functions."""
 
+import tempfile
+import os
+
 from src.db import get_open_position, init_db, log_trade
 
 
@@ -58,3 +61,135 @@ def test_get_open_position_returns_none_when_latest_is_sell() -> None:
 def test_get_open_position_returns_none_when_no_trades() -> None:
     conn = init_db(":memory:")
     assert get_open_position(conn, "AAPL", "US_NASDAQ") is None
+
+
+# ---------------------------------------------------------------------------
+# WAL mode tests (issue #210)
+# ---------------------------------------------------------------------------
+
+
+def test_wal_mode_applied_to_file_db() -> None:
+    """File-based DB must use WAL journal mode for dashboard concurrent reads."""
+    with tempfile.NamedTemporaryFile(suffix=".db", delete=False) as f:
+        db_path = f.name
+    try:
+        conn = init_db(db_path)
+        cursor = conn.execute("PRAGMA journal_mode")
+        mode = cursor.fetchone()[0]
+        assert mode == "wal", f"Expected WAL mode, got {mode}"
+        conn.close()
+    finally:
+        os.unlink(db_path)
+        # Clean up WAL auxiliary files if they exist
+        for ext in ("-wal", "-shm"):
+            path = db_path + ext
+            if os.path.exists(path):
+                os.unlink(path)
+
+
+def test_wal_mode_not_applied_to_memory_db() -> None:
+    """:memory: DB must not apply WAL (SQLite does not support WAL for in-memory)."""
+    conn = init_db(":memory:")
+    cursor = conn.execute("PRAGMA journal_mode")
+    mode = cursor.fetchone()[0]
+    # In-memory DBs default to 'memory' journal mode
+    assert mode != "wal", "WAL should not be set on in-memory database"
+    conn.close()
+
+
+# ---------------------------------------------------------------------------
+# mode column tests (issue #212)
+# ---------------------------------------------------------------------------
+
+
+def test_log_trade_stores_mode_paper() -> None:
+    """log_trade must persist mode='paper' in the trades table."""
+    conn = init_db(":memory:")
+    log_trade(
+        conn=conn,
+        stock_code="005930",
+        action="BUY",
+        confidence=85,
+        rationale="test",
+        mode="paper",
+    )
+    row = conn.execute("SELECT mode FROM trades ORDER BY id DESC LIMIT 1").fetchone()
+    assert row is not None
+    assert row[0] == "paper"
+
+
+def test_log_trade_stores_mode_live() -> None:
+    """log_trade must persist mode='live' in the trades table."""
+    conn = init_db(":memory:")
+    log_trade(
+        conn=conn,
+        stock_code="005930",
+        action="BUY",
+        confidence=85,
+        rationale="test",
+        mode="live",
+    )
+    row = conn.execute("SELECT mode FROM trades ORDER BY id DESC LIMIT 1").fetchone()
+    assert row is not None
+    assert row[0] == "live"
+
+
+def test_log_trade_default_mode_is_paper() -> None:
+    """log_trade without explicit mode must default to 'paper'."""
+    conn = init_db(":memory:")
+    log_trade(
+        conn=conn,
+        stock_code="005930",
+        action="HOLD",
+        confidence=50,
+        rationale="test",
+    )
+    row = conn.execute("SELECT mode FROM trades ORDER BY id DESC LIMIT 1").fetchone()
+    assert row is not None
+    assert row[0] == "paper"
+
+
+def test_mode_column_exists_in_schema() -> None:
+    """trades table must have a mode column after init_db."""
+    conn = init_db(":memory:")
+    cursor = conn.execute("PRAGMA table_info(trades)")
+    columns = {row[1] for row in cursor.fetchall()}
+    assert "mode" in columns
+
+
+def test_mode_migration_adds_column_to_existing_db() -> None:
+    """init_db must add mode column to existing DBs that lack it (migration)."""
+    import sqlite3
+
+    with tempfile.NamedTemporaryFile(suffix=".db", delete=False) as f:
+        db_path = f.name
+    try:
+        # Create DB without mode column (simulate old schema)
+        old_conn = sqlite3.connect(db_path)
+        old_conn.execute(
+            """CREATE TABLE trades (
+                id INTEGER PRIMARY KEY AUTOINCREMENT,
+                timestamp TEXT NOT NULL,
+                stock_code TEXT NOT NULL,
+                action TEXT NOT NULL,
+                confidence INTEGER NOT NULL,
+                rationale TEXT,
+                quantity INTEGER,
+                price REAL,
+                pnl REAL DEFAULT 0.0,
+                market TEXT DEFAULT 'KR',
+                exchange_code TEXT DEFAULT 'KRX',
+                decision_id TEXT
+            )"""
+        )
+        old_conn.commit()
+        old_conn.close()
+
+        # Run init_db — should add mode column via migration
+        conn = init_db(db_path)
+        cursor = conn.execute("PRAGMA table_info(trades)")
+        columns = {row[1] for row in cursor.fetchall()}
+        assert "mode" in columns
+        conn.close()
+    finally:
+        os.unlink(db_path)

tests/test_logging_config.py

@@ -0,0 +1,117 @@
+"""Tests for JSON structured logging configuration."""
+
+from __future__ import annotations
+
+import json
+import logging
+import sys
+
+from src.logging_config import JSONFormatter, setup_logging
+
+
+class TestJSONFormatter:
+    """Test JSONFormatter output."""
+
+    def test_basic_log_record(self) -> None:
+        """JSONFormatter must emit valid JSON with required fields."""
+        formatter = JSONFormatter()
+        record = logging.LogRecord(
+            name="test.logger",
+            level=logging.INFO,
+            pathname="",
+            lineno=0,
+            msg="Hello %s",
+            args=("world",),
+            exc_info=None,
+        )
+        output = formatter.format(record)
+        data = json.loads(output)
+        assert data["level"] == "INFO"
+        assert data["logger"] == "test.logger"
+        assert data["message"] == "Hello world"
+        assert "timestamp" in data
+
+    def test_includes_exception_info(self) -> None:
+        """JSONFormatter must include exception info when present."""
+        formatter = JSONFormatter()
+        try:
+            raise ValueError("test error")
+        except ValueError:
+            exc_info = sys.exc_info()
+        record = logging.LogRecord(
+            name="test",
+            level=logging.ERROR,
+            pathname="",
+            lineno=0,
+            msg="oops",
+            args=(),
+            exc_info=exc_info,
+        )
+        output = formatter.format(record)
+        data = json.loads(output)
+        assert "exception" in data
+        assert "ValueError" in data["exception"]
+
+    def test_extra_trading_fields_included(self) -> None:
+        """Extra trading fields attached to the record must appear in JSON."""
+        formatter = JSONFormatter()
+        record = logging.LogRecord(
+            name="test",
+            level=logging.INFO,
+            pathname="",
+            lineno=0,
+            msg="trade",
+            args=(),
+            exc_info=None,
+        )
+        record.stock_code = "005930"  # type: ignore[attr-defined]
+        record.action = "BUY"  # type: ignore[attr-defined]
+        record.confidence = 85  # type: ignore[attr-defined]
+        record.pnl_pct = -1.5  # type: ignore[attr-defined]
+        record.order_amount = 1_000_000  # type: ignore[attr-defined]
+        output = formatter.format(record)
+        data = json.loads(output)
+        assert data["stock_code"] == "005930"
+        assert data["action"] == "BUY"
+        assert data["confidence"] == 85
+        assert data["pnl_pct"] == -1.5
+        assert data["order_amount"] == 1_000_000
+
+    def test_none_extra_fields_excluded(self) -> None:
+        """Extra fields that are None must not appear in JSON output."""
+        formatter = JSONFormatter()
+        record = logging.LogRecord(
+            name="test",
+            level=logging.INFO,
+            pathname="",
+            lineno=0,
+            msg="no extras",
+            args=(),
+            exc_info=None,
+        )
+        output = formatter.format(record)
+        data = json.loads(output)
+        assert "stock_code" not in data
+        assert "action" not in data
+        assert "confidence" not in data
+
+
+class TestSetupLogging:
+    """Test setup_logging function."""
+
+    def test_configures_root_logger(self) -> None:
+        """setup_logging must attach a JSON handler to the root logger."""
+        setup_logging(level=logging.DEBUG)
+        root = logging.getLogger()
+        json_handlers = [
+            h for h in root.handlers if isinstance(h.formatter, JSONFormatter)
+        ]
+        assert len(json_handlers) == 1
+        assert root.level == logging.DEBUG
+
+    def test_avoids_duplicate_handlers(self) -> None:
+        """Calling setup_logging twice must not add duplicate handlers."""
+        setup_logging()
+        setup_logging()
+        root = logging.getLogger()
+        assert len(root.handlers) == 1

tests/test_overseas_broker.py

@@ -414,7 +414,7 @@ class TestSendOverseasOrder:
 
     @pytest.mark.asyncio
     async def test_sell_limit_order(self, overseas_broker: OverseasBroker) -> None:
-        """Limit sell order should use VTTT1006U and ORD_DVSN=00."""
+        """Limit sell order should use VTTT1001U and ORD_DVSN=00."""
         mock_resp = AsyncMock()
         mock_resp.status = 200
         mock_resp.json = AsyncMock(return_value={"rt_cd": "0"})
@@ -428,7 +428,7 @@ class TestSendOverseasOrder:
         result = await overseas_broker.send_overseas_order("NYSE", "MSFT", "SELL", 5, price=350.0)
         assert result["rt_cd"] == "0"
 
-        overseas_broker._broker._auth_headers.assert_called_with("VTTT1006U")
+        overseas_broker._broker._auth_headers.assert_called_with("VTTT1001U")
 
         call_args = mock_session.post.call_args
         body = call_args[1]["json"]
@@ -640,4 +640,176 @@ class TestPaperOverseasCash:
             GEMINI_API_KEY="g",
         )
         assert settings.PAPER_OVERSEAS_CASH == 0.0
-        del os.environ["PAPER_OVERSEAS_CASH"]
+
+
+# ---------------------------------------------------------------------------
+# TR_ID live/paper branching — overseas (issues #201, #203)
+# ---------------------------------------------------------------------------
+
+
+def _make_overseas_broker_with_mode(mode: str) -> OverseasBroker:
+    s = Settings(
+        KIS_APP_KEY="k",
+        KIS_APP_SECRET="s",
+        KIS_ACCOUNT_NO="12345678-01",
+        GEMINI_API_KEY="g",
+        DB_PATH=":memory:",
+        MODE=mode,
+    )
+    kis = KISBroker(s)
+    kis._access_token = "tok"
+    kis._token_expires_at = float("inf")
+    kis._rate_limiter.acquire = AsyncMock()
+    return OverseasBroker(kis)
+
+
+class TestOverseasTRIDBranching:
+    """get_overseas_balance and send_overseas_order must use correct TR_ID."""
+
+    @pytest.mark.asyncio
+    async def test_get_overseas_balance_paper_uses_vtts3012r(self) -> None:
+        broker = _make_overseas_broker_with_mode("paper")
+        captured: list[str] = []
+
+        async def mock_auth_headers(tr_id: str) -> dict:
+            captured.append(tr_id)
+            return {"tr_id": tr_id, "authorization": "Bearer tok"}
+
+        broker._broker._auth_headers = mock_auth_headers  # type: ignore[method-assign]
+
+        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)
+
+        mock_session = MagicMock()
+        mock_session.get = MagicMock(return_value=mock_resp)
+        broker._broker._get_session = MagicMock(return_value=mock_session)
+
+        await broker.get_overseas_balance("NASD")
+        assert "VTTS3012R" in captured
+
+    @pytest.mark.asyncio
+    async def test_get_overseas_balance_live_uses_ttts3012r(self) -> None:
+        broker = _make_overseas_broker_with_mode("live")
+        captured: list[str] = []
+
+        async def mock_auth_headers(tr_id: str) -> dict:
+            captured.append(tr_id)
+            return {"tr_id": tr_id, "authorization": "Bearer tok"}
+
+        broker._broker._auth_headers = mock_auth_headers  # type: ignore[method-assign]
+
+        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)
+
+        mock_session = MagicMock()
+        mock_session.get = MagicMock(return_value=mock_resp)
+        broker._broker._get_session = MagicMock(return_value=mock_session)
+
+        await broker.get_overseas_balance("NASD")
+        assert "TTTS3012R" in captured
+
+    @pytest.mark.asyncio
+    async def test_send_overseas_order_buy_paper_uses_vttt1002u(self) -> None:
+        broker = _make_overseas_broker_with_mode("paper")
+        captured: list[str] = []
+
+        async def mock_auth_headers(tr_id: str) -> dict:
+            captured.append(tr_id)
+            return {"tr_id": tr_id, "authorization": "Bearer tok"}
+
+        broker._broker._auth_headers = mock_auth_headers  # type: ignore[method-assign]
+        broker._broker._get_hash_key = AsyncMock(return_value="h")  # type: ignore[method-assign]
+
+        mock_resp = AsyncMock()
+        mock_resp.status = 200
+        mock_resp.json = AsyncMock(return_value={"rt_cd": "0", "msg1": "OK"})
+        mock_resp.__aenter__ = AsyncMock(return_value=mock_resp)
+        mock_resp.__aexit__ = AsyncMock(return_value=False)
+
+        mock_session = MagicMock()
+        mock_session.post = MagicMock(return_value=mock_resp)
+        broker._broker._get_session = MagicMock(return_value=mock_session)
+
+        await broker.send_overseas_order("NASD", "AAPL", "BUY", 1)
+        assert "VTTT1002U" in captured
+
+    @pytest.mark.asyncio
+    async def test_send_overseas_order_buy_live_uses_tttt1002u(self) -> None:
+        broker = _make_overseas_broker_with_mode("live")
+        captured: list[str] = []
+
+        async def mock_auth_headers(tr_id: str) -> dict:
+            captured.append(tr_id)
+            return {"tr_id": tr_id, "authorization": "Bearer tok"}
+
+        broker._broker._auth_headers = mock_auth_headers  # type: ignore[method-assign]
+        broker._broker._get_hash_key = AsyncMock(return_value="h")  # type: ignore[method-assign]
+
+        mock_resp = AsyncMock()
+        mock_resp.status = 200
+        mock_resp.json = AsyncMock(return_value={"rt_cd": "0", "msg1": "OK"})
+        mock_resp.__aenter__ = AsyncMock(return_value=mock_resp)
+        mock_resp.__aexit__ = AsyncMock(return_value=False)
+
+        mock_session = MagicMock()
+        mock_session.post = MagicMock(return_value=mock_resp)
+        broker._broker._get_session = MagicMock(return_value=mock_session)
+
+        await broker.send_overseas_order("NASD", "AAPL", "BUY", 1)
+        assert "TTTT1002U" in captured
+
+    @pytest.mark.asyncio
+    async def test_send_overseas_order_sell_paper_uses_vttt1001u(self) -> None:
+        broker = _make_overseas_broker_with_mode("paper")
+        captured: list[str] = []
+
+        async def mock_auth_headers(tr_id: str) -> dict:
+            captured.append(tr_id)
+            return {"tr_id": tr_id, "authorization": "Bearer tok"}
+
+        broker._broker._auth_headers = mock_auth_headers  # type: ignore[method-assign]
+        broker._broker._get_hash_key = AsyncMock(return_value="h")  # type: ignore[method-assign]
+
+        mock_resp = AsyncMock()
+        mock_resp.status = 200
+        mock_resp.json = AsyncMock(return_value={"rt_cd": "0", "msg1": "OK"})
+        mock_resp.__aenter__ = AsyncMock(return_value=mock_resp)
+        mock_resp.__aexit__ = AsyncMock(return_value=False)
+
+        mock_session = MagicMock()
+        mock_session.post = MagicMock(return_value=mock_resp)
+        broker._broker._get_session = MagicMock(return_value=mock_session)
+
+        await broker.send_overseas_order("NASD", "AAPL", "SELL", 1)
+        assert "VTTT1001U" in captured
+
+    @pytest.mark.asyncio
+    async def test_send_overseas_order_sell_live_uses_tttt1006u(self) -> None:
+        broker = _make_overseas_broker_with_mode("live")
+        captured: list[str] = []
+
+        async def mock_auth_headers(tr_id: str) -> dict:
+            captured.append(tr_id)
+            return {"tr_id": tr_id, "authorization": "Bearer tok"}
+
+        broker._broker._auth_headers = mock_auth_headers  # type: ignore[method-assign]
+        broker._broker._get_hash_key = AsyncMock(return_value="h")  # type: ignore[method-assign]
+
+        mock_resp = AsyncMock()
+        mock_resp.status = 200
+        mock_resp.json = AsyncMock(return_value={"rt_cd": "0", "msg1": "OK"})
+        mock_resp.__aenter__ = AsyncMock(return_value=mock_resp)
+        mock_resp.__aexit__ = AsyncMock(return_value=False)
+
+        mock_session = MagicMock()
+        mock_session.post = MagicMock(return_value=mock_resp)
+        broker._broker._get_session = MagicMock(return_value=mock_session)
+
+        await broker.send_overseas_order("NASD", "AAPL", "SELL", 1)
+        assert "TTTT1006U" in captured

tests/test_pre_market_planner.py

@@ -830,3 +830,171 @@ class TestSmartFallbackPlaybook:
         ]
         assert len(buy_scenarios) == 1
         assert buy_scenarios[0].condition.volume_ratio_above == 2.0  # VOL_MULTIPLIER default
+
+
+# ---------------------------------------------------------------------------
+# Holdings in prompt (#170)
+# ---------------------------------------------------------------------------
+
+
+class TestHoldingsInPrompt:
+    """Tests for current_holdings parameter in generate_playbook / _build_prompt."""
+
+    def _make_holdings(self) -> list[dict]:
+        return [
+            {
+                "stock_code": "005930",
+                "name": "Samsung",
+                "qty": 10,
+                "entry_price": 71000.0,
+                "unrealized_pnl_pct": 2.3,
+                "holding_days": 3,
+            }
+        ]
+
+    def test_build_prompt_includes_holdings_section(self) -> None:
+        """Prompt should contain a Current Holdings section when holdings are given."""
+        planner = _make_planner()
+        candidates = [_candidate()]
+        holdings = self._make_holdings()
+
+        prompt = planner._build_prompt(
+            "KR",
+            candidates,
+            context_data={},
+            self_market_scorecard=None,
+            cross_market=None,
+            current_holdings=holdings,
+        )
+
+        assert "## Current Holdings" in prompt
+        assert "005930" in prompt
+        assert "+2.30%" in prompt
+        assert "보유 3일" in prompt
+
+    def test_build_prompt_no_holdings_omits_section(self) -> None:
+        """Prompt should NOT contain a Current Holdings section when holdings=None."""
+        planner = _make_planner()
+        candidates = [_candidate()]
+
+        prompt = planner._build_prompt(
+            "KR",
+            candidates,
+            context_data={},
+            self_market_scorecard=None,
+            cross_market=None,
+            current_holdings=None,
+        )
+
+        assert "## Current Holdings" not in prompt
+
+    def test_build_prompt_empty_holdings_omits_section(self) -> None:
+        """Empty list should also omit the holdings section."""
+        planner = _make_planner()
+        candidates = [_candidate()]
+
+        prompt = planner._build_prompt(
+            "KR",
+            candidates,
+            context_data={},
+            self_market_scorecard=None,
+            cross_market=None,
+            current_holdings=[],
+        )
+
+        assert "## Current Holdings" not in prompt
+
+    def test_build_prompt_holdings_instruction_included(self) -> None:
+        """Prompt should include instruction to generate scenarios for held stocks."""
+        planner = _make_planner()
+        candidates = [_candidate()]
+        holdings = self._make_holdings()
+
+        prompt = planner._build_prompt(
+            "KR",
+            candidates,
+            context_data={},
+            self_market_scorecard=None,
+            cross_market=None,
+            current_holdings=holdings,
+        )
+
+        assert "005930" in prompt
+        assert "SELL/HOLD" in prompt
+
+    @pytest.mark.asyncio
+    async def test_generate_playbook_passes_holdings_to_prompt(self) -> None:
+        """generate_playbook should pass current_holdings through to the prompt."""
+        planner = _make_planner()
+        candidates = [_candidate()]
+        holdings = self._make_holdings()
+
+        # Capture the actual prompt sent to Gemini
+        captured_prompts: list[str] = []
+        original_decide = planner._gemini.decide
+
+        async def capture_and_call(data: dict) -> TradeDecision:
+            captured_prompts.append(data.get("prompt_override", ""))
+            return await original_decide(data)
+
+        planner._gemini.decide = capture_and_call  # type: ignore[method-assign]
+
+        await planner.generate_playbook(
+            "KR", candidates, today=date(2026, 2, 8), current_holdings=holdings
+        )
+
+        assert len(captured_prompts) == 1
+        assert "## Current Holdings" in captured_prompts[0]
+        assert "005930" in captured_prompts[0]
+
+    @pytest.mark.asyncio
+    async def test_holdings_stock_allowed_in_parse_response(self) -> None:
+        """Holdings stocks not in candidates list should be accepted in the response."""
+        holding_code = "000660"  # Not in candidates
+        stocks = [
+            {
+                "stock_code": "005930",  # candidate
+                "scenarios": [
+                    {
+                        "condition": {"rsi_below": 30},
+                        "action": "BUY",
+                        "confidence": 85,
+                        "rationale": "oversold",
+                    }
+                ],
+            },
+            {
+                "stock_code": holding_code,  # holding only
+                "scenarios": [
+                    {
+                        "condition": {"price_change_pct_below": -2.0},
+                        "action": "SELL",
+                        "confidence": 90,
+                        "rationale": "stop-loss",
+                    }
+                ],
+            },
+        ]
+        planner = _make_planner(gemini_response=_gemini_response_json(stocks=stocks))
+        candidates = [_candidate()]  # only 005930
+        holdings = [
+            {
+                "stock_code": holding_code,
+                "name": "SK Hynix",
+                "qty": 5,
+                "entry_price": 180000.0,
+                "unrealized_pnl_pct": -1.5,
+                "holding_days": 7,
+            }
+        ]
+
+        pb = await planner.generate_playbook(
+            "KR",
+            candidates,
+            today=date(2026, 2, 8),
+            current_holdings=holdings,
+        )
+
+        codes = [sp.stock_code for sp in pb.stock_playbooks]
+        assert "005930" in codes
+        assert holding_code in codes

tests/test_scenario_engine.py

@@ -440,3 +440,135 @@ class TestEvaluate:
         assert result.action == ScenarioAction.BUY
         assert result.match_details["rsi"] == 25.0
         assert isinstance(result.match_details["rsi"], float)
+
+
+# ---------------------------------------------------------------------------
+# Position-aware condition tests (#171)
+# ---------------------------------------------------------------------------
+
+
+class TestPositionAwareConditions:
+    """Tests for unrealized_pnl_pct and holding_days condition fields."""
+
+    def test_evaluate_condition_unrealized_pnl_above_matches(
+        self, engine: ScenarioEngine
+    ) -> None:
+        """unrealized_pnl_pct_above should match when P&L exceeds threshold."""
+        condition = StockCondition(unrealized_pnl_pct_above=3.0)
+        assert engine.evaluate_condition(condition, {"unrealized_pnl_pct": 5.0}) is True
+
+    def test_evaluate_condition_unrealized_pnl_above_no_match(
+        self, engine: ScenarioEngine
+    ) -> None:
+        """unrealized_pnl_pct_above should NOT match when P&L is below threshold."""
+        condition = StockCondition(unrealized_pnl_pct_above=3.0)
+        assert engine.evaluate_condition(condition, {"unrealized_pnl_pct": 2.0}) is False
+
+    def test_evaluate_condition_unrealized_pnl_below_matches(
+        self, engine: ScenarioEngine
+    ) -> None:
+        """unrealized_pnl_pct_below should match when P&L is under threshold."""
+        condition = StockCondition(unrealized_pnl_pct_below=-2.0)
+        assert engine.evaluate_condition(condition, {"unrealized_pnl_pct": -3.5}) is True
+
+    def test_evaluate_condition_unrealized_pnl_below_no_match(
+        self, engine: ScenarioEngine
+    ) -> None:
+        """unrealized_pnl_pct_below should NOT match when P&L is above threshold."""
+        condition = StockCondition(unrealized_pnl_pct_below=-2.0)
+        assert engine.evaluate_condition(condition, {"unrealized_pnl_pct": -1.0}) is False
+
+    def test_evaluate_condition_holding_days_above_matches(
+        self, engine: ScenarioEngine
+    ) -> None:
+        """holding_days_above should match when position held longer than threshold."""
+        condition = StockCondition(holding_days_above=5)
+        assert engine.evaluate_condition(condition, {"holding_days": 7}) is True
+
+    def test_evaluate_condition_holding_days_above_no_match(
+        self, engine: ScenarioEngine
+    ) -> None:
+        """holding_days_above should NOT match when position held shorter."""
+        condition = StockCondition(holding_days_above=5)
+        assert engine.evaluate_condition(condition, {"holding_days": 3}) is False
+
+    def test_evaluate_condition_holding_days_below_matches(
+        self, engine: ScenarioEngine
+    ) -> None:
+        """holding_days_below should match when position held fewer days."""
+        condition = StockCondition(holding_days_below=3)
+        assert engine.evaluate_condition(condition, {"holding_days": 1}) is True
+
+    def test_evaluate_condition_holding_days_below_no_match(
+        self, engine: ScenarioEngine
+    ) -> None:
+        """holding_days_below should NOT match when held more days."""
+        condition = StockCondition(holding_days_below=3)
+        assert engine.evaluate_condition(condition, {"holding_days": 5}) is False
+
+    def test_combined_pnl_and_holding_days(self, engine: ScenarioEngine) -> None:
+        """Combined position-aware conditions should AND-evaluate correctly."""
+        condition = StockCondition(
+            unrealized_pnl_pct_above=3.0,
+            holding_days_above=5,
+        )
+        # Both met → match
+        assert engine.evaluate_condition(
+            condition,
+            {"unrealized_pnl_pct": 4.5, "holding_days": 7},
+        ) is True
+        # Only pnl met → no match
+        assert engine.evaluate_condition(
+            condition,
+            {"unrealized_pnl_pct": 4.5, "holding_days": 3},
+        ) is False
+
+    def test_missing_unrealized_pnl_does_not_match(
+        self, engine: ScenarioEngine
+    ) -> None:
+        """Missing unrealized_pnl_pct key should not match the condition."""
+        condition = StockCondition(unrealized_pnl_pct_above=3.0)
+        assert engine.evaluate_condition(condition, {}) is False
+
+    def test_missing_holding_days_does_not_match(
+        self, engine: ScenarioEngine
+    ) -> None:
+        """Missing holding_days key should not match the condition."""
+        condition = StockCondition(holding_days_above=5)
+        assert engine.evaluate_condition(condition, {}) is False
+
+    def test_match_details_includes_position_fields(
+        self, engine: ScenarioEngine
+    ) -> None:
+        """match_details should include position fields when condition specifies them."""
+        pb = _playbook(
+            scenarios=[
+                StockScenario(
+                    condition=StockCondition(unrealized_pnl_pct_above=3.0),
+                    action=ScenarioAction.SELL,
+                    confidence=90,
+                    rationale="Take profit",
+                )
+            ]
+        )
+        result = engine.evaluate(
+            pb,
+            "005930",
+            {"unrealized_pnl_pct": 5.0},
+            {},
+        )
+        assert result.action == ScenarioAction.SELL
+        assert "unrealized_pnl_pct" in result.match_details
+        assert result.match_details["unrealized_pnl_pct"] == 5.0
+
+    def test_position_conditions_parse_from_planner(self) -> None:
+        """StockCondition should accept and store new fields from JSON parsing."""
+        condition = StockCondition(
+            unrealized_pnl_pct_above=3.0,
+            unrealized_pnl_pct_below=None,
+            holding_days_above=5,
+            holding_days_below=None,
+        )
+        assert condition.unrealized_pnl_pct_above == 3.0
+        assert condition.holding_days_above == 5
+        assert condition.has_any_condition() is True

tests/test_smart_scanner.py

@@ -350,6 +350,42 @@ class TestSmartVolatilityScanner:
         assert [c.stock_code for c in candidates] == ["ABCD"]
 
 
+class TestImpliedRSIFormula:
+    """Test the implied_rsi formula in SmartVolatilityScanner (issue #181)."""
+
+    def test_neutral_change_gives_neutral_rsi(self) -> None:
+        """0% change → implied_rsi = 50 (neutral)."""
+        # formula: 50 + (change_rate * 2.0)
+        rsi = max(0.0, min(100.0, 50.0 + (0.0 * 2.0)))
+        assert rsi == 50.0
+
+    def test_10pct_change_gives_rsi_70(self) -> None:
+        """10% upward change → implied_rsi = 70 (momentum signal)."""
+        rsi = max(0.0, min(100.0, 50.0 + (10.0 * 2.0)))
+        assert rsi == 70.0
+
+    def test_minus_10pct_gives_rsi_30(self) -> None:
+        """-10% change → implied_rsi = 30 (oversold signal)."""
+        rsi = max(0.0, min(100.0, 50.0 + (-10.0 * 2.0)))
+        assert rsi == 30.0
+
+    def test_saturation_at_25pct(self) -> None:
+        """Saturation occurs at >=25% change (not 12.5% as with old coefficient 4.0)."""
+        rsi_12pct = max(0.0, min(100.0, 50.0 + (12.5 * 2.0)))
+        rsi_25pct = max(0.0, min(100.0, 50.0 + (25.0 * 2.0)))
+        rsi_30pct = max(0.0, min(100.0, 50.0 + (30.0 * 2.0)))
+        # At 12.5% change: RSI = 75 (not 100, unlike old formula)
+        assert rsi_12pct == 75.0
+        # At 25%+ saturation
+        assert rsi_25pct == 100.0
+        assert rsi_30pct == 100.0  # Capped
+
+    def test_negative_saturation(self) -> None:
+        """Saturation at -25% gives RSI = 0."""
+        rsi = max(0.0, min(100.0, 50.0 + (-25.0 * 2.0)))
+        assert rsi == 0.0
+
+
 class TestRSICalculation:
     """Test RSI calculation in VolatilityAnalyzer."""
 

tests/test_strategies_base.py

@@ -0,0 +1,32 @@
+"""Tests for BaseStrategy abstract class."""
+
+from __future__ import annotations
+
+from typing import Any
+
+import pytest
+
+from src.strategies.base import BaseStrategy
+
+
+class ConcreteStrategy(BaseStrategy):
+    """Minimal concrete strategy for testing."""
+
+    def evaluate(self, market_data: dict[str, Any]) -> dict[str, Any]:
+        return {"action": "HOLD", "confidence": 50, "rationale": "test"}
+
+
+def test_base_strategy_cannot_be_instantiated() -> None:
+    """BaseStrategy cannot be instantiated directly (it's abstract)."""
+    with pytest.raises(TypeError):
+        BaseStrategy()  # type: ignore[abstract]
+
+
+def test_concrete_strategy_evaluate_returns_decision() -> None:
+    """Concrete subclass must implement evaluate and return a dict."""
+    strategy = ConcreteStrategy()
+    result = strategy.evaluate({"close": [100.0, 101.0]})
+    assert isinstance(result, dict)
+    assert result["action"] == "HOLD"
+    assert result["confidence"] == 50
+    assert "rationale" in result

tests/test_telegram_commands.py

@@ -876,6 +876,54 @@ class TestGetUpdates:
 
             assert updates == []
 
+    @pytest.mark.asyncio
+    async def test_get_updates_409_stops_polling(self) -> None:
+        """409 Conflict response stops the poller (_running = False) and returns empty list."""
+        client = TelegramClient(bot_token="123:abc", chat_id="456", enabled=True)
+        handler = TelegramCommandHandler(client)
+        handler._running = True  # simulate active poller
+
+        mock_resp = AsyncMock()
+        mock_resp.status = 409
+        mock_resp.text = AsyncMock(
+            return_value='{"ok":false,"error_code":409,"description":"Conflict"}'
+        )
+        mock_resp.__aenter__ = AsyncMock(return_value=mock_resp)
+        mock_resp.__aexit__ = AsyncMock(return_value=False)
+
+        with patch("aiohttp.ClientSession.post", return_value=mock_resp):
+            updates = await handler._get_updates()
+
+        assert updates == []
+        assert handler._running is False  # poller stopped
+
+    @pytest.mark.asyncio
+    async def test_poll_loop_exits_after_409(self) -> None:
+        """_poll_loop exits naturally after _running is set to False by a 409 response."""
+        import asyncio as _asyncio
+
+        client = TelegramClient(bot_token="123:abc", chat_id="456", enabled=True)
+        handler = TelegramCommandHandler(client)
+
+        call_count = 0
+
+        async def mock_get_updates_409() -> list[dict]:
+            nonlocal call_count
+            call_count += 1
+            # Simulate 409 stopping the poller
+            handler._running = False
+            return []
+
+        handler._get_updates = mock_get_updates_409  # type: ignore[method-assign]
+
+        handler._running = True
+        task = _asyncio.create_task(handler._poll_loop())
+        await _asyncio.wait_for(task, timeout=2.0)
+
+        # _get_updates called exactly once, then loop exited
+        assert call_count == 1
+        assert handler._running is False
+
 
 class TestCommandWithArgs:
     """Test register_command_with_args and argument dispatch."""