新增市場情報 AI summary output receipt gate

TODO_NEXT_STEPS.txt

@@ -163,7 +163,7 @@
    - Phase 79 candidate queue review archive summary：新增 `services/market_intel/candidate_queue_review_archive_summary.py`、POST `/api/market_intel/manual_sample_review/candidate_queue_review_archive_summary` 與 UI summary 按鈕，在 review completion archive 後整理可供摘要/報表審核的結構化輸入；API/UI 不呼叫 LLM、不派送 Telegram、不寫檔、不讀 token、不執行 CLI、不更新 review_state、不寫 DB、不 commit、不掛 scheduler；版本同步至 V10.276。
    - Phase 80 candidate queue review AI summary preflight：新增 `services/market_intel/candidate_queue_review_ai_summary_preflight.py`、POST `/api/market_intel/manual_sample_review/candidate_queue_review_ai_summary_preflight` 與 UI preflight 按鈕，在 archive summary 後檢查 Ollama-first 三主機級聯與 Gemini 備援邊界；API/UI 不呼叫 LLM、不派送 Telegram、不寫檔、不讀 token、不執行 CLI、不更新 review_state、不寫 DB、不 commit、不掛 scheduler；版本同步至 V10.278。
    - Phase 81 candidate queue review AI summary run package：新增 `services/market_intel/candidate_queue_review_ai_summary_run_package.py`、POST `/api/market_intel/manual_sample_review/candidate_queue_review_ai_summary_run_package` 與 UI package 按鈕，在 AI summary preflight 後整理手動 Ollama 摘要任務包、prompt contract 與輸出 schema；API/UI 不呼叫 LLM、不派送 Telegram、不寫 run package、不讀 token、不執行 CLI、不更新 review_state、不寫 DB、不 commit、不掛 scheduler；版本同步至 V10.280。
-   - Phase 82 candidate queue review AI summary output receipt：新增 `services/market_intel/candidate_queue_review_ai_summary_output_receipt.py`、POST `/api/market_intel/manual_sample_review/candidate_queue_review_ai_summary_output_receipt` 與 UI receipt 按鈕，在 run package 後驗收人工 Ollama 摘要輸出的 schema、evidence_refs 與 model_route；API/UI 不呼叫 LLM、不派送 Telegram、不寫 receipt、不讀 token、不執行 CLI、不更新 review_state、不寫 DB、不 commit、不掛 scheduler；版本同步至 V10.282。
+   - Phase 82 candidate queue review AI summary output receipt：新增 `services/market_intel/candidate_queue_review_ai_summary_output_receipt.py`、POST `/api/market_intel/manual_sample_review/candidate_queue_review_ai_summary_output_receipt` 與 UI receipt 按鈕，在 run package 後驗收人工 Ollama 摘要輸出的 schema、evidence_refs 與 model_route；API/UI 不呼叫 LLM、不派送 Telegram、不寫 receipt、不讀 token、不執行 CLI、不更新 review_state、不寫 DB、不 commit、不掛 scheduler；版本同步至 V10.285。
    - V10.248 補市場情報 390px preview panel QA：sample review 工具列改為 textarea + 可換行 action rail，移除舊的硬編 8 欄 grid；`check_responsive_overflow` 新增 `--screenshot-all`，本機 390x844 `/market_intel` 真頁面 QA 通過且 overflow=0。
    - V10.250 補 Code Review Gemini 備援遙測護欄：Ollama 主路徑失敗時 `fallback_to` 明確指向 `code_review_openclaw_gemini`，測試鎖住「Gemini 不得記成 `code_review_openclaw` 主 caller」；AI Calls 觀測台會把 legacy `code_review_openclaw + gemini` 顯示成 Gemini 備援，避免誤判 Gemini-first。
    - Schema smoke：`tests/test_market_intel_skeleton.py` 檢查 `Base.metadata` 內含 ADR-035 八張 `market_*` tables。

config.py

@@ -320,7 +320,7 @@ YOUTUBE_API_KEY = os.getenv('YOUTUBE_API_KEY', '')
 # ==========================================
 # 系統版本與路徑
 # ==========================================
-SYSTEM_VERSION = "V10.284"
+SYSTEM_VERSION = "V10.285"
 LOG_FILE_PATH = os.path.join(BASE_DIR, 'logs/system.log')
 public_url = PUBLIC_URL  # 用於模板顯示
 

docs/adr/ADR-035-cross-platform-market-campaign-intelligence.md

@@ -207,6 +207,7 @@ EwoooC 目前已有 MOMO EDM / 節慶活動資料、`promo_products`、PChome
 - 2026-05-19 追加 candidate queue review archive summary：`services.market_intel.candidate_queue_review_archive_summary` 與 `/api/market_intel/manual_sample_review/candidate_queue_review_archive_summary` 在 review completion archive 後整理可供摘要/報表審核的結構化輸入。此 summary gate 不產生 AI 摘要；API/UI 不呼叫 LLM、不派送 Telegram、不寫檔、不讀 approval token、不執行 CLI、不更新 `review_state`、不寫 DB、不 commit、不掛 scheduler。
 - 2026-05-19 追加 candidate queue review AI summary preflight：`services.market_intel.candidate_queue_review_ai_summary_preflight` 與 `/api/market_intel/manual_sample_review/candidate_queue_review_ai_summary_preflight` 在 archive summary 後檢查未來 AI 摘要前置條件、Ollama-first 三主機級聯與 Gemini 備援邊界。此 preflight 不產生 AI 摘要；API/UI 不呼叫 LLM、不派送 Telegram、不寫檔、不讀 approval token、不執行 CLI、不更新 `review_state`、不寫 DB、不 commit、不掛 scheduler；188 不可作為 Ollama 節點。
 - 2026-05-19 追加 candidate queue review AI summary run package：`services.market_intel.candidate_queue_review_ai_summary_run_package` 與 `/api/market_intel/manual_sample_review/candidate_queue_review_ai_summary_run_package` 在 AI summary preflight 後整理手動 Ollama 摘要任務包、prompt contract、輸出 schema 與 artifact path contract。此 package gate 不產生 AI 摘要；API/UI 不呼叫 LLM、不派送 Telegram、不寫檔、不讀 approval token、不執行 CLI、不更新 `review_state`、不寫 DB、不 commit、不掛 scheduler；Gemini 僅能作為 Ollama cascade 全失敗後的備援。
+- 2026-05-19 追加 candidate queue review AI summary output receipt：`services.market_intel.candidate_queue_review_ai_summary_output_receipt` 與 `/api/market_intel/manual_sample_review/candidate_queue_review_ai_summary_output_receipt` 在 run package 後驗收人工 Ollama 摘要輸出的 schema、`evidence_refs` 與 `model_route`。此 receipt gate 不產生 AI 摘要；API/UI 不呼叫 LLM、不派送 Telegram、不寫檔、不讀 approval token、不執行 CLI、不更新 `review_state`、不寫 DB、不 commit、不掛 scheduler；摘要持久化與 Telegram 派送必須另開後續 gate。
 
 ### Phase 4：Coupang / Shopee Adapter
 

docs/memory/code_modularization_inventory_20260430.md

@@ -14,6 +14,7 @@
 - 2026-05-19 追記：同步治理測試盤點，校正 `routes/admin_observability_routes.py` 行數；此處只更新 inventory，不變更觀測台功能。
 - 2026-05-19 追記：V10.229 之後 `services/ppt_vision_service.py` 進入 800 行治理清單；本次只補 inventory 讓守門測試反映現況，不變更 PPT 視覺 QA 功能。
 - 2026-05-19 追記：同步背景 V10.276 ElephantAlpha 更新後的 `services/elephant_alpha_autonomous_engine.py` 行數；此處只更新 inventory，不變更 AI engine 行為。
+- 2026-05-19 追記：同步背景 V10.281/V10.282 dashboard 與 Code Review pipeline 更新後的行數；此處只更新 inventory，不變更 dashboard 或 code review 行為。
 
 ## 達到或超過 800 行檔案清單
 
@@ -27,7 +28,7 @@
 | 3681 | `routes/admin_observability_routes.py` | P0 觀測台巨型 Blueprint | `services/observability_query_service.py` / `services/observability_action_service.py` / route glue |
 | 1796 | `routes/ai_routes.py` | P1 AI Blueprint | route glue / AI orchestration service / prompt builders |
 | 1721 | `services/nemoton_dispatcher_service.py` | P1 NemoTron service | NIM client / tool-call parser / action dispatcher |
-| 1507 | `routes/dashboard_routes.py` | P1 Dashboard Blueprint | competitor decision overview / dashboard query service；首頁資料整併需抽 service |
+| 1777 | `routes/dashboard_routes.py` | P1 Dashboard Blueprint | competitor decision overview / dashboard query service；首頁資料整併需抽 service |
 | 1485 | `routes/vendor_routes.py` | P1 Vendor Blueprint | route glue / stockout mutation/email；V2 page query、stockout list/batches API query、vendor list/detail query 已抽到 `services/vendor_stockout_query_service.py` |
 | 1390 | `services/telegram_bot_service.py` | P1 Telegram service | command handlers / message formatters / bot client |
 | 1237 | `app.py` | P1 bootstrap | 保持只做 app setup；繼續往 app_factory / extension setup 抽；Phase 42 只做 metadata table name 對齊 |
@@ -41,7 +42,7 @@
 | 867 | `services/token_report_service.py` | P2 token report service | query / aggregation / chart payload / notification formatting |
 | 865 | `routes/daily_sales_routes.py` | P2 Daily Sales Blueprint | route glue / export helpers / daily query and formatting service |
 | 844 | `services/ollama_service.py` | P2 Ollama client | host health / request client / fallback policy / response parsing |
-| 837 | `services/code_review_pipeline_service.py` | P2 Code review pipeline service | scan orchestration / finding normalization / persistence adapter |
+| 1042 | `services/code_review_pipeline_service.py` | P2 Code review pipeline service | scan orchestration / finding normalization / persistence adapter |
 | 832 | `routes/export_routes.py` | P2 Export flow | export command/router glue / file path / download orchestration |
 | 816 | `services/ppt_vision_service.py` | P2 PPT vision QA service | runtime state / queue status / model probe / audit execution 分離 |
 | 904 | `services/competitor_price_feeder.py` | P2 competitor price feeder | crawler scheduling / price normalization / cache strategy |

routes/README.md

@@ -19,9 +19,9 @@
 | `edm_routes.py` | EDM 與節慶儀表板 | `/edm`, `/festival` |
 | `monthly_routes.py` | 月結分析 | `/monthly_summary_analysis`, `/api/monthly_summary_data` |
 | `daily_sales_routes.py` | 當日業績 | `/daily_sales`, `/daily_sales/export*` |
-| `market_intel_routes.py` | 市場情報 Phase 81 candidate queue review AI summary run package 主路由 | `/market_intel`, `/market_intel/*`, `/api/market_intel/status`, `/api/market_intel/schema`, `/api/market_intel/schema_smoke`, `/api/market_intel/schema_db_probe`, `/api/market_intel/platform_seed_db_diff`, `/api/market_intel/legacy_source_bridge`, `/api/market_intel/mcp_readiness`, `/api/market_intel/mcp_tool_contract`, `/api/market_intel/mcp_deploy_preflight`, `/api/market_intel/mcp_activation_runbook`, `/api/market_intel/mcp_fetch_gate`, `/api/market_intel/scheduler_plan`, `/api/market_intel/manual_sample_plan`, `/api/market_intel/manual_sample_acceptance`, `/api/market_intel/manual_sample_review`, `/api/market_intel/manual_sample_review/evaluate`, `/api/market_intel/manual_sample_review/candidate_handoff`, `/api/market_intel/manual_sample_review/candidate_queue_draft`, `/api/market_intel/manual_sample_review/candidate_queue_approval`, `/api/market_intel/manual_sample_review/candidate_queue_transaction`, `/api/market_intel/manual_sample_review/candidate_queue_writer_status`, `/api/market_intel/manual_sample_review/candidate_queue_writer_preflight`, `/api/market_intel/manual_sample_review/candidate_queue_writer_postwrite_smoke`, `/api/market_intel/manual_sample_review/candidate_queue_writer_operator_drill`, `/api/market_intel/manual_sample_review/candidate_queue_writer_run_package`, `/api/market_intel/manual_sample_review/candidate_queue_writer_run_readiness`, `/api/market_intel/manual_sample_review/candidate_queue_writer_run_receipt`, `/api/market_intel/manual_sample_review/candidate_queue_writer_run_closeout`, `/api/market_intel/manual_sample_review/candidate_queue_review_handoff`, `/api/market_intel/match_review_plan`, `/api/market_intel/opportunity_plan`, `/api/market_intel/opportunity_scoring_plan`, `/api/market_intel/opportunity_evidence_plan`, `/api/market_intel/opportunity_alert_plan`, `/api/market_intel/adapters`, `/api/market_intel/dry_run_plan`, `/api/market_intel/discovery_plan`, `/api/market_intel/manual_discovery`, `/api/market_intel/candidate_preview`, `/api/market_intel/platform_seed_plan`, `/api/market_intel/platform_seed_write_guard`, `/api/market_intel/platform_seed_writer_plan`, `/api/market_intel/migration_blueprint`, `/api/market_intel/migration_apply_drill`, `/api/market_intel/migration_catalog_review`, `/api/market_intel/migration_live_smoke`, `/api/market_intel/live_db_inventory`, `/api/market_intel/seed_writer_cli_status`, `/api/market_intel/write_approval_runbook`, `/api/market_intel/deployment_readiness` |
+| `market_intel_routes.py` | 市場情報 Phase 82 candidate queue review AI summary output receipt 主路由 | `/market_intel`, `/market_intel/*`, `/api/market_intel/status`, `/api/market_intel/schema`, `/api/market_intel/schema_smoke`, `/api/market_intel/schema_db_probe`, `/api/market_intel/platform_seed_db_diff`, `/api/market_intel/legacy_source_bridge`, `/api/market_intel/mcp_readiness`, `/api/market_intel/mcp_tool_contract`, `/api/market_intel/mcp_deploy_preflight`, `/api/market_intel/mcp_activation_runbook`, `/api/market_intel/mcp_fetch_gate`, `/api/market_intel/scheduler_plan`, `/api/market_intel/manual_sample_plan`, `/api/market_intel/manual_sample_acceptance`, `/api/market_intel/manual_sample_review`, `/api/market_intel/manual_sample_review/evaluate`, `/api/market_intel/manual_sample_review/candidate_handoff`, `/api/market_intel/manual_sample_review/candidate_queue_draft`, `/api/market_intel/manual_sample_review/candidate_queue_approval`, `/api/market_intel/manual_sample_review/candidate_queue_transaction`, `/api/market_intel/manual_sample_review/candidate_queue_writer_status`, `/api/market_intel/manual_sample_review/candidate_queue_writer_preflight`, `/api/market_intel/manual_sample_review/candidate_queue_writer_postwrite_smoke`, `/api/market_intel/manual_sample_review/candidate_queue_writer_operator_drill`, `/api/market_intel/manual_sample_review/candidate_queue_writer_run_package`, `/api/market_intel/manual_sample_review/candidate_queue_writer_run_readiness`, `/api/market_intel/manual_sample_review/candidate_queue_writer_run_receipt`, `/api/market_intel/manual_sample_review/candidate_queue_writer_run_closeout`, `/api/market_intel/manual_sample_review/candidate_queue_review_handoff`, `/api/market_intel/match_review_plan`, `/api/market_intel/opportunity_plan`, `/api/market_intel/opportunity_scoring_plan`, `/api/market_intel/opportunity_evidence_plan`, `/api/market_intel/opportunity_alert_plan`, `/api/market_intel/adapters`, `/api/market_intel/dry_run_plan`, `/api/market_intel/discovery_plan`, `/api/market_intel/manual_discovery`, `/api/market_intel/candidate_preview`, `/api/market_intel/platform_seed_plan`, `/api/market_intel/platform_seed_write_guard`, `/api/market_intel/platform_seed_writer_plan`, `/api/market_intel/migration_blueprint`, `/api/market_intel/migration_apply_drill`, `/api/market_intel/migration_catalog_review`, `/api/market_intel/migration_live_smoke`, `/api/market_intel/live_db_inventory`, `/api/market_intel/seed_writer_cli_status`, `/api/market_intel/write_approval_runbook`, `/api/market_intel/deployment_readiness` |
 | `market_intel_review_routes.py` | 市場情報人工 queue review 只讀延伸 API | `/api/market_intel/manual_sample_review/candidate_queue_review_inventory`, `/api/market_intel/manual_sample_review/candidate_queue_review_decision`, `/api/market_intel/manual_sample_review/candidate_queue_review_decision_approval`, `/api/market_intel/manual_sample_review/candidate_queue_review_decision_transaction`, `/api/market_intel/manual_sample_review/candidate_queue_review_decision_writer_status`, `/api/market_intel/manual_sample_review/candidate_queue_review_decision_writer_preflight`, `/api/market_intel/manual_sample_review/candidate_queue_review_decision_writer_postwrite_smoke`, `/api/market_intel/manual_sample_review/candidate_queue_review_decision_writer_operator_drill`, `/api/market_intel/manual_sample_review/candidate_queue_review_decision_writer_run_package`, `/api/market_intel/manual_sample_review/candidate_queue_review_decision_writer_run_readiness`, `/api/market_intel/manual_sample_review/candidate_queue_review_decision_writer_run_receipt`, `/api/market_intel/manual_sample_review/candidate_queue_review_decision_writer_run_closeout` |
-| `market_intel_review_post_routes.py` | 市場情報 review_state closeout 後只讀延伸 API（掛在 `market_intel_review_bp`） | `/api/market_intel/manual_sample_review/candidate_queue_review_decision_post_closeout_inventory`, `/api/market_intel/manual_sample_review/candidate_queue_review_completion_archive`, `/api/market_intel/manual_sample_review/candidate_queue_review_archive_summary`, `/api/market_intel/manual_sample_review/candidate_queue_review_ai_summary_preflight`, `/api/market_intel/manual_sample_review/candidate_queue_review_ai_summary_run_package` |
+| `market_intel_review_post_routes.py` | 市場情報 review_state closeout 後只讀延伸 API（掛在 `market_intel_review_bp`） | `/api/market_intel/manual_sample_review/candidate_queue_review_decision_post_closeout_inventory`, `/api/market_intel/manual_sample_review/candidate_queue_review_completion_archive`, `/api/market_intel/manual_sample_review/candidate_queue_review_archive_summary`, `/api/market_intel/manual_sample_review/candidate_queue_review_ai_summary_preflight`, `/api/market_intel/manual_sample_review/candidate_queue_review_ai_summary_run_package`, `/api/market_intel/manual_sample_review/candidate_queue_review_ai_summary_output_receipt` |
 | `api_routes.py` | 通用任務與查詢 API | `/api/run_task`, `/api/history/*` |
 | `export_routes.py` | 匯出功能 | `/api/export/*` |
 | `import_routes.py` | 匯入功能 | `/api/import_excel`, `/api/import/monthly_summary` |

routes/market_intel_review_post_routes.py

@@ -14,6 +14,9 @@ from services.market_intel import MarketIntelService
 from services.market_intel.candidate_queue_review_ai_summary_preflight import (
     build_candidate_queue_review_ai_summary_preflight,
 )
+from services.market_intel.candidate_queue_review_ai_summary_output_receipt import (
+    build_candidate_queue_review_ai_summary_output_receipt,
+)
 from services.market_intel.candidate_queue_review_ai_summary_run_package import (
     build_candidate_queue_review_ai_summary_run_package,
 )
@@ -182,6 +185,53 @@ def _build_review_ai_summary_preflight_stack(
     )
 
 
+def _build_review_ai_summary_run_package_stack(
+    *,
+    service,
+    sample_result,
+    payload_error,
+    operator_evidence,
+    writer_output,
+    smoke_result,
+    limit,
+    execute_requested,
+):
+    (
+        transaction,
+        receipt,
+        closeout,
+        inventory,
+        archive,
+        archive_summary,
+        ai_summary_preflight,
+    ) = _build_review_ai_summary_preflight_stack(
+        service=service,
+        sample_result=sample_result,
+        payload_error=payload_error,
+        operator_evidence=operator_evidence,
+        writer_output=writer_output,
+        smoke_result=smoke_result,
+        limit=limit,
+        execute_requested=execute_requested,
+    )
+    ai_summary_run_package = build_candidate_queue_review_ai_summary_run_package(
+        archive_summary=archive_summary,
+        ai_summary_preflight=ai_summary_preflight,
+        operator_evidence=operator_evidence,
+        execute_requested=execute_requested,
+    )
+    return (
+        transaction,
+        receipt,
+        closeout,
+        inventory,
+        archive,
+        archive_summary,
+        ai_summary_preflight,
+        ai_summary_run_package,
+    )
+
+
 @market_intel_review_bp.route(
     "/api/market_intel/manual_sample_review/"
     "candidate_queue_review_decision_post_closeout_inventory",
@@ -293,7 +343,8 @@ def market_intel_manual_sample_candidate_queue_review_ai_summary_run_package():
         archive,
         archive_summary,
         ai_summary_preflight,
-    ) = _build_review_ai_summary_preflight_stack(
+        ai_summary_run_package,
+    ) = _build_review_ai_summary_run_package_stack(
         service=service,
         sample_result=sample_result,
         payload_error=payload_error,
@@ -303,9 +354,44 @@ def market_intel_manual_sample_candidate_queue_review_ai_summary_run_package():
         limit=limit,
         execute_requested=execute_requested,
     )
-    data = build_candidate_queue_review_ai_summary_run_package(
-        archive_summary=archive_summary,
-        ai_summary_preflight=ai_summary_preflight,
+    data = ai_summary_run_package
+    data["phase"] = service.phase
+    return jsonify(data), 400 if payload_error else 200
+
+
+@market_intel_review_bp.route(
+    "/api/market_intel/manual_sample_review/"
+    "candidate_queue_review_ai_summary_output_receipt",
+    methods=["POST"],
+)
+@login_required
+def market_intel_manual_sample_candidate_queue_review_ai_summary_output_receipt():
+    service = MarketIntelService()
+    execute_requested = request.args.get("execute", "false").lower() == "true"
+    sample_result, operator_evidence, writer_output, smoke_result, payload_error, limit = (
+        _extract_run_payload()
+    )
+    (
+        transaction,
+        receipt,
+        closeout,
+        inventory,
+        archive,
+        archive_summary,
+        ai_summary_preflight,
+        ai_summary_run_package,
+    ) = _build_review_ai_summary_run_package_stack(
+        service=service,
+        sample_result=sample_result,
+        payload_error=payload_error,
+        operator_evidence=operator_evidence,
+        writer_output=writer_output,
+        smoke_result=smoke_result,
+        limit=limit,
+        execute_requested=execute_requested,
+    )
+    data = build_candidate_queue_review_ai_summary_output_receipt(
+        ai_summary_run_package=ai_summary_run_package,
         operator_evidence=operator_evidence,
         execute_requested=execute_requested,
     )

services/market_intel/candidate_queue_review_ai_summary_output_receipt.py

@@ -0,0 +1,451 @@
+"""候選審核 queue AI summary output receipt 預覽。
+
+本模組只驗收人工 Ollama 摘要輸出是否符合 run package schema 與 evidence contract；
+不呼叫 LLM、不派送 Telegram、不讀 approval token、不執行 CLI、不寫檔、
+不寫 DB、不更新 review_state、不 commit、不掛 scheduler。
+"""
+
+
+FORBIDDEN_TOKEN_KEYWORDS = (
+    "approval_token",
+    "approval-token",
+    "market_intel_queue_write_approval",
+)
+SAFE_TOKEN_METADATA_KEYS = {
+    "approval_token_present",
+    "approval_token_valid",
+    "approval_token_secret_configured",
+}
+SAFE_APPROVAL_ENV_VAR = "MARKET_INTEL_QUEUE_WRITE_APPROVAL"
+TARGET_TABLE = "market_alert_review_queue"
+
+
+def _as_dict(value):
+    return value if isinstance(value, dict) else {}
+
+
+def _as_list(value):
+    if value is None:
+        return []
+    if isinstance(value, (list, tuple, set)):
+        return list(value)
+    return [value]
+
+
+def _contains_forbidden_token_key(value):
+    if isinstance(value, dict):
+        for key, nested in value.items():
+            normalized_key = str(key).lower()
+            if normalized_key in SAFE_TOKEN_METADATA_KEYS and isinstance(nested, bool):
+                continue
+            if normalized_key == "approval_env_var" and nested == SAFE_APPROVAL_ENV_VAR:
+                continue
+            if any(token_key in normalized_key for token_key in FORBIDDEN_TOKEN_KEYWORDS):
+                return True
+            if _contains_forbidden_token_key(nested):
+                return True
+    elif isinstance(value, list):
+        return any(_contains_forbidden_token_key(item) for item in value)
+    return False
+
+
+def _safe_text(value, limit=300):
+    if value is None:
+        return None
+    text = str(value).strip()
+    return text[:limit] if text else None
+
+
+def _side_effects_clear(*payloads):
+    blocked_keys = (
+        "api_executes_cli",
+        "api_executes_llm",
+        "api_reads_approval_token",
+        "api_writes_file",
+        "api_writes_database",
+        "api_updates_review_state",
+        "run_package_file_written",
+        "summary_file_written",
+        "summary_record_written",
+        "summary_manifest_written",
+        "summary_receipt_file_written",
+        "database_write_executed",
+        "database_commit_executed",
+        "review_state_update_executed",
+        "llm_call_executed",
+        "ollama_call_executed",
+        "gemini_call_executed",
+        "telegram_dispatched",
+        "ai_summary_generated",
+        "external_network_executed",
+        "scheduler_attached",
+        "writes_executed",
+        "would_write_database",
+    )
+    return all(
+        not _as_dict(payload).get(key)
+        for payload in payloads
+        for key in blocked_keys
+    )
+
+
+def _operator_summary(operator_evidence):
+    operator_evidence = _as_dict(operator_evidence)
+    output = operator_evidence.get("ai_summary_output") or operator_evidence.get(
+        "summary_output"
+    )
+    return {
+        "provided_keys": sorted(operator_evidence.keys()),
+        "operator_confirmed_ai_summary_output_receipt": bool(
+            operator_evidence.get("operator_confirmed_ai_summary_output_receipt")
+            or operator_evidence.get("operator_confirmed_summary_output_receipt")
+        ),
+        "operator_confirmed_manual_output_only": bool(
+            operator_evidence.get("operator_confirmed_manual_output_only")
+            or operator_evidence.get("operator_confirmed_manual_ollama_summary_run")
+        ),
+        "operator_confirmed_ollama_first": bool(
+            operator_evidence.get("operator_confirmed_ollama_first")
+            or operator_evidence.get("operator_confirmed_ollama_cascade")
+        ),
+        "operator_confirmed_gemini_backup_only": bool(
+            operator_evidence.get("operator_confirmed_gemini_backup_only")
+            or operator_evidence.get("operator_confirmed_gemini_fallback_only")
+        ),
+        "operator_confirmed_no_api_llm_call": bool(
+            operator_evidence.get("operator_confirmed_no_api_llm_call")
+            or operator_evidence.get("operator_confirmed_no_llm_call")
+        ),
+        "operator_confirmed_no_telegram_dispatch": bool(
+            operator_evidence.get("operator_confirmed_no_telegram_dispatch")
+        ),
+        "operator_confirmed_no_scheduler_attach": bool(
+            operator_evidence.get("operator_confirmed_no_scheduler_attach")
+        ),
+        "operator_confirmed_no_api_db_write": bool(
+            operator_evidence.get("operator_confirmed_no_api_db_write")
+        ),
+        "ai_summary_output_artifact_path": _safe_text(
+            operator_evidence.get("ai_summary_output_artifact_path")
+            or operator_evidence.get("summary_output_artifact_path")
+        ),
+        "ai_summary_receipt_artifact_path": _safe_text(
+            operator_evidence.get("ai_summary_receipt_artifact_path")
+            or operator_evidence.get("summary_receipt_artifact_path")
+        ),
+        "manual_ai_summary_output_provided": isinstance(output, dict),
+        "approval_token_submitted_to_api": _contains_forbidden_token_key(
+            operator_evidence
+        ),
+    }
+
+
+def _run_package_payload(ai_summary_run_package):
+    package = _as_dict(ai_summary_run_package)
+    prompt = _as_dict(package.get("prompt_contract"))
+    input_contract = _as_dict(prompt.get("input_contract"))
+    output_schema = _as_dict(package.get("summary_output_schema"))
+    return {
+        "provided": bool(package),
+        "mode": package.get("mode"),
+        "ai_summary_run_package_ready": bool(
+            package.get("ai_summary_run_package_ready")
+        ),
+        "ready_for_manual_ollama_summary_run": bool(
+            package.get("ready_for_manual_ollama_summary_run")
+        ),
+        "model_route_policy": _as_dict(package.get("model_route_policy")),
+        "prompt_contract_version": prompt.get("contract_version"),
+        "summary_output_schema": output_schema,
+        "required_fields": _as_list(output_schema.get("required_fields")),
+        "expected_dedupe_keys": _as_list(package.get("expected_dedupe_keys")),
+        "summary_section_keys": _as_list(package.get("summary_section_keys")),
+        "artifact_paths": _as_dict(input_contract.get("artifact_paths")),
+        "safe_boundaries": _as_list(package.get("safe_boundaries")),
+        "blocked_reasons": _as_list(package.get("blocked_reasons")),
+        "read_only_query_executed": bool(package.get("read_only_query_executed")),
+        "database_connection_opened": bool(package.get("database_connection_opened")),
+        "database_write_executed": bool(package.get("database_write_executed")),
+        "database_commit_executed": bool(package.get("database_commit_executed")),
+        "review_state_update_executed": bool(
+            package.get("review_state_update_executed")
+        ),
+        "llm_call_executed": bool(package.get("llm_call_executed")),
+        "ollama_call_executed": bool(package.get("ollama_call_executed")),
+        "gemini_call_executed": bool(package.get("gemini_call_executed")),
+        "telegram_dispatched": bool(package.get("telegram_dispatched")),
+        "ai_summary_generated": bool(package.get("ai_summary_generated")),
+        "scheduler_attached": bool(package.get("scheduler_attached")),
+    }
+
+
+def _summary_output(operator_evidence):
+    output = _as_dict(operator_evidence).get("ai_summary_output") or _as_dict(
+        operator_evidence
+    ).get("summary_output")
+    return _as_dict(output)
+
+
+def _required_field_status(*, output, required_fields):
+    missing = [field for field in required_fields if field not in output]
+    empty_allowed = {"risk_flags"}
+    empty = [
+        field
+        for field in required_fields
+        if field in output
+        and field not in empty_allowed
+        and output.get(field) in (None, "", [], {})
+    ]
+    return missing, empty
+
+
+def _output_shape_valid(output):
+    if not output:
+        return False
+    if not isinstance(output.get("headline"), str):
+        return False
+    if not isinstance(output.get("executive_summary"), str):
+        return False
+    for key in ("key_findings", "risk_flags", "recommended_actions", "evidence_refs"):
+        if not isinstance(output.get(key), list):
+            return False
+    return isinstance(output.get("model_route"), dict)
+
+
+def _allowed_evidence_refs(package):
+    artifact_paths = package["artifact_paths"]
+    return sorted(
+        set(str(item) for item in package["expected_dedupe_keys"])
+        | set(str(item) for item in package["summary_section_keys"])
+        | set(str(key) for key in artifact_paths.keys())
+        | set(str(value) for value in artifact_paths.values() if value)
+    )
+
+
+def _evidence_refs_grounded(*, output, package):
+    refs = [str(item) for item in _as_list(output.get("evidence_refs")) if item]
+    allowed = set(_allowed_evidence_refs(package))
+    return bool(refs) and all(ref in allowed for ref in refs), refs
+
+
+def _model_route_accepted(*, output, package):
+    model_route = _as_dict(output.get("model_route"))
+    route_policy = package["model_route_policy"]
+    cascade = _as_list(route_policy.get("primary_cascade"))
+    allowed_hosts = {str(_as_dict(node).get("host")) for node in cascade}
+    provider = model_route.get("provider")
+    host = model_route.get("host")
+    fallback_reason = _safe_text(model_route.get("fallback_reason"))
+    failed_hosts = set(str(host) for host in _as_list(model_route.get("ollama_failed_hosts")))
+    if route_policy.get("primary_policy") != "ollama_first":
+        return False
+    if route_policy.get("app_host_forbidden_as_ollama_node") == host:
+        return False
+    if provider == "ollama":
+        return bool(host in allowed_hosts)
+    if provider == "gemini_fallback":
+        return bool(
+            route_policy.get("fallback_policy")
+            == "gemini_backup_only_after_ollama_cascade_failure"
+            and fallback_reason
+            and len(failed_hosts & allowed_hosts) == len(allowed_hosts)
+        )
+    return False
+
+
+def _receipt_gates(*, package, output, operator):
+    missing_fields, empty_fields = _required_field_status(
+        output=output,
+        required_fields=package["required_fields"],
+    )
+    grounded, refs = _evidence_refs_grounded(output=output, package=package)
+    route_accepted = _model_route_accepted(output=output, package=package)
+    return [
+        {
+            "key": "ai_summary_run_package_ready",
+            "label": "AI summary run package 必須已通過",
+            "passed": bool(
+                package["provided"]
+                and package["mode"]
+                == "candidate_queue_review_ai_summary_run_package_preview"
+                and package["ai_summary_run_package_ready"]
+                and package["ready_for_manual_ollama_summary_run"]
+            ),
+        },
+        {
+            "key": "manual_ai_summary_output_provided",
+            "label": "必須提供人工 Ollama 摘要輸出 JSON",
+            "passed": bool(output and operator["manual_ai_summary_output_provided"]),
+        },
+        {
+            "key": "summary_output_required_fields_present",
+            "label": "摘要輸出必須包含 schema 必要欄位",
+            "passed": bool(package["required_fields"] and not missing_fields and not empty_fields),
+        },
+        {
+            "key": "summary_output_shape_valid",
+            "label": "摘要輸出欄位型別必須符合 contract",
+            "passed": _output_shape_valid(output),
+        },
+        {
+            "key": "summary_output_evidence_refs_grounded",
+            "label": "摘要 evidence_refs 必須對應 dedupe key、section 或 artifact",
+            "passed": grounded,
+        },
+        {
+            "key": "summary_output_model_route_accepted",
+            "label": "摘要 model_route 必須符合 Ollama-first / Gemini fallback-only",
+            "passed": route_accepted,
+        },
+        {
+            "key": "operator_confirmed_ai_summary_output_receipt",
+            "label": "操作員確認本步只驗收人工摘要輸出",
+            "passed": bool(
+                operator["operator_confirmed_ai_summary_output_receipt"]
+                and operator["operator_confirmed_manual_output_only"]
+                and operator["operator_confirmed_ollama_first"]
+                and operator["operator_confirmed_gemini_backup_only"]
+                and operator["operator_confirmed_no_api_llm_call"]
+                and operator["operator_confirmed_no_telegram_dispatch"]
+                and operator["operator_confirmed_no_api_db_write"]
+                and operator["operator_confirmed_no_scheduler_attach"]
+            ),
+        },
+        {
+            "key": "ai_summary_output_receipt_artifacts_recorded",
+            "label": "summary output 與 receipt artifact path 必須由操作員提供",
+            "passed": bool(
+                operator["ai_summary_output_artifact_path"]
+                and operator["ai_summary_receipt_artifact_path"]
+            ),
+        },
+        {
+            "key": "ai_summary_output_receipt_no_approval_token_submitted_to_api",
+            "label": "payload 不得包含一次性 approval token key",
+            "passed": not operator["approval_token_submitted_to_api"],
+        },
+        {
+            "key": "ai_summary_output_receipt_has_no_side_effects",
+            "label": "receipt 不得呼叫 LLM、派送 Telegram、寫檔、寫 DB 或掛 scheduler",
+            "passed": _side_effects_clear(package),
+        },
+    ], {
+        "missing_fields": missing_fields,
+        "empty_fields": empty_fields,
+        "evidence_refs": refs,
+        "allowed_evidence_refs": _allowed_evidence_refs(package),
+        "model_route_accepted": route_accepted,
+    }
+
+
+def build_candidate_queue_review_ai_summary_output_receipt(
+    *,
+    ai_summary_run_package,
+    operator_evidence=None,
+    execute_requested=False,
+):
+    """建立人工 AI summary output receipt 預覽；不呼叫模型或執行副作用。"""
+    package = _run_package_payload(ai_summary_run_package)
+    operator = _operator_summary(operator_evidence)
+    output = _summary_output(operator_evidence)
+    gates, validation = _receipt_gates(
+        package=package,
+        output=output,
+        operator=operator,
+    )
+    blocked_reasons = [gate["key"] for gate in gates if not gate["passed"]]
+    receipt_ready = bool(not blocked_reasons)
+
+    return {
+        "mode": "candidate_queue_review_ai_summary_output_receipt_preview",
+        "target_table": TARGET_TABLE,
+        "target_operation": "review_manual_ollama_ai_summary_output",
+        "execute_requested": bool(execute_requested),
+        "ai_summary_output_receipt_ready": receipt_ready,
+        "ready_for_summary_persistence_review": receipt_ready,
+        "ready_for_next_manual_phase": receipt_ready,
+        "manual_ai_summary_output_provided": operator[
+            "manual_ai_summary_output_provided"
+        ],
+        "summary_output_schema_valid": bool(
+            output and not validation["missing_fields"] and not validation["empty_fields"]
+        ),
+        "summary_output_evidence_refs_grounded": bool(
+            validation["evidence_refs"]
+            and all(
+                ref in set(validation["allowed_evidence_refs"])
+                for ref in validation["evidence_refs"]
+            )
+        ),
+        "summary_output_model_route_accepted": validation["model_route_accepted"],
+        "ready_for_ai_summary_generation": False,
+        "ready_for_llm_call": False,
+        "ready_for_telegram_dispatch": False,
+        "ready_for_api_review_state_update": False,
+        "ready_for_api_database_write": False,
+        "ready_for_scheduler_attach": False,
+        "api_executes_cli": False,
+        "api_executes_llm": False,
+        "api_reads_approval_token": False,
+        "api_writes_file": False,
+        "api_writes_database": False,
+        "api_updates_review_state": False,
+        "approval_record_written": False,
+        "decision_record_written": False,
+        "run_package_file_written": False,
+        "summary_file_written": False,
+        "summary_record_written": False,
+        "summary_manifest_written": False,
+        "summary_receipt_file_written": False,
+        "ai_summary_generated": False,
+        "llm_call_executed": False,
+        "ollama_call_executed": False,
+        "gemini_call_executed": False,
+        "telegram_dispatched": False,
+        "review_state_update_executed": False,
+        "read_only_query_executed": package["read_only_query_executed"],
+        "database_connection_opened": package["database_connection_opened"],
+        "database_session_created": False,
+        "explicit_transaction_opened": False,
+        "transaction_opened": False,
+        "transaction_committed": False,
+        "database_write_executed": False,
+        "database_commit_executed": False,
+        "database_rollback_executed": False,
+        "external_network_executed": False,
+        "scheduler_attached": False,
+        "writes_executed": False,
+        "would_write_database": False,
+        "expected_dedupe_keys": package["expected_dedupe_keys"],
+        "summary_section_keys": package["summary_section_keys"],
+        "summary_output_required_fields": package["required_fields"],
+        "summary_output_validation": validation,
+        "model_route_policy": package["model_route_policy"],
+        "operator_ai_summary_output_receipt": operator,
+        "blocked_reasons": blocked_reasons,
+        "gates": gates,
+        "next_operator_steps": [
+            "確認人工摘要 output JSON 已符合 schema 與 evidence_refs",
+            "若要持久化 summary 或派送 Telegram，必須另開 persistence/dispatch gate",
+            "保留 Ollama host 或 Gemini fallback reason 供後續稽核",
+            "本 receipt 不寫檔、不寫 DB、不呼叫模型、不發 Telegram",
+        ],
+        "safe_boundaries": [
+            "do_not_call_llm_from_ai_summary_output_receipt",
+            "do_not_dispatch_telegram_from_ai_summary_output_receipt",
+            "do_not_write_summary_receipt_file_from_api",
+            "do_not_write_summary_file_from_api",
+            "do_not_insert_summary_record_from_api",
+            "do_not_update_review_state_from_ai_summary_output_receipt",
+            "do_not_read_approval_token_from_ai_summary_output_receipt",
+            "do_not_execute_cli_from_ai_summary_output_receipt",
+            "do_not_attach_scheduler_from_ai_summary_output_receipt",
+            "manual_ai_summary_output_only",
+            "ollama_cascade_must_use_gcp_a_then_gcp_b_then_111",
+            "gemini_backup_only_after_ollama_cascade_failure",
+            "host_188_forbidden_as_ollama_node",
+            "ai_summary_output_receipt_preview_only",
+            "no_remove_orphans",
+            "no_momo_db_lifecycle_change",
+        ],
+    }

services/market_intel/deployment_readiness.py

@@ -28,6 +28,7 @@ from services.market_intel.candidate_queue_review_decision_post_closeout_invento
 from services.market_intel.candidate_queue_review_completion_archive import build_candidate_queue_review_completion_archive
 from services.market_intel.candidate_queue_review_archive_summary import build_candidate_queue_review_archive_summary
 from services.market_intel.candidate_queue_review_ai_summary_preflight import build_candidate_queue_review_ai_summary_preflight
+from services.market_intel.candidate_queue_review_ai_summary_output_receipt import build_candidate_queue_review_ai_summary_output_receipt
 from services.market_intel.candidate_queue_review_ai_summary_run_package import build_candidate_queue_review_ai_summary_run_package
 
 
@@ -49,6 +50,7 @@ BLOCKED_RUN_REVIEW_KEYS = (
     "summary_file_written",
     "summary_record_written",
     "summary_manifest_written",
+    "summary_receipt_file_written",
     "ai_summary_generated",
     "llm_call_executed",
     "ollama_call_executed",
@@ -67,7 +69,7 @@ BLOCKED_RUN_REVIEW_KEYS = (
     "writes_executed",
     "would_write_database",
 )
-PRODUCTION_SMOKE_TARGETS = ("/health", "/market_intel", "/api/market_intel/status", "/api/market_intel/deployment_readiness", "/api/market_intel/schema_smoke", "/api/market_intel/schema_db_probe", "/api/market_intel/platform_seed_db_diff", "/api/market_intel/legacy_source_bridge", "/api/market_intel/mcp_readiness", "/api/market_intel/mcp_tool_contract", "/api/market_intel/mcp_deploy_preflight", "/api/market_intel/mcp_activation_runbook", "/api/market_intel/mcp_fetch_gate", "/api/market_intel/scheduler_plan", "/api/market_intel/manual_sample_plan", "/api/market_intel/manual_sample_acceptance", "/api/market_intel/manual_sample_review", "/api/market_intel/match_review_plan", "/api/market_intel/opportunity_plan", "/api/market_intel/opportunity_scoring_plan", "/api/market_intel/opportunity_evidence_plan", "/api/market_intel/opportunity_alert_plan", "/api/market_intel/migration_apply_drill", "/api/market_intel/migration_catalog_review", "/api/market_intel/migration_live_smoke", "/api/market_intel/live_db_inventory", "/api/market_intel/manual_sample_review/candidate_queue_writer_postwrite_smoke", "/api/market_intel/manual_sample_review/candidate_queue_writer_operator_drill", "/api/market_intel/manual_sample_review/candidate_queue_writer_run_package", "/api/market_intel/manual_sample_review/candidate_queue_writer_run_readiness", "/api/market_intel/manual_sample_review/candidate_queue_writer_run_receipt", "/api/market_intel/manual_sample_review/candidate_queue_writer_run_closeout", "/api/market_intel/manual_sample_review/candidate_queue_review_handoff", "/api/market_intel/manual_sample_review/candidate_queue_review_inventory", "/api/market_intel/manual_sample_review/candidate_queue_review_decision", "/api/market_intel/manual_sample_review/candidate_queue_review_decision_approval", "/api/market_intel/manual_sample_review/candidate_queue_review_decision_transaction", "/api/market_intel/manual_sample_review/candidate_queue_review_decision_writer_preflight", "/api/market_intel/manual_sample_review/candidate_queue_review_decision_writer_postwrite_smoke", "/api/market_intel/manual_sample_review/candidate_queue_review_decision_writer_operator_drill", "/api/market_intel/manual_sample_review/candidate_queue_review_decision_writer_run_package", "/api/market_intel/manual_sample_review/candidate_queue_review_decision_writer_run_readiness", "/api/market_intel/manual_sample_review/candidate_queue_review_decision_writer_run_receipt", "/api/market_intel/manual_sample_review/candidate_queue_review_decision_writer_run_closeout", "/api/market_intel/manual_sample_review/candidate_queue_review_decision_post_closeout_inventory", "/api/market_intel/manual_sample_review/candidate_queue_review_completion_archive", "/api/market_intel/manual_sample_review/candidate_queue_review_archive_summary", "/api/market_intel/manual_sample_review/candidate_queue_review_ai_summary_preflight", "/api/market_intel/manual_sample_review/candidate_queue_review_ai_summary_run_package", "/api/market_intel/manual_sample_review/candidate_queue_review_decision_writer_status")
+PRODUCTION_SMOKE_TARGETS = ("/health", "/market_intel", "/api/market_intel/status", "/api/market_intel/deployment_readiness", "/api/market_intel/schema_smoke", "/api/market_intel/schema_db_probe", "/api/market_intel/platform_seed_db_diff", "/api/market_intel/legacy_source_bridge", "/api/market_intel/mcp_readiness", "/api/market_intel/mcp_tool_contract", "/api/market_intel/mcp_deploy_preflight", "/api/market_intel/mcp_activation_runbook", "/api/market_intel/mcp_fetch_gate", "/api/market_intel/scheduler_plan", "/api/market_intel/manual_sample_plan", "/api/market_intel/manual_sample_acceptance", "/api/market_intel/manual_sample_review", "/api/market_intel/match_review_plan", "/api/market_intel/opportunity_plan", "/api/market_intel/opportunity_scoring_plan", "/api/market_intel/opportunity_evidence_plan", "/api/market_intel/opportunity_alert_plan", "/api/market_intel/migration_apply_drill", "/api/market_intel/migration_catalog_review", "/api/market_intel/migration_live_smoke", "/api/market_intel/live_db_inventory", "/api/market_intel/manual_sample_review/candidate_queue_writer_postwrite_smoke", "/api/market_intel/manual_sample_review/candidate_queue_writer_operator_drill", "/api/market_intel/manual_sample_review/candidate_queue_writer_run_package", "/api/market_intel/manual_sample_review/candidate_queue_writer_run_readiness", "/api/market_intel/manual_sample_review/candidate_queue_writer_run_receipt", "/api/market_intel/manual_sample_review/candidate_queue_writer_run_closeout", "/api/market_intel/manual_sample_review/candidate_queue_review_handoff", "/api/market_intel/manual_sample_review/candidate_queue_review_inventory", "/api/market_intel/manual_sample_review/candidate_queue_review_decision", "/api/market_intel/manual_sample_review/candidate_queue_review_decision_approval", "/api/market_intel/manual_sample_review/candidate_queue_review_decision_transaction", "/api/market_intel/manual_sample_review/candidate_queue_review_decision_writer_preflight", "/api/market_intel/manual_sample_review/candidate_queue_review_decision_writer_postwrite_smoke", "/api/market_intel/manual_sample_review/candidate_queue_review_decision_writer_operator_drill", "/api/market_intel/manual_sample_review/candidate_queue_review_decision_writer_run_package", "/api/market_intel/manual_sample_review/candidate_queue_review_decision_writer_run_readiness", "/api/market_intel/manual_sample_review/candidate_queue_review_decision_writer_run_receipt", "/api/market_intel/manual_sample_review/candidate_queue_review_decision_writer_run_closeout", "/api/market_intel/manual_sample_review/candidate_queue_review_decision_post_closeout_inventory", "/api/market_intel/manual_sample_review/candidate_queue_review_completion_archive", "/api/market_intel/manual_sample_review/candidate_queue_review_archive_summary", "/api/market_intel/manual_sample_review/candidate_queue_review_ai_summary_preflight", "/api/market_intel/manual_sample_review/candidate_queue_review_ai_summary_run_package", "/api/market_intel/manual_sample_review/candidate_queue_review_ai_summary_output_receipt", "/api/market_intel/manual_sample_review/candidate_queue_review_decision_writer_status")
 
 
 def _run_review_preview_safe(payload, mode):
@@ -222,6 +224,9 @@ def build_deployment_readiness_preview(
         archive_summary=candidate_queue_review_archive_summary,
         ai_summary_preflight=candidate_queue_review_ai_summary_preflight,
     )
+    candidate_queue_review_ai_summary_output_receipt = build_candidate_queue_review_ai_summary_output_receipt(
+        ai_summary_run_package=candidate_queue_review_ai_summary_run_package,
+    )
     checks = {
         "schema_smoke_passed": bool(schema_smoke["passed"]),
         "feature_flags_default_safe": bool(
@@ -508,6 +513,10 @@ def build_deployment_readiness_preview(
             candidate_queue_review_ai_summary_run_package,
             "candidate_queue_review_ai_summary_run_package_preview",
         ),
+        "candidate_queue_review_ai_summary_output_receipt_preview_safe": _run_review_preview_safe(
+            candidate_queue_review_ai_summary_output_receipt,
+            "candidate_queue_review_ai_summary_output_receipt_preview",
+        ),
         "candidate_queue_review_decision_writer_cli_status_safe": _run_review_preview_safe(
             candidate_queue_review_decision_writer_status,
             "candidate_queue_review_decision_writer_cli_blocked",
@@ -756,6 +765,7 @@ def build_deployment_readiness_preview(
         "candidate_queue_review_archive_summary": candidate_queue_review_archive_summary,
         "candidate_queue_review_ai_summary_preflight": candidate_queue_review_ai_summary_preflight,
         "candidate_queue_review_ai_summary_run_package": candidate_queue_review_ai_summary_run_package,
+        "candidate_queue_review_ai_summary_output_receipt": candidate_queue_review_ai_summary_output_receipt,
         "candidate_queue_review_decision_writer_status": candidate_queue_review_decision_writer_status,
         "match_review_plan": match_review_plan,
         "opportunity_plan": opportunity_plan,

services/market_intel/phase.py

@@ -1,3 +1,3 @@
 """市場情報 rollout phase 單一來源。"""
 
-MARKET_INTEL_PHASE = "phase_81_candidate_queue_review_ai_summary_run_package"
+MARKET_INTEL_PHASE = "phase_82_candidate_queue_review_ai_summary_output_receipt"

templates/market_intel/disabled.html

@@ -682,6 +682,9 @@
                 <button class="market-intel-icon-button" type="button" title="產生 queue review AI summary run package" data-market-intel-sample-candidate-queue-review-ai-summary-run-package>
                     <i class="fas fa-box-open" aria-hidden="true"></i>
                 </button>
+                <button class="market-intel-icon-button" type="button" title="驗收 queue review AI summary output receipt" data-market-intel-sample-candidate-queue-review-ai-summary-output-receipt>
+                    <i class="fas fa-file-circle-check" aria-hidden="true"></i>
+                </button>
             </div>
         </div>
     </div>
@@ -1037,6 +1040,7 @@
         const sampleCandidateQueueReviewArchiveSummary = sampleReviewRoot ? sampleReviewRoot.querySelector('[data-market-intel-sample-candidate-queue-review-archive-summary]') : null;
         const sampleCandidateQueueReviewAiSummaryPreflight = sampleReviewRoot ? sampleReviewRoot.querySelector('[data-market-intel-sample-candidate-queue-review-ai-summary-preflight]') : null;
         const sampleCandidateQueueReviewAiSummaryRunPackage = sampleReviewRoot ? sampleReviewRoot.querySelector('[data-market-intel-sample-candidate-queue-review-ai-summary-run-package]') : null;
+        const sampleCandidateQueueReviewAiSummaryOutputReceipt = sampleReviewRoot ? sampleReviewRoot.querySelector('[data-market-intel-sample-candidate-queue-review-ai-summary-output-receipt]') : null;
         const sampleReviewEndpoint = "{{ url_for('market_intel.market_intel_manual_sample_review') }}";
         const sampleReviewEvaluateEndpoint = "{{ url_for('market_intel.market_intel_manual_sample_review_evaluate') }}";
         const sampleCandidateHandoffEndpoint = "{{ url_for('market_intel.market_intel_manual_sample_candidate_handoff') }}";
@@ -1069,6 +1073,7 @@
         const sampleCandidateQueueReviewArchiveSummaryEndpoint = "{{ url_for('market_intel_review.market_intel_manual_sample_candidate_queue_review_archive_summary') }}";
         const sampleCandidateQueueReviewAiSummaryPreflightEndpoint = "{{ url_for('market_intel_review.market_intel_manual_sample_candidate_queue_review_ai_summary_preflight') }}";
         const sampleCandidateQueueReviewAiSummaryRunPackageEndpoint = "{{ url_for('market_intel_review.market_intel_manual_sample_candidate_queue_review_ai_summary_run_package') }}";
+        const sampleCandidateQueueReviewAiSummaryOutputReceiptEndpoint = "{{ url_for('market_intel_review.market_intel_manual_sample_candidate_queue_review_ai_summary_output_receipt') }}";
         const schedulerMeta = schedulerRoot ? schedulerRoot.querySelector('[data-market-intel-scheduler-meta]') : null;
         const schedulerBody = schedulerRoot ? schedulerRoot.querySelector('[data-market-intel-scheduler-body]') : null;
         const schedulerRefresh = schedulerRoot ? schedulerRoot.querySelector('[data-market-intel-scheduler-refresh]') : null;
@@ -4949,6 +4954,108 @@
             }
         };
 
+        const renderCandidateQueueReviewAiSummaryOutputReceipt = data => {
+            const blockers = (data.blocked_reasons || []).join(' / ');
+            const gates = data.gates || [];
+            const validation = data.summary_output_validation || {};
+            const routePolicy = data.model_route_policy || {};
+            sampleReviewMeta.innerHTML = [
+                `mode=${data.mode || 'unknown'}`,
+                `receipt=${data.ai_summary_output_receipt_ready ? 'ready' : 'blocked'}`,
+                `schema=${data.summary_output_schema_valid ? 'pass' : 'blocked'}`,
+                `refs=${data.summary_output_evidence_refs_grounded ? 'pass' : 'blocked'}`,
+                `llm=${data.llm_call_executed ? 'called' : 'blocked'}`
+            ].map(item => `<span class="market-intel-pill">${escapeHtml(item)}</span>`).join('');
+            sampleReviewBody.innerHTML = `
+                <div class="market-intel-empty mb-3">此卡只驗收人工 Ollama 摘要輸出；API/UI 不呼叫模型、不寫 receipt、不派送 Telegram、不更新 review_state、不寫 DB。${blockers ? `阻擋：${escapeHtml(blockers)}` : ''}</div>
+                <div class="market-intel-deploy-grid">
+                    <div>
+                        <p class="market-intel-deploy-section-title">RECEIPT GATES</p>
+                        <div class="market-intel-check-list">${
+                            gates.map(gate => `
+                                <div class="market-intel-check">
+                                    <div>
+                                        <strong>${escapeHtml(gate.key)}</strong>
+                                        <small>${escapeHtml(gate.label)}</small>
+                                    </div>
+                                    <span>${gate.passed ? 'PASS' : 'BLOCK'}</span>
+                                </div>
+                            `).join('') || '<div class="market-intel-empty">尚未提供 output receipt gate。</div>'
+                        }</div>
+                    </div>
+                    <div>
+                        <p class="market-intel-deploy-section-title">OUTPUT VALIDATION</p>
+                        <div class="market-intel-check-list">
+                            ${[
+                                ['missing_fields', (validation.missing_fields || []).join(', ') || 'none'],
+                                ['empty_fields', (validation.empty_fields || []).join(', ') || 'none'],
+                                ['evidence_refs', (validation.evidence_refs || []).join(', ') || 'none'],
+                                ['route_accepted', validation.model_route_accepted],
+                                ['primary_policy', routePolicy.primary_policy || 'missing']
+                            ].map(([key, value]) => `
+                                <div class="market-intel-check">
+                                    <div><strong>${escapeHtml(key)}</strong></div>
+                                    <span>${escapeHtml(String(value))}</span>
+                                </div>
+                            `).join('')}
+                        </div>
+                    </div>
+                    <div>
+                        <p class="market-intel-deploy-section-title">NO-SIDE-EFFECTS</p>
+                        <div class="market-intel-check-list">
+                            ${[
+                                ['api_executes_llm', data.api_executes_llm],
+                                ['llm_call_executed', data.llm_call_executed],
+                                ['ollama_call_executed', data.ollama_call_executed],
+                                ['gemini_call_executed', data.gemini_call_executed],
+                                ['telegram_dispatched', data.telegram_dispatched],
+                                ['summary_receipt_file_written', data.summary_receipt_file_written],
+                                ['summary_file_written', data.summary_file_written],
+                                ['database_write_executed', data.database_write_executed],
+                                ['scheduler_attached', data.scheduler_attached]
+                            ].map(([key, value]) => `
+                                <div class="market-intel-check">
+                                    <div><strong>${escapeHtml(key)}</strong></div>
+                                    <span>${escapeHtml(String(value))}</span>
+                                </div>
+                            `).join('')}
+                        </div>
+                    </div>
+                </div>
+            `;
+        };
+
+        const loadCandidateQueueReviewAiSummaryOutputReceipt = async () => {
+            if (!sampleReviewMeta || !sampleReviewBody || !sampleReviewInput) return;
+            let parsed;
+            try {
+                parsed = JSON.parse(sampleReviewInput.value || '{}');
+            } catch (error) {
+                sampleReviewMeta.innerHTML = '<span class="market-intel-pill">json_error</span>';
+                sampleReviewBody.innerHTML = `<div class="market-intel-empty">JSON 格式錯誤：${escapeHtml(error.message)}</div>`;
+                return;
+            }
+            const body = parsed && parsed.sample_result ? parsed : { sample_result: parsed };
+            sampleReviewBody.innerHTML = '<div class="market-intel-empty">驗收 queue review AI summary output receipt 中...</div>';
+            try {
+                const response = await fetch(sampleCandidateQueueReviewAiSummaryOutputReceiptEndpoint, {
+                    method: 'POST',
+                    credentials: 'same-origin',
+                    headers: {
+                        'Content-Type': 'application/json',
+                        'X-CSRFToken': csrfToken
+                    },
+                    body: JSON.stringify(body)
+                });
+                const data = await response.json();
+                if (!response.ok && !data.mode) throw new Error(`HTTP ${response.status}`);
+                renderCandidateQueueReviewAiSummaryOutputReceipt(data);
+            } catch (error) {
+                sampleReviewMeta.innerHTML = '<span class="market-intel-pill">error</span>';
+                sampleReviewBody.innerHTML = `<div class="market-intel-empty">queue review AI summary output receipt 失敗：${escapeHtml(error.message)}</div>`;
+            }
+        };
+
         const renderCandidateQueueReviewDecisionWriter = data => {
             const blockers = (data.blocked_reasons || []).join(' / ');
             const summary = data.statement_summary || {};
@@ -6646,6 +6753,9 @@
         if (sampleCandidateQueueReviewAiSummaryRunPackage) {
             sampleCandidateQueueReviewAiSummaryRunPackage.addEventListener('click', loadCandidateQueueReviewAiSummaryRunPackage);
         }
+        if (sampleCandidateQueueReviewAiSummaryOutputReceipt) {
+            sampleCandidateQueueReviewAiSummaryOutputReceipt.addEventListener('click', loadCandidateQueueReviewAiSummaryOutputReceipt);
+        }
         if (schedulerRefresh) {
             schedulerRefresh.addEventListener('click', loadScheduler);
         }