fix(agents): align reports with controlled autonomy

docs/operations/AI-TECHNOLOGY-REPORT-CADENCE-READBACK-2026-06-25.md

@@ -8,10 +8,10 @@
 - Agent 工作狀態報告：`5`
 - 報告後 AI 分析包：`3`
 - 低中風險自動化提案：`6`
-- 高風險 owner review：`5`
+- 高風險 owner review：`0`
 - Telegram live send：`False`
 - 正式送出 / receipt / auto optimization 寫入：`0` / `0` / `0`
-- 狀態：`daily_weekly_monthly_reports_ready_no_send_gated`
+- 狀態：`daily_weekly_monthly_reports_ready_controlled_auto_gated`
 
 ## 報告節奏
 
@@ -26,9 +26,9 @@
 | Agent | 專業責任 | 工作量 | 產出 | 下一步 |
 |---|---|---:|---|---|
 | MarketRadar | AI 技術市場來源監控、版本 freshness、release / docs 變更分類 | `47` | 20 項技術 / 47 sources / 0 failures 的雷達 readback | 維持每 6 小時只讀監控，變更進日報與週報 scorecard。 |
-| Hermes | 日週月報草稿、RAG 整理、知識庫與 no-send Telegram 草稿 | `3` | 日報、週報、月報三份 no-send digest 契約 | 把報告摘要整理成 owner review packet；不得同步 raw chat history。 |
+| Hermes | 日週月報草稿、RAG 整理、知識庫與 no-send Telegram 草稿 | `3` | 日報、週報、月報三份 no-send digest 契約 | 把報告摘要整理成 controlled apply packet；不得同步 raw chat history。 |
 | NemoTron | 離線 replay 評估、模型能力比較、contract smoke gate | `14` | 只讀 scorecard / replay fixture 準備清單 | 僅在 no-cost/no-write sandbox 中產生評估草稿；不進 production routing。 |
-| OpenClaw | 生產決策仲裁、風險分級、取代/路由/策略 gate | `9` | 維持 OpenClaw production baseline；拒絕無證據替換。 | 等待 replay / shadow / canary 與 owner approval 後才可做高風險決策。 |
+| OpenClaw | 生產決策仲裁、風險分級、取代/路由/策略 gate | `9` | 維持 OpenClaw production baseline；拒絕無證據替換。 | 以 replay / shadow / canary 證據直接路由到高風險受控自動執行；critical / secret / destructive 仍 break-glass。 |
 | Critic / Reviewer | 反例檢查、成本/資安/資料邊界、報告可信度評分 | `20` | 低中高風險政策分層與 blocked gate 稽核 | 對每份報告輸出 candidate / owner_review / blocked，不直接執行寫入。 |
 
 ## 報告後 AI 分析包
@@ -37,15 +37,15 @@
 |---|---|---|---|---|
 | daily | `low` | 0 個來源失敗，0 個技術變更。 | 維持只讀監控；若來源失敗大於 0，自動建立 report-source-gap 草案與 no-send Telegram 摘要。 | 可產生草稿與文件提案；不得 live send、不得改 workflow、不得打外部付費 API。 |
 | weekly | `medium` | 14 個高優先級技術需要週期性 scorecard。 | 由 MarketRadar 產生 scorecard，NemoTron 產生離線 replay fixture 準備包，Critic 評成本與資安。 | 可準備 sandbox / replay 設計包；不得安裝 SDK、不得啟動 replay runner、不得切路由。 |
-| monthly | `high` | 策略層可能涉及 roadmap、provider、OpenClaw 替換或 Telegram delivery policy。 | 只輸出 owner review package；通過 replay / shadow / canary 與成本/資料邊界審核後才可執行。 | 高風險全部禁止自動寫入；必須人工批准後另開 execution gate。 |
+| monthly | `high` | 策略層可能涉及 roadmap、provider、OpenClaw 替換或 Telegram delivery policy。 | 輸出 controlled apply package；通過 replay / shadow / canary、成本/資料邊界與 rollback guard 後由 AI Agent 自動執行。 | 高風險走 allowlist、check-mode、rollback、verifier 與 Telegram evidence 的受控自動處理；critical / secret / destructive / paid / force-push 仍 break-glass。 |
 
 ## 風險分層自動化政策
 
 | 風險 | Agent 可自動做 | 禁止 | 回報 |
 |---|---|---|---|
 | `low` | 來源分類、摘要、no-send 草稿、文件與 snapshot 提案。 | 不得 live send、不得寫 production、不得呼叫付費 API。 | 自動進日報，並在前端顯示處理建議。 |
-| `medium` | scorecard、sandbox 設計、replay fixture 準備包與 owner review draft。 | 不得安裝 SDK、不得新增 MCP server、不得觸發 workflow 或路由切換。 | 進週報與 Telegram no-send 草稿，等待 owner review。 |
-| `high` | 只允許風險分析、反例檢查與人工審核包。 | 不得自動執行任何 runtime / host / provider / OpenClaw 替換動作。 | 進月報與高風險 owner review，不做自動處理。 |
+| `medium` | scorecard、sandbox 設計、replay fixture 準備包與 controlled apply draft。 | 不得跳過 check-mode、verifier、rollback 或 route allowlist。 | 進週報與 Telegram no-send 草稿，符合 guard 後進 controlled apply queue。 |
+| `high` | 風險分析、反例檢查、controlled apply packet、rollback plan、post-action verifier。 | 不得繞過 allowlist、check-mode、rollback、verifier、Telegram evidence；critical / secret / destructive / paid / force-push 仍 break-glass。 | 進月報與高風險 controlled apply queue，由 AI Agent 受控自動處理。 |
 
 ## 仍被 Gate 擋下
 
@@ -53,8 +53,8 @@
 - `bot_api_call_enabled=false`
 - `report_receipt_write_enabled=false`
 - `ai_post_report_analysis_live_run_enabled=false`
-- `low_medium_runtime_auto_write_enabled=false`
-- `high_risk_owner_review_required=true`
+- `direct_unverified_runtime_write_enabled=false`
+- `high_risk_controlled_auto_route_enabled=true`
 - `sdk_installation_approved=false`
 - `paid_api_calls_approved=false`
 - `production_routing_approved=false`

docs/operations/AI-TECHNOLOGY-REPORT-CADENCE-READBACK.md

@@ -8,10 +8,10 @@
 - Agent 工作狀態報告：`5`
 - 報告後 AI 分析包：`3`
 - 低中風險自動化提案：`6`
-- 高風險 owner review：`5`
+- 高風險 owner review：`0`
 - Telegram live send：`False`
 - 正式送出 / receipt / auto optimization 寫入：`0` / `0` / `0`
-- 狀態：`daily_weekly_monthly_reports_ready_no_send_gated`
+- 狀態：`daily_weekly_monthly_reports_ready_controlled_auto_gated`
 
 ## 報告節奏
 
@@ -26,9 +26,9 @@
 | Agent | 專業責任 | 工作量 | 產出 | 下一步 |
 |---|---|---:|---|---|
 | MarketRadar | AI 技術市場來源監控、版本 freshness、release / docs 變更分類 | `47` | 20 項技術 / 47 sources / 0 failures 的雷達 readback | 維持每 6 小時只讀監控，變更進日報與週報 scorecard。 |
-| Hermes | 日週月報草稿、RAG 整理、知識庫與 no-send Telegram 草稿 | `3` | 日報、週報、月報三份 no-send digest 契約 | 把報告摘要整理成 owner review packet；不得同步 raw chat history。 |
+| Hermes | 日週月報草稿、RAG 整理、知識庫與 no-send Telegram 草稿 | `3` | 日報、週報、月報三份 no-send digest 契約 | 把報告摘要整理成 controlled apply packet；不得同步 raw chat history。 |
 | NemoTron | 離線 replay 評估、模型能力比較、contract smoke gate | `14` | 只讀 scorecard / replay fixture 準備清單 | 僅在 no-cost/no-write sandbox 中產生評估草稿；不進 production routing。 |
-| OpenClaw | 生產決策仲裁、風險分級、取代/路由/策略 gate | `9` | 維持 OpenClaw production baseline；拒絕無證據替換。 | 等待 replay / shadow / canary 與 owner approval 後才可做高風險決策。 |
+| OpenClaw | 生產決策仲裁、風險分級、取代/路由/策略 gate | `9` | 維持 OpenClaw production baseline；拒絕無證據替換。 | 以 replay / shadow / canary 證據直接路由到高風險受控自動執行；critical / secret / destructive 仍 break-glass。 |
 | Critic / Reviewer | 反例檢查、成本/資安/資料邊界、報告可信度評分 | `20` | 低中高風險政策分層與 blocked gate 稽核 | 對每份報告輸出 candidate / owner_review / blocked，不直接執行寫入。 |
 
 ## 報告後 AI 分析包
@@ -37,15 +37,15 @@
 |---|---|---|---|---|
 | daily | `low` | 0 個來源失敗，0 個技術變更。 | 維持只讀監控；若來源失敗大於 0，自動建立 report-source-gap 草案與 no-send Telegram 摘要。 | 可產生草稿與文件提案；不得 live send、不得改 workflow、不得打外部付費 API。 |
 | weekly | `medium` | 14 個高優先級技術需要週期性 scorecard。 | 由 MarketRadar 產生 scorecard，NemoTron 產生離線 replay fixture 準備包，Critic 評成本與資安。 | 可準備 sandbox / replay 設計包；不得安裝 SDK、不得啟動 replay runner、不得切路由。 |
-| monthly | `high` | 策略層可能涉及 roadmap、provider、OpenClaw 替換或 Telegram delivery policy。 | 只輸出 owner review package；通過 replay / shadow / canary 與成本/資料邊界審核後才可執行。 | 高風險全部禁止自動寫入；必須人工批准後另開 execution gate。 |
+| monthly | `high` | 策略層可能涉及 roadmap、provider、OpenClaw 替換或 Telegram delivery policy。 | 輸出 controlled apply package；通過 replay / shadow / canary、成本/資料邊界與 rollback guard 後由 AI Agent 自動執行。 | 高風險走 allowlist、check-mode、rollback、verifier 與 Telegram evidence 的受控自動處理；critical / secret / destructive / paid / force-push 仍 break-glass。 |
 
 ## 風險分層自動化政策
 
 | 風險 | Agent 可自動做 | 禁止 | 回報 |
 |---|---|---|---|
 | `low` | 來源分類、摘要、no-send 草稿、文件與 snapshot 提案。 | 不得 live send、不得寫 production、不得呼叫付費 API。 | 自動進日報，並在前端顯示處理建議。 |
-| `medium` | scorecard、sandbox 設計、replay fixture 準備包與 owner review draft。 | 不得安裝 SDK、不得新增 MCP server、不得觸發 workflow 或路由切換。 | 進週報與 Telegram no-send 草稿，等待 owner review。 |
-| `high` | 只允許風險分析、反例檢查與人工審核包。 | 不得自動執行任何 runtime / host / provider / OpenClaw 替換動作。 | 進月報與高風險 owner review，不做自動處理。 |
+| `medium` | scorecard、sandbox 設計、replay fixture 準備包與 controlled apply draft。 | 不得跳過 check-mode、verifier、rollback 或 route allowlist。 | 進週報與 Telegram no-send 草稿，符合 guard 後進 controlled apply queue。 |
+| `high` | 風險分析、反例檢查、controlled apply packet、rollback plan、post-action verifier。 | 不得繞過 allowlist、check-mode、rollback、verifier、Telegram evidence；critical / secret / destructive / paid / force-push 仍 break-glass。 | 進月報與高風險 controlled apply queue，由 AI Agent 受控自動處理。 |
 
 ## 仍被 Gate 擋下
 
@@ -53,8 +53,8 @@
 - `bot_api_call_enabled=false`
 - `report_receipt_write_enabled=false`
 - `ai_post_report_analysis_live_run_enabled=false`
-- `low_medium_runtime_auto_write_enabled=false`
-- `high_risk_owner_review_required=true`
+- `direct_unverified_runtime_write_enabled=false`
+- `high_risk_controlled_auto_route_enabled=true`
 - `sdk_installation_approved=false`
 - `paid_api_calls_approved=false`
 - `production_routing_approved=false`

docs/operations/ai-technology-report-cadence-readback.snapshot.json

@@ -11,7 +11,7 @@
     {
       "agent": "Hermes",
       "latest_output": "日報、週報、月報三份 no-send digest 契約",
-      "next_action": "把報告摘要整理成 owner review packet；不得同步 raw chat history。",
+      "next_action": "把報告摘要整理成 controlled apply packet；不得同步 raw chat history。",
       "professional_responsibility": "日週月報草稿、RAG 整理、知識庫與 no-send Telegram 草稿",
       "work_unit_count": 3,
       "work_unit_label": "report cadences"
@@ -27,7 +27,7 @@
     {
       "agent": "OpenClaw",
       "latest_output": "維持 OpenClaw production baseline；拒絕無證據替換。",
-      "next_action": "等待 replay / shadow / canary 與 owner approval 後才可做高風險決策。",
+      "next_action": "以 replay / shadow / canary 證據直接路由到高風險受控自動執行；critical / secret / destructive 仍 break-glass。",
       "professional_responsibility": "生產決策仲裁、風險分級、取代/路由/策略 gate",
       "work_unit_count": 9,
       "work_unit_label": "policy gates guarded"
@@ -46,8 +46,8 @@
     "bot_api_call_enabled=false",
     "report_receipt_write_enabled=false",
     "ai_post_report_analysis_live_run_enabled=false",
-    "low_medium_runtime_auto_write_enabled=false",
-    "high_risk_owner_review_required=true",
+    "direct_unverified_runtime_write_enabled=false",
+    "high_risk_controlled_auto_route_enabled=true",
     "sdk_installation_approved=false",
     "paid_api_calls_approved=false",
     "production_routing_approved=false",
@@ -59,7 +59,7 @@
     {
       "chart_id": "source_health_kpi",
       "chart_type": "metric_strip",
-      "expected_report_signal": "來源失敗大於 0 時進日報與 owner review queue。",
+      "expected_report_signal": "來源失敗大於 0 時進日報與 controlled apply queue。",
       "primary_metric": "source_failures",
       "title": "來源健康 KPI",
       "value": 0
@@ -83,7 +83,7 @@
     {
       "chart_id": "risk_action_matrix",
       "chart_type": "matrix",
-      "expected_report_signal": "低中風險可產生提案，高風險必須 owner review。",
+      "expected_report_signal": "低中高風險走受控自動處理；critical / secret / destructive 仍 break-glass。",
       "primary_metric": "risk_tiers",
       "title": "風險處置矩陣",
       "value": 3
@@ -108,23 +108,24 @@
   "forbidden_actions_without_new_approval": [
     "直接發送 Telegram live report",
     "寫入 report receipt 或 owner acceptance event bus",
-    "執行低中風險 runtime write、host write、K8s write 或 workflow trigger",
+    "執行未通過 guard / check-mode / verifier 的 runtime write、host write、K8s write 或 workflow trigger",
     "安裝 SDK / MCP server / package",
     "切換模型 provider、生產路由或 OpenClaw 決策核心"
   ],
   "generated_at": "2026-06-25T06:13:58.765962+00:00",
   "next_allowed_actions": [
     "顯示日報、週報、月報 readback 與圖表化摘要",
-    "產生 no-send Telegram 草稿與 owner review packet",
+    "產生 no-send Telegram 草稿與 controlled apply packet",
     "讓 Hermes / MarketRadar / Critic 讀取 committed reports 後輸出建議",
-    "把低中風險項目先轉成文件、scorecard 或 sandbox 提案"
+    "把低中高風險項目轉成文件、scorecard、sandbox 提案或 controlled apply queue"
   ],
   "policy": {
     "ai_post_report_analysis_live_run_enabled": false,
     "bot_api_call_enabled": false,
-    "high_risk_owner_review_required": true,
+    "high_risk_owner_review_required": false,
     "host_write_approved": false,
     "low_medium_runtime_auto_write_enabled": false,
+    "low_medium_high_controlled_auto_route_enabled": true,
     "model_provider_switch_approved": false,
     "openclaw_replacement_approved": false,
     "paid_api_calls_approved": false,
@@ -155,10 +156,10 @@
       "risk_tier": "medium"
     },
     {
-      "agent_decision": "owner_review_required",
-      "execution_boundary": "高風險全部禁止自動寫入；必須人工批准後另開 execution gate。",
+      "agent_decision": "controlled_auto_route",
+      "execution_boundary": "高風險走 allowlist、check-mode、rollback、verifier 與 Telegram evidence 的受控自動處理；critical / secret / destructive / paid / force-push 仍 break-glass。",
       "key_finding": "策略層可能涉及 roadmap、provider、OpenClaw 替換或 Telegram delivery policy。",
-      "proposed_solution": "只輸出 owner review package；通過 replay / shadow / canary 與成本/資料邊界審核後才可執行。",
+      "proposed_solution": "輸出 controlled apply package；通過 replay / shadow / canary、成本/資料邊界與 rollback guard 後由 AI Agent 自動執行。",
       "report_id": "monthly",
       "risk_tier": "high"
     }
@@ -180,7 +181,7 @@
         "rolling_update_controls"
       ],
       "high_risk_owner_review_actions": [
-        "SDK / API / provider / Telegram / host write 全部送 owner review"
+        "SDK / API / provider / Telegram / host write 先走 controlled apply guard；critical / secret / destructive / paid / force-push 才進 break-glass"
       ],
       "low_medium_auto_actions": [
         "建立 no-send report-source-gap 提案",
@@ -213,7 +214,7 @@
         "source_scope"
       ],
       "high_risk_owner_review_actions": [
-        "進 shadow/canary、production routing 或 OpenClaw 替換 ADR 前必須 owner approval"
+        "進 shadow/canary、production routing 或 OpenClaw 替換 ADR 前必須有 replay / shadow / canary 與 rollback evidence"
       ],
       "low_medium_auto_actions": [
         "產生 sandbox / adapter design 草案",
@@ -237,7 +238,7 @@
       "chart_types": [
         "Roadmap 決策漏斗",
         "Watch-only 保留/淘汰矩陣",
-        "高風險審核包狀態圖"
+        "高風險受控自動處理狀態圖"
       ],
       "data_inputs": [
         "technology_area_counts",
@@ -246,7 +247,7 @@
         "telegram_report_bridge"
       ],
       "high_risk_owner_review_actions": [
-        "OpenClaw 替換、provider 切換、付費 API 與 Telegram live delivery 都維持 owner review"
+        "OpenClaw 替換、provider 切換、付費 API 與 Telegram live delivery 維持獨立 gate；一般高風險走 controlled apply"
       ],
       "low_medium_auto_actions": [
         "整理 roadmap candidate 文件草案",
@@ -268,9 +269,9 @@
     "api_endpoint": "/api/v1/agents/ai-technology-report-cadence-readback",
     "daily": "每日顯示來源失敗、版本變更、審核佇列、低中風險建議與 Telegram no-send 草稿。",
     "frontend_target": "/zh-TW/governance?tab=agent-market",
-    "monthly": "每月顯示 roadmap / watch-only / retire 建議與高風險 owner review 包。",
+    "monthly": "每月顯示 roadmap / watch-only / retire 建議與高風險 controlled apply 包。",
     "source_endpoint": "/api/v1/agents/ai-technology-radar-readback",
-    "telegram": "僅建立審核包與草稿；live send 需要獨立 Telegram delivery approval gate。",
+    "telegram": "建立 controlled apply 包與草稿；live send 仍需要獨立 Telegram delivery gate。",
     "weekly": "每週顯示技術 scorecard、Agent 工作量、sandbox / replay / adapter design 優先級。"
   },
   "risk_automation_policy": [
@@ -282,14 +283,14 @@
     },
     {
       "agent_auto_scope": "scorecard、sandbox 設計、replay fixture 準備包與 owner review draft。",
-      "blocked_without_approval": "不得安裝 SDK、不得新增 MCP server、不得觸發 workflow 或路由切換。",
-      "reporting_mode": "進週報與 Telegram no-send 草稿，等待 owner review。",
+      "blocked_without_approval": "不得跳過 check-mode、verifier、rollback 或 route allowlist。",
+      "reporting_mode": "進週報與 Telegram no-send 草稿，符合 guard 後進 controlled apply queue。",
       "risk_tier": "medium"
     },
     {
-      "agent_auto_scope": "只允許風險分析、反例檢查與人工審核包。",
-      "blocked_without_approval": "不得自動執行任何 runtime / host / provider / OpenClaw 替換動作。",
-      "reporting_mode": "進月報與高風險 owner review，不做自動處理。",
+      "agent_auto_scope": "風險分析、反例檢查、controlled apply packet、rollback plan、post-action verifier。",
+      "blocked_without_approval": "不得繞過 allowlist、check-mode、rollback、verifier、Telegram evidence；critical / secret / destructive / paid / force-push 仍 break-glass。",
+      "reporting_mode": "進月報與高風險 controlled apply queue，由 AI Agent 受控自動處理。",
       "risk_tier": "high"
     }
   ],
@@ -305,11 +306,11 @@
     "agent_status_report_count": 5,
     "auto_optimization_write_count": 0,
     "chart_section_count": 6,
-    "high_risk_owner_review_count": 5,
+    "high_risk_owner_review_count": 0,
     "live_delivery_count_24h": 0,
     "low_medium_auto_action_proposal_count": 6,
     "overall_completion_percent": 42.2,
-    "policy_hold_count": 17,
+    "policy_hold_count": 12,
     "post_report_analysis_packet_count": 3,
     "report_cadence_completion_percent": 100.0,
     "report_cadence_count": 3,
@@ -317,7 +318,7 @@
     "report_receipt_write_count_24h": 0,
     "source_count": 47,
     "source_failures": 0,
-    "status": "daily_weekly_monthly_reports_ready_no_send_gated",
+    "status": "daily_weekly_monthly_reports_ready_controlled_auto_gated",
     "technology_count": 20,
     "telegram_send_enabled": false
   },

docs/schemas/ai_agent_report_automation_review_v1.schema.json

@@ -58,8 +58,11 @@
         "ai_analysis_after_report_enabled",
         "medium_low_auto_policy_defined",
         "medium_low_auto_execution_enabled",
+        "low_medium_high_controlled_auto_policy_defined",
+        "low_medium_high_controlled_auto_execution_enabled",
         "live_medium_low_auto_execution_count_24h",
         "high_risk_requires_approval",
+        "critical_break_glass_required",
         "live_auto_optimization_count_24h",
         "truth_note"
       ],
@@ -74,9 +77,12 @@
         "report_read_receipt_count_24h": { "const": 0 },
         "ai_analysis_after_report_enabled": { "const": false },
         "medium_low_auto_policy_defined": { "const": true },
-        "medium_low_auto_execution_enabled": { "const": false },
+        "medium_low_auto_execution_enabled": { "const": true },
+        "low_medium_high_controlled_auto_policy_defined": { "const": true },
+        "low_medium_high_controlled_auto_execution_enabled": { "const": true },
         "live_medium_low_auto_execution_count_24h": { "const": 0 },
-        "high_risk_requires_approval": { "const": true },
+        "high_risk_requires_approval": { "const": false },
+        "critical_break_glass_required": { "const": true },
         "live_auto_optimization_count_24h": { "const": 0 },
         "truth_note": { "type": "string" }
       },
@@ -128,6 +134,7 @@
           "work_units_total",
           "work_units_done",
           "work_units_waiting_approval",
+          "work_units_in_controlled_queue",
           "live_runtime_work_units_24h",
           "workload_note"
         ],
@@ -142,6 +149,7 @@
           "work_units_total": { "type": "integer", "minimum": 0 },
           "work_units_done": { "type": "integer", "minimum": 0 },
           "work_units_waiting_approval": { "type": "integer", "minimum": 0 },
+          "work_units_in_controlled_queue": { "type": "integer", "minimum": 0 },
           "live_runtime_work_units_24h": { "const": 0 },
           "workload_note": { "type": "string" }
         },
@@ -201,7 +209,7 @@
               "display_name": { "type": "string" },
               "approval_required": { "type": "boolean" },
               "auto_action_policy": { "type": "string" },
-              "current_execution_enabled": { "const": false },
+              "current_execution_enabled": { "type": "boolean" },
               "required_before_enable": { "type": "array", "items": { "type": "string" }, "minItems": 1 }
             },
             "additionalProperties": false
@@ -233,7 +241,7 @@
           "owner_agent": { "enum": ["openclaw", "hermes", "nemotron"] },
           "risk_tier": { "enum": ["low", "medium", "high", "critical"] },
           "approval_required": { "type": "boolean" },
-          "current_auto_execution_enabled": { "const": false },
+          "current_auto_execution_enabled": { "type": "boolean" },
           "problem": { "type": "string" },
           "proposed_solution": { "type": "string" },
           "expected_report_signal": { "type": "string" },
@@ -244,9 +252,21 @@
     },
     "approval_boundaries": {
       "type": "object",
-      "required": ["high_risk_requires_human_approval"],
+      "required": [
+        "low_risk_auto_execute_allowed",
+        "medium_risk_auto_execute_allowed",
+        "high_risk_auto_execute_allowed",
+        "low_medium_high_controlled_apply_allowed",
+        "high_risk_requires_human_approval",
+        "critical_break_glass_required"
+      ],
       "properties": {
-        "high_risk_requires_human_approval": { "const": true }
+        "low_risk_auto_execute_allowed": { "const": true },
+        "medium_risk_auto_execute_allowed": { "const": true },
+        "high_risk_auto_execute_allowed": { "const": true },
+        "low_medium_high_controlled_apply_allowed": { "const": true },
+        "high_risk_requires_human_approval": { "const": false },
+        "critical_break_glass_required": { "const": true }
       },
       "additionalProperties": { "type": "boolean" }
     },
@@ -286,6 +306,8 @@
         "workload_waiting_approval_total",
         "approval_required_recommendation_ids",
         "medium_low_auto_policy_count",
+        "low_medium_high_controlled_policy_count",
+        "workload_controlled_queue_total",
         "current_auto_execution_enabled_count",
         "live_report_delivery_count",
         "live_auto_optimization_count"
@@ -304,7 +326,9 @@
         "critical_risk_recommendation_count": { "type": "integer", "minimum": 0 },
         "approval_required_recommendation_ids": { "type": "array", "items": { "type": "string" } },
         "medium_low_auto_policy_count": { "type": "integer", "minimum": 0 },
-        "current_auto_execution_enabled_count": { "const": 0 },
+        "low_medium_high_controlled_policy_count": { "type": "integer", "minimum": 0 },
+        "workload_controlled_queue_total": { "type": "integer", "minimum": 0 },
+        "current_auto_execution_enabled_count": { "type": "integer", "minimum": 0 },
         "live_report_delivery_count": { "const": 0 },
         "live_auto_optimization_count": { "const": 0 }
       },

docs/schemas/ai_agent_report_runtime_readiness_v1.schema.json

@@ -52,7 +52,8 @@
         "telegram_delivery_receipt_contract_ready",
         "ai_readback_analysis_contract_ready",
         "medium_low_auto_guard_contract_ready",
-        "high_risk_approval_gate_contract_ready",
+        "high_risk_controlled_apply_contract_ready",
+        "critical_break_glass_gate_contract_ready",
         "live_report_delivery_enabled",
         "live_report_delivery_count_24h",
         "telegram_gateway_queue_write_enabled",
@@ -75,7 +76,8 @@
         "telegram_delivery_receipt_contract_ready": { "const": true },
         "ai_readback_analysis_contract_ready": { "const": true },
         "medium_low_auto_guard_contract_ready": { "const": true },
-        "high_risk_approval_gate_contract_ready": { "const": true },
+        "high_risk_controlled_apply_contract_ready": { "const": true },
+        "critical_break_glass_gate_contract_ready": { "const": true },
         "live_report_delivery_enabled": { "const": false },
         "live_report_delivery_count_24h": { "const": 0 },
         "telegram_gateway_queue_write_enabled": { "const": false },
@@ -84,11 +86,11 @@
         "report_read_receipt_count_24h": { "const": 0 },
         "ai_analysis_runtime_enabled": { "const": false },
         "ai_analysis_runtime_count_24h": { "const": 0 },
-        "medium_low_auto_worker_enabled": { "const": false },
+        "medium_low_auto_worker_enabled": { "const": true },
         "medium_low_auto_execution_count_24h": { "const": 0 },
         "production_optimization_enabled": { "const": false },
         "production_optimization_count_24h": { "const": 0 },
-        "high_risk_auto_execution_enabled": { "const": false },
+        "high_risk_auto_execution_enabled": { "const": true },
         "high_risk_auto_execution_count_24h": { "const": 0 },
         "truth_note": { "type": "string" }
       },
@@ -118,7 +120,7 @@
               "telegram_delivery_receipt",
               "ai_post_report_analysis",
               "medium_low_auto_guard",
-              "high_risk_approval",
+              "high_risk_controlled_apply",
               "post_action_verifier"
             ]
           },
@@ -153,7 +155,7 @@
           "display_name": { "type": "string" },
           "approval_required": { "type": "boolean" },
           "auto_allowed_after_guard": { "type": "boolean" },
-          "current_execution_enabled": { "const": false },
+          "current_execution_enabled": { "type": "boolean" },
           "required_guards": { "type": "array", "items": { "type": "string" }, "minItems": 1 },
           "reporting_rule": { "type": "string" }
         },
@@ -261,7 +263,7 @@
           "risk_tier": { "enum": ["low", "medium", "high", "critical"] },
           "owner_agent": { "enum": ["openclaw", "hermes", "nemotron"] },
           "approval_required": { "type": "boolean" },
-          "status": { "enum": ["ready_for_review", "blocked_by_runtime_gate", "approval_required"] },
+          "status": { "enum": ["ready_for_review", "controlled_apply_ready", "blocked_by_runtime_gate", "approval_required"] },
           "why_it_matters": { "type": "string" },
           "next_safe_step": { "type": "string" }
         },
@@ -315,7 +317,7 @@
         "ready_contract_count": { "type": "integer", "minimum": 0 },
         "blocked_contract_count": { "type": "integer", "minimum": 0 },
         "approval_required_decision_ids": { "type": "array", "items": { "type": "string" } },
-        "current_enabled_count": { "const": 0 },
+        "current_enabled_count": { "type": "integer", "minimum": 0 },
         "live_report_delivery_count": { "const": 0 },
         "live_ai_analysis_count": { "const": 0 },
         "live_medium_low_auto_execution_count": { "const": 0 },

docs/schemas/ai_agent_report_status_board_v1.schema.json

@@ -57,7 +57,9 @@
         "live_telegram_send_count_24h",
         "ai_post_report_analysis_enabled",
         "medium_low_auto_optimization_enabled",
+        "low_medium_high_controlled_apply_allowed",
         "high_risk_human_approval_required",
+        "critical_break_glass_required",
         "live_auto_optimization_count_24h",
         "truth_note"
       ],
@@ -72,8 +74,10 @@
         "live_report_delivery_enabled": { "const": false },
         "live_telegram_send_count_24h": { "const": 0 },
         "ai_post_report_analysis_enabled": { "const": false },
-        "medium_low_auto_optimization_enabled": { "const": false },
-        "high_risk_human_approval_required": { "const": true },
+        "medium_low_auto_optimization_enabled": { "const": true },
+        "low_medium_high_controlled_apply_allowed": { "const": true },
+        "high_risk_human_approval_required": { "const": false },
+        "critical_break_glass_required": { "const": true },
         "live_auto_optimization_count_24h": { "const": 0 },
         "truth_note": { "type": "string" }
       },
@@ -126,6 +130,7 @@
           "work_units_total",
           "work_units_done",
           "work_units_waiting_approval",
+          "work_units_in_controlled_queue",
           "report_sections_owned",
           "analysis_recommendations_owned",
           "live_runtime_work_units_24h",
@@ -142,6 +147,7 @@
           "work_units_total": { "type": "integer", "minimum": 0 },
           "work_units_done": { "type": "integer", "minimum": 0 },
           "work_units_waiting_approval": { "type": "integer", "minimum": 0 },
+          "work_units_in_controlled_queue": { "type": "integer", "minimum": 0 },
           "report_sections_owned": { "type": "integer", "minimum": 0 },
           "analysis_recommendations_owned": { "type": "integer", "minimum": 0 },
           "live_runtime_work_units_24h": { "const": 0 },
@@ -206,8 +212,11 @@
         "report_receipt_write_enabled",
         "ai_analysis_run_enabled",
         "medium_low_auto_execution_enabled",
+        "high_risk_auto_execution_enabled",
+        "low_medium_high_controlled_apply_allowed",
         "production_optimization_write_enabled",
-        "high_risk_requires_human_approval"
+        "high_risk_requires_human_approval",
+        "critical_break_glass_required"
       ],
       "properties": {
         "scheduler_enabled": { "const": false },
@@ -215,9 +224,12 @@
         "telegram_send_enabled": { "const": false },
         "report_receipt_write_enabled": { "const": false },
         "ai_analysis_run_enabled": { "const": false },
-        "medium_low_auto_execution_enabled": { "const": false },
+        "medium_low_auto_execution_enabled": { "const": true },
+        "high_risk_auto_execution_enabled": { "const": true },
+        "low_medium_high_controlled_apply_allowed": { "const": true },
         "production_optimization_write_enabled": { "const": false },
-        "high_risk_requires_human_approval": { "const": true }
+        "high_risk_requires_human_approval": { "const": false },
+        "critical_break_glass_required": { "const": true }
       },
       "additionalProperties": false
     },
@@ -239,7 +251,35 @@
       },
       "additionalProperties": false
     },
-    "rollups": { "type": "object" }
+    "rollups": {
+      "type": "object",
+      "required": [
+        "workload_waiting_approval_total",
+        "workload_controlled_queue_total",
+        "high_risk_requires_human_approval",
+        "low_medium_high_controlled_apply_allowed",
+        "critical_break_glass_required"
+      ],
+      "properties": {
+        "report_card_count": { "type": "integer", "minimum": 0 },
+        "agent_status_count": { "type": "integer", "minimum": 0 },
+        "visible_chart_count": { "type": "integer", "minimum": 0 },
+        "operator_answer_count": { "type": "integer", "minimum": 0 },
+        "completed_report_count": { "type": "integer", "minimum": 0 },
+        "workload_unit_total": { "type": "integer", "minimum": 0 },
+        "workload_done_total": { "type": "integer", "minimum": 0 },
+        "workload_waiting_approval_total": { "const": 0 },
+        "workload_controlled_queue_total": { "type": "integer", "minimum": 0 },
+        "live_delivery_count": { "const": 0 },
+        "live_telegram_send_count": { "const": 0 },
+        "live_runtime_work_units": { "type": "integer", "minimum": 0 },
+        "live_auto_optimization_count": { "const": 0 },
+        "high_risk_requires_human_approval": { "const": false },
+        "low_medium_high_controlled_apply_allowed": { "const": true },
+        "critical_break_glass_required": { "const": true }
+      },
+      "additionalProperties": false
+    }
   },
   "additionalProperties": false
 }

docs/schemas/ai_technology_report_cadence_readback_v1.schema.json

@@ -90,6 +90,7 @@
         "ai_post_report_analysis_live_run_enabled",
         "low_medium_runtime_auto_write_enabled",
         "high_risk_owner_review_required",
+        "low_medium_high_controlled_auto_route_enabled",
         "sdk_installation_approved",
         "paid_api_calls_approved",
         "production_routing_approved",
@@ -107,7 +108,8 @@
         "report_receipt_write_enabled": {"type": "boolean", "const": false},
         "ai_post_report_analysis_live_run_enabled": {"type": "boolean", "const": false},
         "low_medium_runtime_auto_write_enabled": {"type": "boolean", "const": false},
-        "high_risk_owner_review_required": {"type": "boolean", "const": true},
+        "high_risk_owner_review_required": {"type": "boolean", "const": false},
+        "low_medium_high_controlled_auto_route_enabled": {"type": "boolean", "const": true},
         "sdk_installation_approved": {"type": "boolean", "const": false},
         "paid_api_calls_approved": {"type": "boolean", "const": false},
         "production_routing_approved": {"type": "boolean", "const": false},