diff --git a/apps/api/src/services/ai_agent_professional_task_expansion.py b/apps/api/src/services/ai_agent_professional_task_expansion.py index da658fcc6..696cf9f55 100644 --- a/apps/api/src/services/ai_agent_professional_task_expansion.py +++ b/apps/api/src/services/ai_agent_professional_task_expansion.py @@ -1,10 +1,10 @@ """ AI Agent professional task expansion and Telegram runtime bridge snapshot. -Loads the latest committed P2-405A read-only contract. The contract expands -professional AI Agent work and defines Telegram bridge stages, but it does not -write Telegram Gateway queues, send Telegram messages, call the Bot API, read -secrets, or execute production changes. +Loads the latest committed P2-405B read-only contract. The contract expands +professional AI Agent work and defines Telegram no-send previews, but it does +not write Telegram Gateway queues, send Telegram messages, call the Bot API, +read secrets, or execute production changes. """ from __future__ import annotations @@ -24,6 +24,10 @@ _EXPECTED_TASK_COUNT = 24 _EXPECTED_DOMAIN_COUNT = 8 _EXPECTED_STAGE_COUNT = 5 _EXPECTED_MESSAGE_TYPE_COUNT = 6 +_EXPECTED_NO_SEND_PREVIEW_COUNT = 6 +_EXPECTED_DEDUP_KEY_COUNT = 6 +_EXPECTED_RECEIPT_EXPECTATION_COUNT = 6 +_EXPECTED_CANARY_PACKAGE_COUNT = 1 _ZERO_ROLLUP_FIELDS = { "current_live_count", "gateway_queue_write_count", @@ -35,6 +39,11 @@ _ZERO_ROLLUP_FIELDS = { "paid_api_call_count", "host_write_count", "kubectl_action_count", + "preview_send_enabled_count", + "preview_queue_write_enabled_count", + "preview_bot_api_call_enabled_count", + "receipt_live_write_enabled_count", + "canary_live_send_enabled_count", } _FORBIDDEN_PUBLIC_TERMS = { "work_window_transcript", @@ -82,11 +91,11 @@ def _require_schema(payload: dict[str, Any], label: str) -> None: status = payload.get("program_status") or {} expected = { "current_priority": "P2", - "current_task_id": "P2-405A", - "next_task_id": "P2-405B", + "current_task_id": "P2-405B", + "next_task_id": "P2-405C", "read_only_mode": True, "runtime_authority": _RUNTIME_AUTHORITY, - "overall_completion_percent": 82, + "overall_completion_percent": 88, } mismatches = _mismatches(status, expected) if mismatches: @@ -124,6 +133,105 @@ def _require_telegram_bridge(payload: dict[str, Any], label: str) -> None: if len(message_types) != _EXPECTED_MESSAGE_TYPE_COUNT: raise ValueError(f"{label}: expected {_EXPECTED_MESSAGE_TYPE_COUNT} message types") + _require_no_send_previews(bridge, label) + _require_receipt_and_canary_package(bridge, label) + + +def _require_no_send_previews(bridge: dict[str, Any], label: str) -> None: + previews = bridge.get("no_send_message_previews") or [] + if len(previews) != _EXPECTED_NO_SEND_PREVIEW_COUNT: + raise ValueError(f"{label}: expected {_EXPECTED_NO_SEND_PREVIEW_COUNT} no-send previews") + + message_types = {item.get("message_type") for item in bridge.get("message_types") or []} + preview_message_types = [preview.get("message_type") for preview in previews] + if set(preview_message_types) != message_types: + raise ValueError(f"{label}: no-send previews must cover every message type exactly once") + + dedup_keys = [preview.get("dedup_key") for preview in previews] + if len(set(dedup_keys)) != len(dedup_keys): + raise ValueError(f"{label}: no-send preview dedup_key values must be unique") + + for preview in previews: + preview_id = preview.get("preview_id") + if preview.get("status") != "preview_ready_no_send": + raise ValueError(f"{label}: {preview_id}.status must be preview_ready_no_send") + if preview.get("send_enabled") is not False: + raise ValueError(f"{label}: {preview_id}.send_enabled must remain false") + if preview.get("gateway_queue_write_enabled") is not False: + raise ValueError(f"{label}: {preview_id}.gateway_queue_write_enabled must remain false") + if preview.get("bot_api_call_enabled") is not False: + raise ValueError(f"{label}: {preview_id}.bot_api_call_enabled must remain false") + if preview.get("delivery_receipt_write_enabled") is not False: + raise ValueError(f"{label}: {preview_id}.delivery_receipt_write_enabled must remain false") + if not preview.get("sanitized_body_lines"): + raise ValueError(f"{label}: {preview_id}.sanitized_body_lines must not be empty") + + dedup_policy = bridge.get("dedup_policy") or {} + keys = dedup_policy.get("keys") or [] + if dedup_policy.get("required") is not True: + raise ValueError(f"{label}: dedup_policy.required must be true") + if dedup_policy.get("live_cache_write_enabled") is not False: + raise ValueError(f"{label}: dedup_policy.live_cache_write_enabled must remain false") + if len(keys) != _EXPECTED_DEDUP_KEY_COUNT: + raise ValueError(f"{label}: expected {_EXPECTED_DEDUP_KEY_COUNT} dedup keys") + + queue_preview = bridge.get("queue_preview_readback") or {} + expected_queue = { + "enabled": True, + "preview_only": True, + "write_enabled": False, + "readback_enabled": True, + } + mismatches = _mismatches(queue_preview, expected_queue) + if mismatches: + raise ValueError(f"{label}: queue_preview_readback mismatch: {mismatches}") + + +def _require_receipt_and_canary_package(bridge: dict[str, Any], label: str) -> None: + previews = bridge.get("no_send_message_previews") or [] + preview_ids = {preview.get("preview_id") for preview in previews} + preview_receipts = {preview.get("receipt_expectation_id") for preview in previews} + + receipts = bridge.get("receipt_expectations") or [] + if len(receipts) != _EXPECTED_RECEIPT_EXPECTATION_COUNT: + raise ValueError(f"{label}: expected {_EXPECTED_RECEIPT_EXPECTATION_COUNT} receipt expectations") + + receipt_ids = {receipt.get("receipt_id") for receipt in receipts} + if receipt_ids != preview_receipts: + raise ValueError(f"{label}: receipt expectations must match preview receipt ids") + + for receipt in receipts: + receipt_id = receipt.get("receipt_id") + if receipt.get("preview_id") not in preview_ids: + raise ValueError(f"{label}: {receipt_id}.preview_id must reference a no-send preview") + if receipt.get("receipt_write_enabled") is not False: + raise ValueError(f"{label}: {receipt_id}.receipt_write_enabled must remain false") + if receipt.get("production_receipt_readback_enabled") is not False: + raise ValueError( + f"{label}: {receipt_id}.production_receipt_readback_enabled must remain false" + ) + if not receipt.get("required_evidence_refs"): + raise ValueError(f"{label}: {receipt_id}.required_evidence_refs must not be empty") + + canary = bridge.get("canary_approval_package") or {} + expected_canary = { + "package_ready": True, + "approval_required": True, + "status": "blocked_until_explicit_approval", + "live_send_enabled": False, + "gateway_queue_write_enabled": False, + "bot_api_call_enabled": False, + "delivery_receipt_write_enabled": False, + "production_write_enabled": False, + } + mismatches = _mismatches(canary, expected_canary) + if mismatches: + raise ValueError(f"{label}: canary_approval_package mismatch: {mismatches}") + if not canary: + raise ValueError(f"{label}: expected {_EXPECTED_CANARY_PACKAGE_COUNT} canary package") + if not canary.get("approval_checklist"): + raise ValueError(f"{label}: canary_approval_package.approval_checklist is required") + def _require_professional_tasks(payload: dict[str, Any], label: str) -> None: domains = payload.get("professional_task_domains") or [] @@ -212,6 +320,12 @@ def _require_rollups(payload: dict[str, Any], label: str) -> None: "medium_risk_task_count": sum(1 for task in tasks if task.get("risk_tier") == "medium"), "high_risk_task_count": sum(1 for task in tasks if task.get("risk_tier") == "high"), "critical_risk_task_count": sum(1 for task in tasks if task.get("risk_tier") == "critical"), + "no_send_preview_count": len(bridge.get("no_send_message_previews") or []), + "dedup_key_count": len((bridge.get("dedup_policy") or {}).get("keys") or []), + "receipt_expectation_count": len(bridge.get("receipt_expectations") or []), + "canary_approval_package_count": 1 + if bridge.get("canary_approval_package") + else 0, } mismatches = _mismatches(rollups, expected) if mismatches: diff --git a/apps/api/tests/test_ai_agent_professional_task_expansion.py b/apps/api/tests/test_ai_agent_professional_task_expansion.py index d20c6430a..52d4656e1 100644 --- a/apps/api/tests/test_ai_agent_professional_task_expansion.py +++ b/apps/api/tests/test_ai_agent_professional_task_expansion.py @@ -18,9 +18,9 @@ def test_load_latest_ai_agent_professional_task_expansion_snapshot() -> None: snapshot = load_latest_ai_agent_professional_task_expansion() assert snapshot["schema_version"] == "ai_agent_professional_task_expansion_v1" - assert snapshot["program_status"]["current_task_id"] == "P2-405A" - assert snapshot["program_status"]["next_task_id"] == "P2-405B" - assert snapshot["program_status"]["overall_completion_percent"] == 82 + assert snapshot["program_status"]["current_task_id"] == "P2-405B" + assert snapshot["program_status"]["next_task_id"] == "P2-405C" + assert snapshot["program_status"]["overall_completion_percent"] == 88 assert snapshot["program_status"]["runtime_authority"] == ( "professional_task_expansion_and_telegram_bridge_read_only_no_send" ) @@ -36,6 +36,12 @@ def test_load_latest_ai_agent_professional_task_expansion_snapshot() -> None: assert bridge["telegram_send_enabled"] is False assert len(bridge["stages"]) == 5 assert len(bridge["message_types"]) == 6 + assert len(bridge["no_send_message_previews"]) == 6 + assert len(bridge["dedup_policy"]["keys"]) == 6 + assert len(bridge["receipt_expectations"]) == 6 + assert bridge["queue_preview_readback"]["write_enabled"] is False + assert bridge["canary_approval_package"]["status"] == "blocked_until_explicit_approval" + assert bridge["canary_approval_package"]["live_send_enabled"] is False rollups = snapshot["rollups"] assert rollups["professional_task_count"] == 24 @@ -57,6 +63,15 @@ def test_load_latest_ai_agent_professional_task_expansion_snapshot() -> None: assert rollups["paid_api_call_count"] == 0 assert rollups["host_write_count"] == 0 assert rollups["kubectl_action_count"] == 0 + assert rollups["no_send_preview_count"] == 6 + assert rollups["dedup_key_count"] == 6 + assert rollups["receipt_expectation_count"] == 6 + assert rollups["canary_approval_package_count"] == 1 + assert rollups["preview_send_enabled_count"] == 0 + assert rollups["preview_queue_write_enabled_count"] == 0 + assert rollups["preview_bot_api_call_enabled_count"] == 0 + assert rollups["receipt_live_write_enabled_count"] == 0 + assert rollups["canary_live_send_enabled_count"] == 0 def test_professional_tasks_cover_required_agents_and_reporting() -> None: @@ -81,6 +96,54 @@ def test_professional_tasks_cover_required_agents_and_reporting() -> None: assert snapshot["redaction_contract"]["raw_prompt_display_allowed"] is False assert snapshot["redaction_contract"]["private_reasoning_display_allowed"] is False assert snapshot["redaction_contract"]["secret_value_display_allowed"] is False + assert snapshot["redaction_contract"]["frontend_display_policy"] + + +def test_no_send_previews_have_unique_dedup_and_no_live_flags() -> None: + snapshot = load_latest_ai_agent_professional_task_expansion() + bridge = snapshot["telegram_runtime_bridge"] + + message_types = {item["message_type"] for item in bridge["message_types"]} + previews = bridge["no_send_message_previews"] + receipts = bridge["receipt_expectations"] + + assert {preview["message_type"] for preview in previews} == message_types + assert len({preview["dedup_key"] for preview in previews}) == 6 + assert {preview["receipt_expectation_id"] for preview in previews} == { + receipt["receipt_id"] for receipt in receipts + } + + for preview in previews: + assert preview["status"] == "preview_ready_no_send" + assert preview["send_enabled"] is False + assert preview["gateway_queue_write_enabled"] is False + assert preview["bot_api_call_enabled"] is False + assert preview["delivery_receipt_write_enabled"] is False + assert preview["sanitized_body_lines"] + + assert bridge["dedup_policy"]["required"] is True + assert bridge["dedup_policy"]["live_cache_write_enabled"] is False + assert bridge["queue_preview_readback"]["preview_only"] is True + assert bridge["queue_preview_readback"]["write_enabled"] is False + + +def test_receipts_and_canary_package_remain_no_send() -> None: + snapshot = load_latest_ai_agent_professional_task_expansion() + bridge = snapshot["telegram_runtime_bridge"] + + for receipt in bridge["receipt_expectations"]: + assert receipt["receipt_write_enabled"] is False + assert receipt["production_receipt_readback_enabled"] is False + assert receipt["required_evidence_refs"] + + canary = bridge["canary_approval_package"] + assert canary["package_ready"] is True + assert canary["approval_required"] is True + assert canary["live_send_enabled"] is False + assert canary["gateway_queue_write_enabled"] is False + assert canary["bot_api_call_enabled"] is False + assert canary["delivery_receipt_write_enabled"] is False + assert canary["production_write_enabled"] is False def test_rejects_telegram_send_enabled(tmp_path: Path) -> None: @@ -101,6 +164,36 @@ def test_rejects_gateway_queue_write_count(tmp_path: Path) -> None: load_latest_ai_agent_professional_task_expansion(tmp_path) +def test_rejects_preview_send_enabled(tmp_path: Path) -> None: + snapshot = copy.deepcopy(load_latest_ai_agent_professional_task_expansion()) + snapshot["telegram_runtime_bridge"]["no_send_message_previews"][0]["send_enabled"] = True + snapshot["rollups"]["preview_send_enabled_count"] = 1 + _write_snapshot(tmp_path, snapshot) + + with pytest.raises(ValueError, match="send_enabled must remain false"): + load_latest_ai_agent_professional_task_expansion(tmp_path) + + +def test_rejects_duplicate_dedup_key(tmp_path: Path) -> None: + snapshot = copy.deepcopy(load_latest_ai_agent_professional_task_expansion()) + previews = snapshot["telegram_runtime_bridge"]["no_send_message_previews"] + previews[1]["dedup_key"] = previews[0]["dedup_key"] + _write_snapshot(tmp_path, snapshot) + + with pytest.raises(ValueError, match="dedup_key values must be unique"): + load_latest_ai_agent_professional_task_expansion(tmp_path) + + +def test_rejects_canary_live_send_enabled(tmp_path: Path) -> None: + snapshot = copy.deepcopy(load_latest_ai_agent_professional_task_expansion()) + snapshot["telegram_runtime_bridge"]["canary_approval_package"]["live_send_enabled"] = True + snapshot["rollups"]["canary_live_send_enabled_count"] = 1 + _write_snapshot(tmp_path, snapshot) + + with pytest.raises(ValueError, match="canary_approval_package mismatch"): + load_latest_ai_agent_professional_task_expansion(tmp_path) + + def test_rejects_high_risk_without_approval(tmp_path: Path) -> None: snapshot = copy.deepcopy(load_latest_ai_agent_professional_task_expansion()) high_task = next(task for task in snapshot["professional_tasks"] if task["risk_tier"] == "high") diff --git a/apps/api/tests/test_ai_agent_professional_task_expansion_api.py b/apps/api/tests/test_ai_agent_professional_task_expansion_api.py index 28d1fa2f2..b2c9eccbb 100644 --- a/apps/api/tests/test_ai_agent_professional_task_expansion_api.py +++ b/apps/api/tests/test_ai_agent_professional_task_expansion_api.py @@ -17,9 +17,9 @@ def test_ai_agent_professional_task_expansion_endpoint() -> None: assert response.status_code == 200 payload = response.json() assert payload["schema_version"] == "ai_agent_professional_task_expansion_v1" - assert payload["program_status"]["current_task_id"] == "P2-405A" - assert payload["program_status"]["next_task_id"] == "P2-405B" - assert payload["program_status"]["overall_completion_percent"] == 82 + assert payload["program_status"]["current_task_id"] == "P2-405B" + assert payload["program_status"]["next_task_id"] == "P2-405C" + assert payload["program_status"]["overall_completion_percent"] == 88 assert payload["program_status"]["runtime_authority"] == ( "professional_task_expansion_and_telegram_bridge_read_only_no_send" ) @@ -37,3 +37,11 @@ def test_ai_agent_professional_task_expansion_endpoint() -> None: assert payload["rollups"]["telegram_send_count"] == 0 assert payload["rollups"]["bot_api_call_count"] == 0 assert payload["rollups"]["production_write_count"] == 0 + assert payload["rollups"]["no_send_preview_count"] == 6 + assert payload["rollups"]["dedup_key_count"] == 6 + assert payload["rollups"]["receipt_expectation_count"] == 6 + assert payload["rollups"]["canary_approval_package_count"] == 1 + assert payload["rollups"]["preview_send_enabled_count"] == 0 + assert payload["telegram_runtime_bridge"]["canary_approval_package"]["live_send_enabled"] is False + assert len(payload["telegram_runtime_bridge"]["no_send_message_previews"]) == 6 + assert len(payload["telegram_runtime_bridge"]["receipt_expectations"]) == 6 diff --git a/apps/web/messages/en.json b/apps/web/messages/en.json index 48c5bf1ee..94555a122 100644 --- a/apps/web/messages/en.json +++ b/apps/web/messages/en.json @@ -6448,7 +6448,7 @@ } }, "professionalTaskExpansion": { - "title": "P2-405A AI Agent 專業任務擴展", + "title": "P2-405B AI Agent 專業任務與 TG 預覽", "source": "產生 {generated}；目前 {current}；下一步 {next}", "runtime": "runtime={value}", "telegramTitle": "Telegram Runtime Bridge", @@ -6462,8 +6462,14 @@ "messageTypes": "訊息類型", "approvals": "需批准", "highCritical": "高風險 / critical", - "liveWrites": "live / send / write" + "liveWrites": "live / send / write", + "noSendPreviews": "no-send 預覽", + "dedupKeys": "dedup key", + "receipts": "receipt expectation", + "previewLiveWrites": "preview 實發" }, + "previewTitle": "Telegram no-send 訊息預覽", + "canaryTitle": "Canary approval package", "labels": { "gateway": "Gateway required={value}", "noSend": "no-send preview={value}", @@ -6472,10 +6478,19 @@ "send": "Telegram send={value}", "redaction": "脫敏鎖定={value}", "gatewayWrites": "Gateway 寫入={value}", + "queueWrites": "queue 寫入={value}", "botCalls": "Bot API={value}", "productionWrites": "正式寫入={value}", "approval": "需批准={value}", - "liveCount": "live={value}" + "liveCount": "live={value}", + "dedup": "dedup={value}", + "receipt": "receipt={value}", + "previewSend": "preview send={value}", + "packageReady": "package ready={value}", + "canaryApproval": "canary approval={value}", + "arbiter": "仲裁={value}", + "owner": "owner={value}", + "canarySend": "canary send={value}" }, "riskTiers": { "low": "低風險", diff --git a/apps/web/messages/zh-TW.json b/apps/web/messages/zh-TW.json index 48c5bf1ee..94555a122 100644 --- a/apps/web/messages/zh-TW.json +++ b/apps/web/messages/zh-TW.json @@ -6448,7 +6448,7 @@ } }, "professionalTaskExpansion": { - "title": "P2-405A AI Agent 專業任務擴展", + "title": "P2-405B AI Agent 專業任務與 TG 預覽", "source": "產生 {generated}；目前 {current}；下一步 {next}", "runtime": "runtime={value}", "telegramTitle": "Telegram Runtime Bridge", @@ -6462,8 +6462,14 @@ "messageTypes": "訊息類型", "approvals": "需批准", "highCritical": "高風險 / critical", - "liveWrites": "live / send / write" + "liveWrites": "live / send / write", + "noSendPreviews": "no-send 預覽", + "dedupKeys": "dedup key", + "receipts": "receipt expectation", + "previewLiveWrites": "preview 實發" }, + "previewTitle": "Telegram no-send 訊息預覽", + "canaryTitle": "Canary approval package", "labels": { "gateway": "Gateway required={value}", "noSend": "no-send preview={value}", @@ -6472,10 +6478,19 @@ "send": "Telegram send={value}", "redaction": "脫敏鎖定={value}", "gatewayWrites": "Gateway 寫入={value}", + "queueWrites": "queue 寫入={value}", "botCalls": "Bot API={value}", "productionWrites": "正式寫入={value}", "approval": "需批准={value}", - "liveCount": "live={value}" + "liveCount": "live={value}", + "dedup": "dedup={value}", + "receipt": "receipt={value}", + "previewSend": "preview send={value}", + "packageReady": "package ready={value}", + "canaryApproval": "canary approval={value}", + "arbiter": "仲裁={value}", + "owner": "owner={value}", + "canarySend": "canary send={value}" }, "riskTiers": { "low": "低風險", diff --git a/apps/web/src/app/[locale]/governance/tabs/automation-inventory-tab.tsx b/apps/web/src/app/[locale]/governance/tabs/automation-inventory-tab.tsx index 5d0e96490..cf9ef358f 100644 --- a/apps/web/src/app/[locale]/governance/tabs/automation-inventory-tab.tsx +++ b/apps/web/src/app/[locale]/governance/tabs/automation-inventory-tab.tsx @@ -17,13 +17,16 @@ import { BookOpenCheck, Boxes, CalendarClock, + ClipboardCheck, Database, FileText, + Fingerprint, GitBranch, Gauge, HardDrive, Layers3, Lock, + MessageSquareText, PackageCheck, RefreshCw, Route, @@ -3700,6 +3703,15 @@ export function AutomationInventoryTab() { && professionalTaskExpansion.redaction_contract.redaction_required === true ) const visibleProfessionalTasks = professionalTaskExpansion.professional_tasks.slice(0, 8) + const visibleTelegramPreviews = professionalTaskExpansion.telegram_runtime_bridge.no_send_message_previews.slice(0, 6) + const professionalTaskCanaryPackage = professionalTaskExpansion.telegram_runtime_bridge.canary_approval_package + const professionalTaskPreviewLiveWrites = ( + professionalTaskExpansion.rollups.preview_send_enabled_count + + professionalTaskExpansion.rollups.preview_queue_write_enabled_count + + professionalTaskExpansion.rollups.preview_bot_api_call_enabled_count + + professionalTaskExpansion.rollups.receipt_live_write_enabled_count + + professionalTaskExpansion.rollups.canary_live_send_enabled_count + ) const backlogProgressPercent = backlog.progress_summary.overall_percent const explicitApprovalItemCount = backlog.item_approval_boundary_rollup.items_requiring_explicit_approval.length const taskBoundaryCount = snapshot.task_approval_boundary_rollup.total_tasks @@ -4121,6 +4133,10 @@ export function AutomationInventoryTab() { } /> } /> } /> + } /> + } /> + } /> + } />
@@ -4152,6 +4168,61 @@ export function AutomationInventoryTab() {
+ {t('professionalTaskExpansion.previewTitle')} +
+ {visibleTelegramPreviews.map(preview => { + const previewTone = preview.risk_tier === 'critical' ? 'danger' : preview.risk_tier === 'high' ? 'warn' : 'neutral' + const previewColor = toneColor(previewTone) + return ( +
+
+ + {preview.title} + + +
+
+ {preview.sanitized_body_lines.slice(0, 3).map(line => ( + + {redactPublicText(line)} + + ))} +
+
+ + + + + + +
+
+ ) + })} +
+ +
+ {t('professionalTaskExpansion.canaryTitle')} +

+ {professionalTaskCanaryPackage.package_id} · {professionalTaskCanaryPackage.status} +

+
+ + + + + + +
+
+ {professionalTaskCanaryPackage.approval_checklist.slice(0, 5).map(item => ( +
+ {item} +
+ ))} +
+
+ {t('professionalTaskExpansion.tasksTitle')}
{visibleProfessionalTasks.map(task => { diff --git a/apps/web/src/lib/api-client.ts b/apps/web/src/lib/api-client.ts index e3bd66817..a23553752 100644 --- a/apps/web/src/lib/api-client.ts +++ b/apps/web/src/lib/api-client.ts @@ -1429,8 +1429,8 @@ export interface AiAgentProfessionalTaskExpansionSnapshot { program_status: { overall_completion_percent: number current_priority: 'P0' | 'P1' | 'P2' | 'P3' - current_task_id: 'P2-405A' - next_task_id: 'P2-405B' + current_task_id: 'P2-405B' + next_task_id: 'P2-405C' read_only_mode: true runtime_authority: 'professional_task_expansion_and_telegram_bridge_read_only_no_send' status_note: string @@ -1461,19 +1461,91 @@ export interface AiAgentProfessionalTaskExpansionSnapshot { exit_condition: string }> message_types: Array<{ - message_type_id: string - title: string + message_type: string cadence: string - risk_scope: string - delivery_gate: string + risk_tier: string + owner_agent: string + send_policy: string }> + no_send_message_previews: Array<{ + preview_id: string + message_type: string + title: string + target_room_alias: string + target_room_env: string + owner_agent: string + risk_tier: 'low' | 'medium' | 'high' | 'critical' + status: string + dedup_key: string + receipt_expectation_id: string + approval_required_before_send: boolean + canary_required_before_live: boolean + send_enabled: boolean + gateway_queue_write_enabled: boolean + bot_api_call_enabled: boolean + delivery_receipt_write_enabled: boolean + sanitized_body_lines: string[] + redaction_checks: string[] + }> + dedup_policy: { + required: boolean + key_count: number + collision_policy: string + preview_hash_algorithm: string + live_cache_write_enabled: boolean + keys: Array<{ + message_type: string + key_template: string + owner_agent: string + risk_tier: string + window: string + live_cache_write_enabled: boolean + }> + } + queue_preview_readback: { + enabled: boolean + preview_only: boolean + write_enabled: boolean + readback_enabled: boolean + envelope_fields: string[] + readback_checks: string[] + } + receipt_expectations: Array<{ + receipt_id: string + message_type: string + preview_id: string + expected_signal: string + required_evidence_refs: string[] + receipt_write_enabled: boolean + production_receipt_readback_enabled: boolean + missing_receipt_policy: string + }> + canary_approval_package: { + package_id: string + status: string + package_ready: boolean + approval_required: boolean + target_room_alias: string + target_room_env: string + arbiter: string + owner_agent: string + reviewers: string[] + message_type_scope: string[] + required_approval_inputs: string[] + approval_checklist: string[] + live_send_enabled: boolean + gateway_queue_write_enabled: boolean + bot_api_call_enabled: boolean + delivery_receipt_write_enabled: boolean + production_write_enabled: boolean + } + no_send_preview_completion_percent: number + canary_approval_package_completion_percent: number } professional_task_domains: Array<{ domain_id: string - title: string - summary: string - primary_owner: string - task_count: number + label: string + owner_agent: string }> professional_tasks: Array<{ task_id: string @@ -1491,8 +1563,8 @@ export interface AiAgentProfessionalTaskExpansionSnapshot { }> reporting_contract: Record<'daily' | 'weekly' | 'monthly' | 'action_required', { required: boolean - delivery_mode: string - sections: string[] + owner_agent: string + telegram_stage: string }> redaction_contract: { redaction_required: boolean @@ -1503,6 +1575,7 @@ export interface AiAgentProfessionalTaskExpansionSnapshot { raw_runtime_payload_display_allowed: boolean telegram_message_must_be_sanitized: boolean frontend_display_policy: string + message_preview_redaction_checks: string[] } rollups: { professional_task_count: number @@ -1524,6 +1597,18 @@ export interface AiAgentProfessionalTaskExpansionSnapshot { paid_api_call_count: number host_write_count: number kubectl_action_count: number + no_send_preview_count: number + dedup_key_count: number + receipt_expectation_count: number + canary_approval_package_count: number + queue_preview_envelope_field_count: number + canary_required_preview_count: number + approval_required_preview_count: number + preview_send_enabled_count: number + preview_queue_write_enabled_count: number + preview_bot_api_call_enabled_count: number + receipt_live_write_enabled_count: number + canary_live_send_enabled_count: number } } diff --git a/docs/LOGBOOK.md b/docs/LOGBOOK.md index 75282ebaf..e24e79930 100644 --- a/docs/LOGBOOK.md +++ b/docs/LOGBOOK.md @@ -44,6 +44,35 @@ - owner response received / accepted、live conf、rendered diff evidence、`nginx -t` evidence、route smoke evidence、maintenance window、rollback owner、Nginx reload、host write、firewall change、Docker restart、active scan、secret collection、runtime execution、action button 全部維持 `0 / false`。 - 本輪未 SSH、未讀 live conf raw payload、未改 Nginx、未重啟 Docker / service、未修改 firewall / iptables、未執行 active scan、未收 secrets 明文、未 force push。 +## 2026-06-15｜P2-405B Telegram no-send 預覽、dedup、receipt 與 canary 批准包 + +**背景**：P2-405A 已把 AI Agent 專業任務擴展與 Telegram Runtime Bridge 固定成只讀契約，並已部署到正式 API。下一步需要讓統帥在治理頁實際看到 AI Agent 將來要送到 Telegram 的訊息長相、去重鍵、回執預期與 canary 前置審核包，但仍不得實發、不得寫 Gateway queue、不得呼叫 Bot API、不得讀 secret。 + +**完成項目**： +- 新增 `docs/evaluations/ai_agent_professional_task_expansion_2026-06-15_1445_p2_405b.json`，把 P2-405B 固定為最新 snapshot。 +- `ai_agent_professional_task_expansion_v1` schema 與 API loader 已升級：目前 `current_task_id=P2-405B`、`next_task_id=P2-405C`、overall `88%`。 +- Telegram no-send preview 已建立 `6` 種訊息：日報、週報、月報、高風險審核卡、failure-only escalation、report receipt gap alert。 +- dedup policy 已建立 `6` 個 key template；receipt expectation 已建立 `6` 組；canary approval package 已建立 `1` 份。 +- `/zh-TW/governance?tab=automation-inventory` 已顯示 P2-405B 卡片：no-send preview、dedup key、receipt expectation、canary approval package、preview 實發計數與所有 live / send / write = `0`。 +- 前端 i18n 維持繁體中文；沒有把工作視窗對話、未遮罩提示、私有推理或未遮罩 runtime payload 放入前端。 + +**完成度與狀態**： +- P2-405B Telegram no-send preview / dedup / receipt / canary package：`100%`。 +- AI Agent 專業任務擴展整體：`82% -> 88%`。 +- Telegram preview 可視化：`0% -> 100%`。 +- Telegram 實發、Gateway queue 寫入、Bot API call、delivery receipt production write：仍維持 `0%` / `0` / `false`。 +- 下一步：P2-405C 只允許產生 approved canary send approval packet 與操作者批准欄位；不得在未批准前真正發送 Telegram。 + +**本地驗證**： +- `python3 -m json.tool docs/evaluations/ai_agent_professional_task_expansion_2026-06-15_1445_p2_405b.json` 通過。 +- `python3 -m json.tool docs/schemas/ai_agent_professional_task_expansion_v1.schema.json` 通過。 +- `python3 -m json.tool apps/web/messages/zh-TW.json` 與 `python3 -m json.tool apps/web/messages/en.json` 通過。 +- `python3 -m py_compile apps/api/src/services/ai_agent_professional_task_expansion.py apps/api/src/api/v1/agents.py` 通過。 +- `DATABASE_URL=postgresql+asyncpg://test:test@localhost/test pytest -q apps/api/tests/test_ai_agent_professional_task_expansion.py apps/api/tests/test_ai_agent_professional_task_expansion_api.py` → `12 passed`。 +- `pnpm --filter @awoooi/web typecheck` 通過。 + +**邊界**：本輪未 SSH、未讀 live conf、未改主機、未修改 Telegram Bot、未發 Telegram、未寫 Gateway queue、未呼叫 Bot API、未讀 secret、未啟用 paid API、未跑 kubectl、未做 production write。 + ## 2026-06-15｜端口 / 防火牆事故型 Gate 與前台手機溢出收斂 **背景**：2026-06-12 14:00 左右的 110 端口關閉事件暴露「端口 / 防火牆變更」不能只用一般配置變更欄位驗收，必須要求事故等級、服務健康影響、外部 route 影響、恢復時間、操作員通知與事後補件 owner。使用者也指出前台不得揭露工作視窗對話、原始 owner / repo namespace、raw blocked 狀態或內部英文識別資訊；因此本階段同步做 production 前台敏感資訊與手機溢出驗證。 diff --git a/docs/ai/AI_AGENT_12_AGENT_WAR_ROOM_2026-06-14.md b/docs/ai/AI_AGENT_12_AGENT_WAR_ROOM_2026-06-14.md index 8f835d17c..9f6cf4c9a 100644 --- a/docs/ai/AI_AGENT_12_AGENT_WAR_ROOM_2026-06-14.md +++ b/docs/ai/AI_AGENT_12_AGENT_WAR_ROOM_2026-06-14.md @@ -184,5 +184,5 @@ Telegram Ops 工位負責 Telegram Gateway、Channel Hub、日報/週報/月報 狀態變更：12 位 Agent 只讀審查全部回收並彙整；已建立 `ai_agent_12_agent_war_room_v1` schema / committed snapshot / API / tests / governance UI 區塊。 證據：`docs/schemas/ai_agent_12_agent_war_room_v1.schema.json`、`docs/evaluations/ai_agent_12_agent_war_room_2026-06-14.json`、`GET /api/v1/agents/agent-12-agent-war-room`、治理頁 12-Agent War Room 區塊、12 位 Agent 只讀回饋。 阻擋：runtime writer、Telegram send、Bot API、production write、SDK 安裝、付費 API、shadow/canary、host update、DB migration、restore 仍未批准。 -下一步：P2-405B 承接 Telegram no-send preview UI；讓統帥先看見 TG 訊息、dedup、receipt 與風險標籤長相。P2-146 owner response receipt preview 仍屬另一條 release gate 主線；兩者都不得直接打開 Gateway queue、Telegram send、Bot API 或 production write。 +下一步：P2-405C 承接 Telegram approved canary send approval packet；讓統帥先審核發送時間窗、單一訊息類型、dedup、receipt、mute / rollback plan 與停止條件。P2-146 owner response receipt preview 仍屬另一條 release gate 主線；兩者都不得直接打開 Gateway queue、Telegram send、Bot API 或 production write。 ``` diff --git a/docs/ai/AI_AGENT_AUTOMATION_WORKLIST_2026-06-04.md b/docs/ai/AI_AGENT_AUTOMATION_WORKLIST_2026-06-04.md index d86ca41e3..ee4854004 100644 --- a/docs/ai/AI_AGENT_AUTOMATION_WORKLIST_2026-06-04.md +++ b/docs/ai/AI_AGENT_AUTOMATION_WORKLIST_2026-06-04.md @@ -15,7 +15,7 @@ | OpenClaw / Hermes / NemoTron 主動溝通、學習與成長證據 | 100% | P2-401A 到 P2-144 已完成只讀證據面、runtime / report / result-capture gates、no-write readback、promotion review、writer implementation review、writer dry-run fixture、writer dry-run readback、owner promotion execution gate、owner-approved execution rehearsal、owner acceptance / maintenance window gate、owner acceptance readback / preflight hold、owner-approved preflight release package、owner-approved release readiness readback、owner release approval gate、post-release verifier / rollback gate、final release candidate readback、release authorization hold / readback gate、release verifier preflight / owner review packet、release decision hold / readback、release decision next handoff、release decision input prep、12-Agent War Room、owner response 預檢與 owner response 回讀；P2-141 基線與 S4.9 owner release packet 補強皆已正式驗證，P2-142 12-Agent War Room 已完成 production readback 與 desktop / mobile smoke，P2-143 owner response 預檢已完成 production readback 與 in-app browser smoke，P2-144 owner response 回讀已完成 production API readback 與 desktop / mobile smoke。runtime worker、DB migration、production Redis consumer group、canonical runtime readback、live query、runtime score、result capture write、Telegram 實發、delivery receipt E2E、live report delivery、reviewer queue write、Gateway queue write、AI analysis runtime、中低風險 auto worker、KM / LOGBOOK / audit DB / timeline / PlayBook trust 寫入、SDK / 付費服務仍未開 gate | `ai_agent_result_capture_release_decision_owner_response_readback_v1`、`GET /api/v1/agents/agent-result-capture-release-decision-owner-response-readback`、`docs/evaluations/ai_agent_result_capture_release_decision_owner_response_readback_2026-06-14.json`、feature commit `8795f100`、deploy marker `ac938037`、Gitea code-review `2965` / CD `2964` success、5 個回覆讀回 lane、18 個 owner 必填欄位、6 個 readback validation check、6 個 rejection guard、5 個 operator action、等待外部回覆 `5`、未收件 lane `5`、正式寫入 / 發送 `0`；P2-142 feature commit `5de4b3f3`、deploy marker `1a2c9e36`、Gitea CD run `4232` success、production API readback、desktop / mobile in-app browser smoke；P2-143 feature commit `755b0a8d`、deploy marker `667d6329`、Gitea code-review `2961` / CD `2960` success、production API readback、desktop / mobile in-app browser smoke；MASTER §3.2.1b / §3.2.1d / §3.4.3 | | AI Agent 主動營運委派與版本生命週期 | 100% | P2-402A / P2-402B / P2-402C / P2-402D / P2-402E / P2-402F / P2-402G 已完成；已建立 repo-only 版本新鮮度快照、工具採用批准包、Telegram action-required digest policy、Gitea PR 草案 lane、host / K3s / stateful 版本只讀盤點、API 與 governance UI。定期排程、外部版本查詢、工具安裝、CI 變更、套件升級、主機更新、container pull、實際 PR creation、auto merge、Telegram 實發、SSH、kubectl、重啟仍未開 gate | `ai_agent_proactive_operations_contract_v1`、`ai_agent_version_freshness_snapshot_v1`、`ai_agent_tool_adoption_approval_package_v1`、`ai_agent_telegram_action_required_digest_policy_v1`、`ai_agent_gitea_pr_draft_lane_v1`、`ai_agent_host_stateful_version_inventory_v1`、`GET /api/v1/agents/agent-proactive-operations-contract`、`GET /api/v1/agents/agent-version-freshness-snapshot`、`GET /api/v1/agents/agent-tool-adoption-approval-package`、`GET /api/v1/agents/agent-telegram-action-required-digest-policy`、`GET /api/v1/agents/agent-gitea-pr-draft-lane`、`GET /api/v1/agents/agent-host-stateful-version-inventory`、`/zh-TW/governance?tab=automation-inventory`、MASTER §3.2.1c | | 12-Agent War Room 編組 | 72% | 12 個邏輯工位與分批派工規則已正式部署；OpenClaw / Hermes / NemoTron / SRE / Security / DevOps / Data/DR / Supply Chain / Product/UI / QA / Market / Telegram 共 12 份只讀審查已回收；schema / committed snapshot / API / tests / governance UI 區塊 / production API readback / desktop + mobile in-app browser smoke 已完成；runtime writer、Telegram send、Bot API、production write 仍未批准 | `ai_agent_12_agent_war_room_v1`、`docs/evaluations/ai_agent_12_agent_war_room_2026-06-14.json`、`GET /api/v1/agents/agent-12-agent-war-room`、feature commit `5de4b3f3`、deploy marker `1a2c9e36`、Gitea CD run `4232` success、`/zh-TW/governance?tab=automation-inventory`、12 份 Codex sub-agent 只讀回饋 | -| AI Agent 專業任務擴展與 Telegram Runtime Bridge | 84% | P2-405A 已完成本地只讀契約與治理頁 P2-405A 卡片；24 類專業任務、8 個領域、5 段 Telegram bridge、6 種訊息類型、MCP/RAG stack、日報 / 週報 / 月報 / action-required 報告契約已固定；Telegram 實發、Gateway queue、Bot API、delivery receipt、production write、secret read、paid API、host write、kubectl action 仍全部關閉 | `ai_agent_professional_task_expansion_v1`、`docs/evaluations/ai_agent_professional_task_expansion_2026-06-15.json`、`GET /api/v1/agents/agent-professional-task-expansion`、`/zh-TW/governance?tab=automation-inventory`、`docs/ai/AI_AGENT_PROFESSIONAL_TASK_EXPANSION_2026-06-15.md`、需批准任務 `19`、low / medium / high / critical = `3 / 10 / 6 / 5`；下一步 P2-405B 完整 no-send message preview / canary approval package | +| AI Agent 專業任務擴展與 Telegram Runtime Bridge | 88% | P2-405B 已完成只讀契約、正式 API、治理頁 P2-405B 卡片、6 種 Telegram no-send preview、6 個 dedup key、6 組 receipt expectation 與 1 份 canary approval package；24 類專業任務、8 個領域、5 段 Telegram bridge、6 種訊息類型、MCP/RAG stack、日報 / 週報 / 月報 / action-required 報告契約已固定；Telegram 實發、Gateway queue、Bot API、delivery receipt production write、secret read、paid API、host write、kubectl action 仍全部關閉 | `ai_agent_professional_task_expansion_v1`、`docs/evaluations/ai_agent_professional_task_expansion_2026-06-15_1445_p2_405b.json`、`GET /api/v1/agents/agent-professional-task-expansion`、`/zh-TW/governance?tab=automation-inventory`、`docs/ai/AI_AGENT_PROFESSIONAL_TASK_EXPANSION_2026-06-15.md`、需批准任務 `19`、no-send preview `6`、dedup key `6`、receipt expectation `6`、canary package `1`、preview live write `0`；下一步 P2-405C approved canary send approval packet | | Owner response 預檢與拒收邊界 | 100% | P2-143 已完成正式部署與 production readback；承接 P2-141 input prep 與 P2-142 War Room，只建立 owner / verifier / rollback / maintenance / live-apply 五類外部回覆的 intake 預檢、必填欄位與拒收規則；正式 owner response 尚未收到、未接受、未寫入 | `ai_agent_result_capture_release_decision_owner_response_preflight_v1`、`GET /api/v1/agents/agent-result-capture-release-decision-owner-response-preflight`、feature commit `755b0a8d`、deploy marker `667d6329`、Gitea code-review `2961` / CD `2960` success、5 個 response intake lane、18 個 required owner field、6 個 validation check、6 個 rejection guard、5 個 operator action；owner response received / accepted / redacted payload / reviewer queue / Gateway / Telegram / Bot API / production write / secret read / destructive operation 全為 `0` | | Owner response 回讀狀態 | 100% | P2-144 已完成正式部署與 production readback；承接 P2-143 preflight，只讀回五類外部回覆仍未收到、未接受、未拒絕、未保存 | `ai_agent_result_capture_release_decision_owner_response_readback_v1`、`GET /api/v1/agents/agent-result-capture-release-decision-owner-response-readback`、feature commit `8795f100`、deploy marker `ac938037`、Gitea code-review `2965` / CD `2964` success、5 個 response readback lane、18 個 required owner field、6 個 readback validation check、6 個 readback rejection guard、5 個 operator action、waiting external response `5`、no external response received `5`；owner response received / accepted / redacted payload / reviewer queue / Gateway / Telegram / Bot API / production write / secret read / destructive operation 全為 `0` | | 本工作清單與分析報告 | 100% | 已完成 | 本 MD 文件 | @@ -1417,7 +1417,7 @@ UI： ## 13. 立即執行順序 1. P2-004：依賴 / 供應鏈漂移監控，保持只讀觀察與批准包邊界。 -2. P2-405B：AI Agent Telegram 完整 no-send message preview / canary approval package，先讓統帥看見未來 TG 訊息、dedup、receipt 與風險標籤長相；仍不得實發。 +2. P2-405C：AI Agent Telegram approved canary send approval packet，先讓統帥審核發送時間窗、單一訊息類型、dedup、receipt、mute / rollback plan 與停止條件；仍不得在未批准前實發。 3. P3-001：外部 Agent / SDK / API 相關能力仍需證據、費用批准與 shadow / canary 關卡。 ## 14. 目前風險 diff --git a/docs/ai/AI_AGENT_PROFESSIONAL_TASK_EXPANSION_2026-06-15.md b/docs/ai/AI_AGENT_PROFESSIONAL_TASK_EXPANSION_2026-06-15.md index c8a428e48..ace5ffd31 100644 --- a/docs/ai/AI_AGENT_PROFESSIONAL_TASK_EXPANSION_2026-06-15.md +++ b/docs/ai/AI_AGENT_PROFESSIONAL_TASK_EXPANSION_2026-06-15.md @@ -1,12 +1,12 @@ # AI Agent 專業任務擴展與 Telegram Runtime Bridge 工作報告 > 日期：2026-06-15（台北時間） -> 狀態：P2-405A 已完成只讀契約、snapshot、API、測試與治理頁卡片；Telegram 實發仍未啟用。 +> 狀態：P2-405B 已完成 no-send 訊息預覽、dedup key、receipt expectation、canary approval package、API guard、測試與治理頁可視化；Telegram 實發仍未啟用。 > 事實來源：`ai_agent_professional_task_expansion_v1` ## 1. 結論 -本輪把「AI Agent 還能處理哪些專業工作」正式產品化成 24 類專業任務，並把 Telegram 群組 / TG Bot 整合拆成 5 段啟動前閘門。 +本輪把「AI Agent 還能處理哪些專業工作」正式產品化成 24 類專業任務，並把 Telegram 群組 / TG Bot 整合拆成 5 段啟動前閘門。P2-405B 進一步把將來要送出的 Telegram 訊息長相、dedup key、receipt expectation 與 canary approval package 顯示到治理頁。 這不是直接讓 AI Agent 發 Telegram 或改 production；目前只允許 no-send preview、queue preview readback、owner review 與批准包。真正送到 **AwoooI SRE 戰情室** 必須先通過 approved canary、dedup、receipt、redaction、OpenClaw 仲裁、Security gate 與 QA verifier。 @@ -16,8 +16,9 @@ |---|---:|---| | 專業任務擴展契約 | 100% | 24 類任務已定義 owner、風險、MCP、RAG、Telegram policy 與 blocked actions | | Telegram Runtime Bridge 契約 | 100% | no-send preview、queue preview、approved canary、日週月報、action-required digest 已分段 | +| Telegram no-send 訊息預覽 | 100% | 6 種訊息預覽、6 個 dedup key、6 組 receipt expectation、1 份 canary approval package 已固定 | | API / loader | 100% | `GET /api/v1/agents/agent-professional-task-expansion` 只讀輸出 | -| 治理頁可視化 | 100% | `/zh-TW/governance?tab=automation-inventory` 顯示任務、風險、TG bridge 與 live/send/write=0 | +| 治理頁可視化 | 100% | `/zh-TW/governance?tab=automation-inventory` 顯示任務、風險、TG bridge、preview、dedup、receipt、canary 與 live/send/write=0 | | Telegram 實發 | 0% | `telegram_send_count=0`、`bot_api_call_count=0`、`gateway_queue_write_count=0` | | Runtime 自動優化 | 0% | production write、host write、kubectl、paid API、secret read 全部維持 0 | @@ -31,7 +32,32 @@ | 日週月報 digest | blocked until canary receipt | Agent 工作量、完成度、阻擋項、圖表摘要 | 不洗版、不發成功噪音 | | action-required digest | blocked until dual gate | 高風險審核卡、failure-only escalation | 不取代 runtime authorization | -## 4. 專業任務總覽 +## 4. P2-405B no-send 預覽 + +| 訊息類型 | 風險 | 主責 | dedup key | 發送狀態 | +|---|---|---|---|---| +| 日報 | low | Hermes | `awoooi:agent-report:daily:{yyyy-mm-dd}:v1` | no-send | +| 週報 | medium | Hermes | `awoooi:agent-report:weekly:{iso-week}:v1` | no-send | +| 月報 | medium | Hermes | `awoooi:agent-report:monthly:{yyyy-mm}:v1` | no-send | +| 高風險審核卡 | high | OpenClaw | `awoooi:agent-approval:{risk}:{work-item-id}:v1` | no-send | +| failure-only escalation | critical | SRE Sentinel | `awoooi:agent-failure:{alertname}:{target}:{window}:v1` | no-send | +| report receipt gap alert | medium | Telegram Ops | `awoooi:agent-receipt-gap:{message-type}:{window}:v1` | no-send | + +每一種 preview 都包含脫敏 body lines、receipt expectation id、target room env ref、risk tier、approval gate 與 send / queue / bot / receipt write false flags。治理頁只顯示 env ref 與脫敏內容，不顯示 chat id value、Bot token、工作視窗對話、未遮罩 runtime payload 或可執行修復指令。 + +## 5. Canary Approval Package + +目前 canary package 狀態為 `blocked_until_explicit_approval`，package ready 但尚未批准。啟用前必須取得： + +| 必填批准輸入 | 目的 | +|---|---| +| 統帥批准 canary 時間窗 | 避免非預期發送 | +| 單一訊息類型 | 限縮第一次實發範圍 | +| mute / rollback plan | 發生錯誤時可停止 | +| receipt readback owner | 明確誰驗證收到 | +| 失敗時停止條件 | 避免重試洗版 | + +## 6. 專業任務總覽 | 領域 | 任務數 | 代表任務 | 主責 | |---|---:|---|---| @@ -44,7 +70,7 @@ | AI Governance / Replay / Market | 4 | market watch、NemoTron replay、cost forecast、runbook/postmortem | OpenClaw / NemoTron / Hermes | | Telegram / Reports / Receipts | 3 | digest preview、report truth gate、post-action verifier | Telegram Ops / Hermes / OpenClaw | -## 5. 專業能力層級 +## 7. 專業能力層級 | 層級 | AI Agent 可自動做 | Gate | |---|---|---| @@ -53,7 +79,7 @@ | 高風險 | 只產批准包、rollback plan、failure-only digest 草案 | 統帥批准 | | Critical | production write、kubectl、ArgoCD sync、Telegram 實發、secret、restore、host write | 預設 blocked | -## 6. MCP / RAG +## 8. MCP / RAG 首批 MCP：Gitea、Browser、Observability、Telegram Gateway、Package Registry、Database Readonly、Backup Status、ArgoCD Readonly、HTTP Probe、Fixture Store。 @@ -61,7 +87,7 @@ 成長指標：KM entries、PlayBook updates、recommendations、replay score delta、blocked action prevented count、receipt missing count。 -## 7. 邊界 +## 9. 邊界 - 不直接發 Telegram。 - 不寫 Telegram Gateway queue。 @@ -70,9 +96,9 @@ - 不把工作視窗對話、未遮罩提示、私人推理或未遮罩 runtime payload 放進前端或 Telegram。 - 不做 production write、host write、kubectl、ArgoCD sync、restore、rollback、paid API、SDK install。 -## 8. 下一步 +## 10. 下一步 -1. P2-405B：把完整 no-send message preview、dedup key、receipt expectation 與 canary approval package 顯示到治理頁，讓統帥先看 TG 訊息會長什麼樣。 -2. P2-405C：產生 approved canary approval package，明確列出發送時間窗、目標群組、dedup、receipt 與 rollback / mute plan。 -3. P2-405D：canary 通過後才開日報 / 週報 / 月報 digest delivery。 -4. P2-405E：Action-required digest 只對 failure / high-risk / approval-required 事件開啟。 +1. P2-405C：產生 approved canary send approval packet，明確列出發送時間窗、單一訊息類型、目標 env ref、dedup、receipt、rollback / mute plan 與停止條件。 +2. P2-405D：canary 通過後才開日報 / 週報 / 月報 digest delivery。 +3. P2-405E：Action-required digest 只對 failure / high-risk / approval-required 事件開啟。 +4. P2-405F：把 receipt readback 與 report status board 串起來，但仍需 canary gate 後才能寫正式 receipt。 diff --git a/docs/evaluations/ai_agent_professional_task_expansion_2026-06-15_1445_p2_405b.json b/docs/evaluations/ai_agent_professional_task_expansion_2026-06-15_1445_p2_405b.json new file mode 100644 index 000000000..cebd2b7d2 --- /dev/null +++ b/docs/evaluations/ai_agent_professional_task_expansion_2026-06-15_1445_p2_405b.json @@ -0,0 +1,1337 @@ +{ + "schema_version": "ai_agent_professional_task_expansion_v1", + "generated_at": "2026-06-15T14:45:00+08:00", + "program_status": { + "current_priority": "P2", + "current_task_id": "P2-405B", + "next_task_id": "P2-405C", + "overall_completion_percent": 88, + "read_only_mode": true, + "runtime_authority": "professional_task_expansion_and_telegram_bridge_read_only_no_send", + "status_note": "P2-405B 已把 Telegram no-send 訊息預覽、dedup key、receipt expectation 與 canary approval package 固定成只讀契約；目前可在治理頁看見訊息長相與驗收門檻，但仍不寫 Gateway queue、不送 Telegram、不呼叫 Bot API、不讀 secret、不做 production write。" + }, + "source_refs": [ + "docs/ai/AI_AGENT_12_AGENT_WAR_ROOM_2026-06-14.md", + "docs/ai/AI_AGENT_PROACTIVE_OPERATIONS_2026-06-11.md", + "docs/ai/AI_AGENT_INTERACTION_LEARNING_PROOF_2026-06-11.md", + "docs/ai/AI_AGENT_TELEGRAM_ACTION_REQUIRED_DIGEST_POLICY_2026-06-11.md", + "docs/ai/AI_AGENT_DEPLOYMENT_LAYOUT_2026-06-11.md", + "docs/security/MONITORING-OWNER-RESPONSE-ACCEPTANCE.md", + "apps/api/src/services/telegram_gateway.py", + ".gitea/workflows/cd.yaml", + ".gitea/workflows/code-review.yaml" + ], + "telegram_runtime_bridge": { + "canonical_room": "AwoooI SRE 戰情室", + "canonical_room_env": "SRE_GROUP_CHAT_ID", + "gateway_required": true, + "gateway_owner": "telegram_ops_liaison", + "arbiter": "openclaw", + "receipt_owner": "hermes", + "replay_owner": "nemotron", + "no_send_preview_ready": true, + "queue_preview_readback_ready": true, + "approved_canary_required": true, + "direct_bot_api_allowed": false, + "bot_api_call_enabled": false, + "gateway_queue_write_enabled": false, + "telegram_send_enabled": false, + "delivery_receipt_write_enabled": false, + "stages": [ + { + "stage_id": "tg_bridge_01_no_send_preview", + "title": "Telegram no-send preview", + "status": "ready_for_owner_review", + "allowed_output": "脫敏訊息預覽、dedup key、目標群組 metadata、風險層級、需審核選項", + "live_send_enabled": false, + "exit_condition": "OpenClaw + Security + QA 確認不含 secret、raw payload、工作視窗內容與誤導性執行按鈕。" + }, + { + "stage_id": "tg_bridge_02_queue_preview_readback", + "title": "Gateway queue preview readback", + "status": "ready_for_owner_review", + "allowed_output": "queue envelope hash、message type、receipt expectation、rate-limit / dedup plan", + "live_send_enabled": false, + "exit_condition": "Hermes 讀回 preview hash，QA 確認不會寫入 production queue。" + }, + { + "stage_id": "tg_bridge_03_approved_canary", + "title": "Approved canary delivery", + "status": "blocked_until_explicit_approval", + "allowed_output": "單一 approved canary 訊息與 receipt 驗證", + "live_send_enabled": false, + "exit_condition": "統帥明確批准 canary、指定時間窗與 rollback / mute plan。" + }, + { + "stage_id": "tg_bridge_04_daily_weekly_monthly_digest", + "title": "日報 / 週報 / 月報 digest delivery", + "status": "blocked_until_canary_receipt", + "allowed_output": "Agent 工作量、完成度、阻擋項、圖表摘要與 low-noise digest", + "live_send_enabled": false, + "exit_condition": "canary receipt 正常、dedup 有效、成功訊息壓制策略通過。" + }, + { + "stage_id": "tg_bridge_05_action_required_digest", + "title": "Action-required / high-risk approval digest", + "status": "blocked_until_dual_gate", + "allowed_output": "高風險審核卡、owner response 缺口、failure-only escalation", + "live_send_enabled": false, + "exit_condition": "OpenClaw 仲裁、Security gate、QA verifier 與 Telegram receipt 全通過。" + } + ], + "message_types": [ + { + "message_type": "daily_agent_workload_digest", + "cadence": "daily", + "risk_tier": "low", + "owner_agent": "hermes", + "send_policy": "send_after_canary_and_dedup_gate" + }, + { + "message_type": "weekly_governance_summary", + "cadence": "weekly", + "risk_tier": "medium", + "owner_agent": "hermes", + "send_policy": "send_after_canary_and_truth_gate" + }, + { + "message_type": "monthly_growth_report", + "cadence": "monthly", + "risk_tier": "medium", + "owner_agent": "hermes", + "send_policy": "send_after_owner_review" + }, + { + "message_type": "high_risk_approval_card", + "cadence": "event", + "risk_tier": "high", + "owner_agent": "openclaw", + "send_policy": "approval_required_before_send" + }, + { + "message_type": "failure_only_escalation", + "cadence": "event", + "risk_tier": "critical", + "owner_agent": "sre_sentinel", + "send_policy": "failure_only_after_route_lock" + }, + { + "message_type": "report_receipt_gap_alert", + "cadence": "event", + "risk_tier": "medium", + "owner_agent": "telegram_ops_liaison", + "send_policy": "no_send_preview_until_receipt_contract" + } + ], + "no_send_message_previews": [ + { + "preview_id": "p2_405b_preview_daily_agent_workload_digest_v1", + "message_type": "daily_agent_workload_digest", + "title": "AI Agent 日報 no-send 預覽", + "target_room_alias": "awoooi_sre_war_room_env_ref", + "target_room_env": "SRE_GROUP_CHAT_ID", + "owner_agent": "hermes", + "risk_tier": "low", + "status": "preview_ready_no_send", + "dedup_key": "awoooi:agent-report:daily:{yyyy-mm-dd}:v1", + "receipt_expectation_id": "p2_405b_receipt_daily_agent_workload_digest_v1", + "approval_required_before_send": false, + "canary_required_before_live": true, + "send_enabled": false, + "gateway_queue_write_enabled": false, + "bot_api_call_enabled": false, + "delivery_receipt_write_enabled": false, + "sanitized_body_lines": [ + "AWOOOI AI Agent 日報", + "今日完成度、工作量、阻擋項與低風險自動處理摘要。", + "所有成功煙霧訊號只彙整，不逐筆洗版。", + "live / send / write 計數維持 0，等待 canary gate。" + ], + "redaction_checks": [ + "不含機密明文", + "不含未遮罩工作上下文", + "不含未遮罩 runtime payload", + "不含可執行修復指令", + "不含誤導性成功宣告" + ] + }, + { + "preview_id": "p2_405b_preview_weekly_governance_summary_v1", + "message_type": "weekly_governance_summary", + "title": "AI Agent 週報 no-send 預覽", + "target_room_alias": "awoooi_sre_war_room_env_ref", + "target_room_env": "SRE_GROUP_CHAT_ID", + "owner_agent": "hermes", + "risk_tier": "medium", + "status": "preview_ready_no_send", + "dedup_key": "awoooi:agent-report:weekly:{iso-week}:v1", + "receipt_expectation_id": "p2_405b_receipt_weekly_governance_summary_v1", + "approval_required_before_send": false, + "canary_required_before_live": true, + "send_enabled": false, + "gateway_queue_write_enabled": false, + "bot_api_call_enabled": false, + "delivery_receipt_write_enabled": false, + "sanitized_body_lines": [ + "AWOOOI AI Agent 週報", + "本週專業任務、模型市場 watch、治理 gate 與未完成 owner response 摘要。", + "中風險項目只列 proposal / dry-run，不列正式執行結果。", + "任何需要批准的項目改送 action-required digest 草案。" + ], + "redaction_checks": [ + "不含機密明文", + "不含未遮罩工作上下文", + "不含未遮罩 runtime payload", + "不含可執行修復指令", + "不含誤導性成功宣告" + ] + }, + { + "preview_id": "p2_405b_preview_monthly_growth_report_v1", + "message_type": "monthly_growth_report", + "title": "AI Agent 月報 no-send 預覽", + "target_room_alias": "awoooi_sre_war_room_env_ref", + "target_room_env": "SRE_GROUP_CHAT_ID", + "owner_agent": "hermes", + "risk_tier": "medium", + "status": "preview_ready_no_send", + "dedup_key": "awoooi:agent-report:monthly:{yyyy-mm}:v1", + "receipt_expectation_id": "p2_405b_receipt_monthly_growth_report_v1", + "approval_required_before_send": true, + "canary_required_before_live": true, + "send_enabled": false, + "gateway_queue_write_enabled": false, + "bot_api_call_enabled": false, + "delivery_receipt_write_enabled": false, + "sanitized_body_lines": [ + "AWOOOI AI Agent 月報", + "智慧成長指標、PlayBook 更新、replay 分數變化與 blocked action 防護數。", + "成本、容量、版本更新與外部 Agent 市場候選以審核建議呈現。", + "月報首次實發前需要 owner review 與 canary receipt。" + ], + "redaction_checks": [ + "不含機密明文", + "不含未遮罩工作上下文", + "不含未遮罩 runtime payload", + "不含可執行修復指令", + "不含誤導性成功宣告" + ] + }, + { + "preview_id": "p2_405b_preview_high_risk_approval_card_v1", + "message_type": "high_risk_approval_card", + "title": "高風險審核卡 no-send 預覽", + "target_room_alias": "awoooi_sre_war_room_env_ref", + "target_room_env": "SRE_GROUP_CHAT_ID", + "owner_agent": "openclaw", + "risk_tier": "high", + "status": "preview_ready_no_send", + "dedup_key": "awoooi:agent-approval:{risk}:{work-item-id}:v1", + "receipt_expectation_id": "p2_405b_receipt_high_risk_approval_card_v1", + "approval_required_before_send": true, + "canary_required_before_live": true, + "send_enabled": false, + "gateway_queue_write_enabled": false, + "bot_api_call_enabled": false, + "delivery_receipt_write_enabled": false, + "sanitized_body_lines": [ + "AWOOOI 高風險 AI Agent 審核卡", + "列出候選動作、風險、rollback plan、verifier plan 與 owner 缺口。", + "審核卡不等於 runtime authorization，按鈕與指令預設不顯示。", + "OpenClaw 仲裁、Security gate 與 QA verifier 必須全部通過。" + ], + "redaction_checks": [ + "不含機密明文", + "不含未遮罩工作上下文", + "不含未遮罩 runtime payload", + "不含可執行修復指令", + "不含誤導性成功宣告" + ] + }, + { + "preview_id": "p2_405b_preview_failure_only_escalation_v1", + "message_type": "failure_only_escalation", + "title": "Failure-only 升級 no-send 預覽", + "target_room_alias": "awoooi_sre_war_room_env_ref", + "target_room_env": "SRE_GROUP_CHAT_ID", + "owner_agent": "sre_sentinel", + "risk_tier": "critical", + "status": "preview_ready_no_send", + "dedup_key": "awoooi:agent-failure:{alertname}:{target}:{window}:v1", + "receipt_expectation_id": "p2_405b_receipt_failure_only_escalation_v1", + "approval_required_before_send": true, + "canary_required_before_live": true, + "send_enabled": false, + "gateway_queue_write_enabled": false, + "bot_api_call_enabled": false, + "delivery_receipt_write_enabled": false, + "sanitized_body_lines": [ + "AWOOOI failure-only 升級通知", + "只在失敗、阻擋或高風險審核缺口時產生，不發成功噪音。", + "包含 blocker summary、下一步、owner 欄位與 verifier plan。", + "任何修復執行仍需獨立批准與 post-action verifier。" + ], + "redaction_checks": [ + "不含機密明文", + "不含未遮罩工作上下文", + "不含未遮罩 runtime payload", + "不含可執行修復指令", + "不含誤導性成功宣告" + ] + }, + { + "preview_id": "p2_405b_preview_report_receipt_gap_alert_v1", + "message_type": "report_receipt_gap_alert", + "title": "報表回執缺口 no-send 預覽", + "target_room_alias": "awoooi_sre_war_room_env_ref", + "target_room_env": "SRE_GROUP_CHAT_ID", + "owner_agent": "telegram_ops_liaison", + "risk_tier": "medium", + "status": "preview_ready_no_send", + "dedup_key": "awoooi:agent-receipt-gap:{message-type}:{window}:v1", + "receipt_expectation_id": "p2_405b_receipt_report_receipt_gap_alert_v1", + "approval_required_before_send": true, + "canary_required_before_live": true, + "send_enabled": false, + "gateway_queue_write_enabled": false, + "bot_api_call_enabled": false, + "delivery_receipt_write_enabled": false, + "sanitized_body_lines": [ + "AWOOOI 報表回執缺口提示", + "用來指出 preview hash、dedup key、canary receipt 或 readback evidence 不完整。", + "此提示本身仍是 no-send preview，不寫 delivery receipt。", + "修復方向只產生 owner review 項目，不自動改 Gateway。" + ], + "redaction_checks": [ + "不含機密明文", + "不含未遮罩工作上下文", + "不含未遮罩 runtime payload", + "不含可執行修復指令", + "不含誤導性成功宣告" + ] + } + ], + "dedup_policy": { + "required": true, + "key_count": 6, + "collision_policy": "drop_duplicate_and_raise_preview_gap_without_send", + "preview_hash_algorithm": "sha256_preview_only", + "live_cache_write_enabled": false, + "keys": [ + { + "message_type": "daily_agent_workload_digest", + "key_template": "awoooi:agent-report:daily:{yyyy-mm-dd}:v1", + "owner_agent": "hermes", + "risk_tier": "low", + "window": "daily", + "live_cache_write_enabled": false + }, + { + "message_type": "weekly_governance_summary", + "key_template": "awoooi:agent-report:weekly:{iso-week}:v1", + "owner_agent": "hermes", + "risk_tier": "medium", + "window": "event_or_cadence_window", + "live_cache_write_enabled": false + }, + { + "message_type": "monthly_growth_report", + "key_template": "awoooi:agent-report:monthly:{yyyy-mm}:v1", + "owner_agent": "hermes", + "risk_tier": "medium", + "window": "event_or_cadence_window", + "live_cache_write_enabled": false + }, + { + "message_type": "high_risk_approval_card", + "key_template": "awoooi:agent-approval:{risk}:{work-item-id}:v1", + "owner_agent": "openclaw", + "risk_tier": "high", + "window": "event_or_cadence_window", + "live_cache_write_enabled": false + }, + { + "message_type": "failure_only_escalation", + "key_template": "awoooi:agent-failure:{alertname}:{target}:{window}:v1", + "owner_agent": "sre_sentinel", + "risk_tier": "critical", + "window": "event_or_cadence_window", + "live_cache_write_enabled": false + }, + { + "message_type": "report_receipt_gap_alert", + "key_template": "awoooi:agent-receipt-gap:{message-type}:{window}:v1", + "owner_agent": "telegram_ops_liaison", + "risk_tier": "medium", + "window": "event_or_cadence_window", + "live_cache_write_enabled": false + } + ] + }, + "queue_preview_readback": { + "enabled": true, + "preview_only": true, + "write_enabled": false, + "readback_enabled": true, + "envelope_fields": [ + "message_type", + "dedup_key", + "sanitized_preview_hash", + "target_room_env_ref", + "risk_tier", + "approval_gate" + ], + "readback_checks": [ + "preview hash 必須與治理頁顯示一致", + "dedup key 必須唯一", + "target room 只能顯示 env ref", + "所有 send / queue / bot flags 必須為 false" + ] + }, + "receipt_expectations": [ + { + "receipt_id": "p2_405b_receipt_daily_agent_workload_digest_v1", + "message_type": "daily_agent_workload_digest", + "preview_id": "p2_405b_preview_daily_agent_workload_digest_v1", + "expected_signal": "preview_hash_and_dedup_key_match_after_approved_canary", + "required_evidence_refs": [ + "sanitized_preview_hash", + "dedup_key", + "target_room_env_ref", + "canary_window_ref", + "gateway_preview_readback_ref" + ], + "receipt_write_enabled": false, + "production_receipt_readback_enabled": false, + "missing_receipt_policy": "create_report_receipt_gap_alert_preview_only" + }, + { + "receipt_id": "p2_405b_receipt_weekly_governance_summary_v1", + "message_type": "weekly_governance_summary", + "preview_id": "p2_405b_preview_weekly_governance_summary_v1", + "expected_signal": "preview_hash_and_dedup_key_match_after_approved_canary", + "required_evidence_refs": [ + "sanitized_preview_hash", + "dedup_key", + "target_room_env_ref", + "canary_window_ref", + "gateway_preview_readback_ref" + ], + "receipt_write_enabled": false, + "production_receipt_readback_enabled": false, + "missing_receipt_policy": "create_report_receipt_gap_alert_preview_only" + }, + { + "receipt_id": "p2_405b_receipt_monthly_growth_report_v1", + "message_type": "monthly_growth_report", + "preview_id": "p2_405b_preview_monthly_growth_report_v1", + "expected_signal": "preview_hash_and_dedup_key_match_after_approved_canary", + "required_evidence_refs": [ + "sanitized_preview_hash", + "dedup_key", + "target_room_env_ref", + "canary_window_ref", + "gateway_preview_readback_ref" + ], + "receipt_write_enabled": false, + "production_receipt_readback_enabled": false, + "missing_receipt_policy": "create_report_receipt_gap_alert_preview_only" + }, + { + "receipt_id": "p2_405b_receipt_high_risk_approval_card_v1", + "message_type": "high_risk_approval_card", + "preview_id": "p2_405b_preview_high_risk_approval_card_v1", + "expected_signal": "preview_hash_and_dedup_key_match_after_approved_canary", + "required_evidence_refs": [ + "sanitized_preview_hash", + "dedup_key", + "target_room_env_ref", + "canary_window_ref", + "gateway_preview_readback_ref" + ], + "receipt_write_enabled": false, + "production_receipt_readback_enabled": false, + "missing_receipt_policy": "create_report_receipt_gap_alert_preview_only" + }, + { + "receipt_id": "p2_405b_receipt_failure_only_escalation_v1", + "message_type": "failure_only_escalation", + "preview_id": "p2_405b_preview_failure_only_escalation_v1", + "expected_signal": "preview_hash_and_dedup_key_match_after_approved_canary", + "required_evidence_refs": [ + "sanitized_preview_hash", + "dedup_key", + "target_room_env_ref", + "canary_window_ref", + "gateway_preview_readback_ref" + ], + "receipt_write_enabled": false, + "production_receipt_readback_enabled": false, + "missing_receipt_policy": "create_report_receipt_gap_alert_preview_only" + }, + { + "receipt_id": "p2_405b_receipt_report_receipt_gap_alert_v1", + "message_type": "report_receipt_gap_alert", + "preview_id": "p2_405b_preview_report_receipt_gap_alert_v1", + "expected_signal": "preview_hash_and_dedup_key_match_after_approved_canary", + "required_evidence_refs": [ + "sanitized_preview_hash", + "dedup_key", + "target_room_env_ref", + "canary_window_ref", + "gateway_preview_readback_ref" + ], + "receipt_write_enabled": false, + "production_receipt_readback_enabled": false, + "missing_receipt_policy": "create_report_receipt_gap_alert_preview_only" + } + ], + "canary_approval_package": { + "package_id": "p2_405b_telegram_canary_approval_package_v1", + "status": "blocked_until_explicit_approval", + "package_ready": true, + "approval_required": true, + "target_room_alias": "awoooi_sre_war_room_env_ref", + "target_room_env": "SRE_GROUP_CHAT_ID", + "arbiter": "openclaw", + "owner_agent": "telegram_ops_liaison", + "reviewers": [ + "security_sentinel", + "qa_verifier", + "hermes" + ], + "message_type_scope": [ + "daily_agent_workload_digest", + "weekly_governance_summary", + "monthly_growth_report", + "high_risk_approval_card", + "failure_only_escalation", + "report_receipt_gap_alert" + ], + "required_approval_inputs": [ + "統帥批准 canary 時間窗", + "可接受的單一訊息類型", + "mute / rollback plan", + "receipt readback owner", + "失敗時停止條件" + ], + "approval_checklist": [ + "OpenClaw 仲裁通過", + "Security redaction gate 通過", + "QA verifier 確認 no-send preview 與 receipt expectation", + "Hermes 確認報告真實性 gate", + "Telegram Ops 確認 dedup 與 rate-limit plan" + ], + "live_send_enabled": false, + "gateway_queue_write_enabled": false, + "bot_api_call_enabled": false, + "delivery_receipt_write_enabled": false, + "production_write_enabled": false + }, + "no_send_preview_completion_percent": 100, + "canary_approval_package_completion_percent": 100 + }, + "professional_task_domains": [ + { + "domain_id": "sre_observability", + "label": "SRE / Observability", + "owner_agent": "sre_sentinel" + }, + { + "domain_id": "security_iwooos", + "label": "Security / IwoooS", + "owner_agent": "security_sentinel" + }, + { + "domain_id": "devops_release", + "label": "DevOps / Release", + "owner_agent": "devops_commander" + }, + { + "domain_id": "data_dr", + "label": "Data / Backup / DR", + "owner_agent": "data_dr_guardian" + }, + { + "domain_id": "supply_chain", + "label": "Supply Chain / Versions", + "owner_agent": "supply_chain_scout" + }, + { + "domain_id": "product_quality", + "label": "Product / UI / QA", + "owner_agent": "product_ui_curator" + }, + { + "domain_id": "ai_governance", + "label": "AI Governance / Replay / Market", + "owner_agent": "openclaw" + }, + { + "domain_id": "telegram_reporting", + "label": "Telegram / Reports / Receipts", + "owner_agent": "telegram_ops_liaison" + } + ], + "professional_tasks": [ + { + "task_id": "pro_task_001_incident_correlation", + "title": "跨來源 incident correlation 與根因候選排序", + "domain_id": "sre_observability", + "owner_agent": "sre_sentinel", + "supporting_agents": [ + "openclaw", + "hermes" + ], + "risk_tier": "medium", + "automation_mode": "no_write_dry_run", + "telegram_policy": "action_required_preview", + "required_mcp": [ + "observability", + "gitea", + "browser" + ], + "required_rag": [ + "incident_traces", + "runbooks" + ], + "approval_required": true, + "current_live_count_24h": 0, + "blocked_actions": [ + "alert_rule_change", + "silence_change", + "restart_service" + ] + }, + { + "task_id": "pro_task_002_alert_noise_budget", + "title": "告警噪音預算與降噪提案", + "domain_id": "sre_observability", + "owner_agent": "sre_sentinel", + "supporting_agents": [ + "hermes", + "openclaw" + ], + "risk_tier": "medium", + "automation_mode": "proposal_only", + "telegram_policy": "weekly_digest", + "required_mcp": [ + "prometheus", + "alertmanager" + ], + "required_rag": [ + "alert_traces", + "runbooks" + ], + "approval_required": true, + "current_live_count_24h": 0, + "blocked_actions": [ + "receiver_route_change", + "silence_policy_change", + "prometheus_reload" + ] + }, + { + "task_id": "pro_task_003_synthetic_check_plan", + "title": "公開路由 synthetic check 與失敗分類計畫", + "domain_id": "sre_observability", + "owner_agent": "qa_verifier", + "supporting_agents": [ + "product_ui_curator", + "sre_sentinel" + ], + "risk_tier": "low", + "automation_mode": "auto_summary_after_guard", + "telegram_policy": "daily_digest", + "required_mcp": [ + "browser", + "http_probe" + ], + "required_rag": [ + "route_runbooks" + ], + "approval_required": false, + "current_live_count_24h": 0, + "blocked_actions": [ + "active_scan", + "production_write" + ] + }, + { + "task_id": "pro_task_004_backup_restore_readiness", + "title": "備份 freshness、restore drill 與 DR readiness 報告", + "domain_id": "data_dr", + "owner_agent": "data_dr_guardian", + "supporting_agents": [ + "hermes", + "openclaw" + ], + "risk_tier": "high", + "automation_mode": "approval_required_before_execution", + "telegram_policy": "action_required_preview", + "required_mcp": [ + "backup_status", + "gitea" + ], + "required_rag": [ + "dr_runbooks" + ], + "approval_required": true, + "current_live_count_24h": 0, + "blocked_actions": [ + "restore_run", + "offsite_sync", + "credential_escrow_marker_write" + ] + }, + { + "task_id": "pro_task_005_non_secret_escrow_evidence", + "title": "credential escrow 非敏感證據補件路由", + "domain_id": "data_dr", + "owner_agent": "data_dr_guardian", + "supporting_agents": [ + "security_sentinel", + "hermes" + ], + "risk_tier": "critical", + "automation_mode": "blocked_until_owner_response", + "telegram_policy": "action_required_preview", + "required_mcp": [ + "gitea" + ], + "required_rag": [ + "dr_runbooks", + "hard_rules" + ], + "approval_required": true, + "current_live_count_24h": 0, + "blocked_actions": [ + "secret_value_collection", + "credential_read", + "marker_write_without_owner" + ] + }, + { + "task_id": "pro_task_006_dependency_cve_triage", + "title": "依賴 CVE / EOL / license 風險分流", + "domain_id": "supply_chain", + "owner_agent": "supply_chain_scout", + "supporting_agents": [ + "openclaw", + "security_sentinel" + ], + "risk_tier": "medium", + "automation_mode": "proposal_only", + "telegram_policy": "weekly_digest", + "required_mcp": [ + "package_registry", + "osv" + ], + "required_rag": [ + "dependency_policy" + ], + "approval_required": true, + "current_live_count_24h": 0, + "blocked_actions": [ + "package_upgrade", + "lockfile_write", + "external_scan_without_approval" + ] + }, + { + "task_id": "pro_task_007_sbom_license_drift", + "title": "SBOM / license / Docker surface drift 批准包", + "domain_id": "supply_chain", + "owner_agent": "supply_chain_scout", + "supporting_agents": [ + "hermes", + "security_sentinel" + ], + "risk_tier": "medium", + "automation_mode": "proposal_only", + "telegram_policy": "weekly_digest", + "required_mcp": [ + "syft", + "trivy" + ], + "required_rag": [ + "supply_chain_runbooks" + ], + "approval_required": true, + "current_live_count_24h": 0, + "blocked_actions": [ + "tool_install", + "image_pull", + "sbom_upload" + ] + }, + { + "task_id": "pro_task_008_image_digest_pin_plan", + "title": "Docker image digest pinning 與 rollback 計畫", + "domain_id": "supply_chain", + "owner_agent": "devops_commander", + "supporting_agents": [ + "supply_chain_scout", + "qa_verifier" + ], + "risk_tier": "high", + "automation_mode": "approval_required_before_execution", + "telegram_policy": "action_required_preview", + "required_mcp": [ + "registry", + "gitea" + ], + "required_rag": [ + "release_runbooks" + ], + "approval_required": true, + "current_live_count_24h": 0, + "blocked_actions": [ + "image_pull", + "image_push", + "kustomization_write_without_review" + ] + }, + { + "task_id": "pro_task_009_gitops_diff_review", + "title": "K8s / ArgoCD manifest diff 與 blast-radius review", + "domain_id": "devops_release", + "owner_agent": "devops_commander", + "supporting_agents": [ + "openclaw", + "qa_verifier" + ], + "risk_tier": "critical", + "automation_mode": "approval_required_before_execution", + "telegram_policy": "action_required_preview", + "required_mcp": [ + "gitea", + "argocd" + ], + "required_rag": [ + "gitops_runbooks" + ], + "approval_required": true, + "current_live_count_24h": 0, + "blocked_actions": [ + "argocd_sync", + "kubectl_apply", + "helm_upgrade" + ] + }, + { + "task_id": "pro_task_010_gateway_route_impact", + "title": "Nginx / public gateway route impact analysis", + "domain_id": "devops_release", + "owner_agent": "devops_commander", + "supporting_agents": [ + "security_sentinel", + "qa_verifier" + ], + "risk_tier": "critical", + "automation_mode": "approval_required_before_execution", + "telegram_policy": "action_required_preview", + "required_mcp": [ + "gitea", + "browser" + ], + "required_rag": [ + "gateway_runbooks" + ], + "approval_required": true, + "current_live_count_24h": 0, + "blocked_actions": [ + "nginx_reload", + "route_change", + "certbot_renew" + ] + }, + { + "task_id": "pro_task_011_public_bundle_redaction", + "title": "前台 bundle / DOM / i18n 敏感資訊防洩漏掃描", + "domain_id": "security_iwooos", + "owner_agent": "security_sentinel", + "supporting_agents": [ + "product_ui_curator", + "qa_verifier" + ], + "risk_tier": "medium", + "automation_mode": "no_write_dry_run", + "telegram_policy": "daily_digest", + "required_mcp": [ + "browser", + "gitea" + ], + "required_rag": [ + "hard_rules" + ], + "approval_required": true, + "current_live_count_24h": 0, + "blocked_actions": [ + "frontend_deploy_without_smoke", + "secret_storage" + ] + }, + { + "task_id": "pro_task_012_secret_metadata_parity", + "title": "Secret name parity、injection route 與 redaction gate", + "domain_id": "security_iwooos", + "owner_agent": "security_sentinel", + "supporting_agents": [ + "devops_commander", + "openclaw" + ], + "risk_tier": "critical", + "automation_mode": "blocked_until_owner_response", + "telegram_policy": "action_required_preview", + "required_mcp": [ + "gitea" + ], + "required_rag": [ + "secret_runbooks" + ], + "approval_required": true, + "current_live_count_24h": 0, + "blocked_actions": [ + "secret_value_read", + "secret_rotation", + "workflow_modification" + ] + }, + { + "task_id": "pro_task_013_telegram_digest_preview", + "title": "AI Agent 日週月報 Telegram no-send preview", + "domain_id": "telegram_reporting", + "owner_agent": "telegram_ops_liaison", + "supporting_agents": [ + "hermes", + "qa_verifier" + ], + "risk_tier": "medium", + "automation_mode": "no_write_dry_run", + "telegram_policy": "no_direct_telegram", + "required_mcp": [ + "telegram_gateway" + ], + "required_rag": [ + "logbook_chunks" + ], + "approval_required": true, + "current_live_count_24h": 0, + "blocked_actions": [ + "telegram_send", + "bot_api_call", + "gateway_queue_write" + ] + }, + { + "task_id": "pro_task_014_report_truth_gate", + "title": "日報 / 週報 / 月報 truth gate 與可處置性審查", + "domain_id": "telegram_reporting", + "owner_agent": "hermes", + "supporting_agents": [ + "openclaw", + "sre_sentinel" + ], + "risk_tier": "medium", + "automation_mode": "auto_summary_after_guard", + "telegram_policy": "weekly_digest", + "required_mcp": [ + "gitea", + "observability" + ], + "required_rag": [ + "logbook_chunks", + "incident_traces" + ], + "approval_required": false, + "current_live_count_24h": 0, + "blocked_actions": [ + "false_green_report", + "report_send_without_truth_gate" + ] + }, + { + "task_id": "pro_task_015_agent_market_watch", + "title": "主流 AI Agent / SDK / 模型版本市場雷達", + "domain_id": "ai_governance", + "owner_agent": "market_scout", + "supporting_agents": [ + "openclaw", + "nemotron" + ], + "risk_tier": "medium", + "automation_mode": "proposal_only", + "telegram_policy": "weekly_digest", + "required_mcp": [ + "web_primary_sources", + "gitea" + ], + "required_rag": [ + "agent_market_evidence" + ], + "approval_required": true, + "current_live_count_24h": 0, + "blocked_actions": [ + "sdk_install", + "paid_api_call", + "production_route_change" + ] + }, + { + "task_id": "pro_task_016_model_replay_scorecard", + "title": "NemoTron / 候選模型 replay scorecard 與 failure mode 分析", + "domain_id": "ai_governance", + "owner_agent": "nemotron", + "supporting_agents": [ + "openclaw", + "qa_verifier" + ], + "risk_tier": "high", + "automation_mode": "no_write_dry_run", + "telegram_policy": "action_required_preview", + "required_mcp": [ + "fixture_store", + "gitea" + ], + "required_rag": [ + "replay_fixtures" + ], + "approval_required": true, + "current_live_count_24h": 0, + "blocked_actions": [ + "shadow_route", + "canary_route", + "full_replay_without_approval" + ] + }, + { + "task_id": "pro_task_017_cost_capacity_forecast", + "title": "AI provider、主機、K3s 與外部服務成本 / 容量預測", + "domain_id": "ai_governance", + "owner_agent": "openclaw", + "supporting_agents": [ + "supply_chain_scout", + "sre_sentinel" + ], + "risk_tier": "high", + "automation_mode": "proposal_only", + "telegram_policy": "monthly_digest", + "required_mcp": [ + "observability", + "billing_metadata" + ], + "required_rag": [ + "cost_policy" + ], + "approval_required": true, + "current_live_count_24h": 0, + "blocked_actions": [ + "paid_provider_switch", + "limit_increase", + "paid_api_call" + ] + }, + { + "task_id": "pro_task_018_db_query_triage", + "title": "PostgreSQL 慢查詢、索引與資料新鮮度異常 triage", + "domain_id": "sre_observability", + "owner_agent": "sre_sentinel", + "supporting_agents": [ + "data_dr_guardian", + "openclaw" + ], + "risk_tier": "high", + "automation_mode": "proposal_only", + "telegram_policy": "action_required_preview", + "required_mcp": [ + "database_readonly", + "observability" + ], + "required_rag": [ + "db_runbooks" + ], + "approval_required": true, + "current_live_count_24h": 0, + "blocked_actions": [ + "migration", + "reindex", + "write_query" + ] + }, + { + "task_id": "pro_task_019_data_freshness_source_reliability", + "title": "資料來源新鮮度、fallback 與來源可信度監控", + "domain_id": "data_dr", + "owner_agent": "hermes", + "supporting_agents": [ + "sre_sentinel", + "qa_verifier" + ], + "risk_tier": "medium", + "automation_mode": "auto_summary_after_guard", + "telegram_policy": "daily_digest", + "required_mcp": [ + "http_probe", + "database_readonly" + ], + "required_rag": [ + "data_quality_runbooks" + ], + "approval_required": false, + "current_live_count_24h": 0, + "blocked_actions": [ + "source_switch_without_owner", + "data_write" + ] + }, + { + "task_id": "pro_task_020_ui_accessibility_regression", + "title": "前後台 mobile overflow、console error、可讀性與 a11y regression", + "domain_id": "product_quality", + "owner_agent": "product_ui_curator", + "supporting_agents": [ + "qa_verifier", + "security_sentinel" + ], + "risk_tier": "low", + "automation_mode": "auto_summary_after_guard", + "telegram_policy": "daily_digest", + "required_mcp": [ + "browser" + ], + "required_rag": [ + "frontend_guidelines" + ], + "approval_required": false, + "current_live_count_24h": 0, + "blocked_actions": [ + "unrelated_redesign", + "public_secret_display" + ] + }, + { + "task_id": "pro_task_021_runbook_postmortem_draft", + "title": "Runbook、postmortem 與 owner handoff 草稿", + "domain_id": "ai_governance", + "owner_agent": "hermes", + "supporting_agents": [ + "openclaw", + "sre_sentinel" + ], + "risk_tier": "low", + "automation_mode": "auto_summary_after_guard", + "telegram_policy": "weekly_digest", + "required_mcp": [ + "gitea" + ], + "required_rag": [ + "logbook_chunks", + "runbooks" + ], + "approval_required": false, + "current_live_count_24h": 0, + "blocked_actions": [ + "canonical_km_write_without_review", + "secret_in_doc" + ] + }, + { + "task_id": "pro_task_022_owner_response_completeness", + "title": "Owner response 欄位完整性、補件與拒收分流", + "domain_id": "security_iwooos", + "owner_agent": "security_sentinel", + "supporting_agents": [ + "hermes", + "openclaw" + ], + "risk_tier": "high", + "automation_mode": "proposal_only", + "telegram_policy": "action_required_preview", + "required_mcp": [ + "gitea" + ], + "required_rag": [ + "owner_response_templates" + ], + "approval_required": true, + "current_live_count_24h": 0, + "blocked_actions": [ + "mark_accepted_without_reviewer", + "runtime_gate_open" + ] + }, + { + "task_id": "pro_task_023_candidate_pr_draft_packaging", + "title": "低 / 中風險候選 PR 草案、測試與 rollback 封包", + "domain_id": "devops_release", + "owner_agent": "devops_commander", + "supporting_agents": [ + "qa_verifier", + "openclaw" + ], + "risk_tier": "medium", + "automation_mode": "proposal_only", + "telegram_policy": "action_required_preview", + "required_mcp": [ + "gitea" + ], + "required_rag": [ + "release_runbooks" + ], + "approval_required": true, + "current_live_count_24h": 0, + "blocked_actions": [ + "branch_push_without_approval", + "auto_merge", + "workflow_trigger" + ] + }, + { + "task_id": "pro_task_024_post_action_verifier", + "title": "Post-action verifier、rollback trigger 與 receipt reconciliation", + "domain_id": "telegram_reporting", + "owner_agent": "openclaw", + "supporting_agents": [ + "telegram_ops_liaison", + "qa_verifier" + ], + "risk_tier": "critical", + "automation_mode": "approval_required_before_execution", + "telegram_policy": "failure_only_preview", + "required_mcp": [ + "telegram_gateway", + "observability", + "gitea" + ], + "required_rag": [ + "verification_runbooks" + ], + "approval_required": true, + "current_live_count_24h": 0, + "blocked_actions": [ + "verifier_live_readback_without_gate", + "rollback_without_owner", + "telegram_send_without_receipt" + ] + } + ], + "mcp_rag_stack": { + "mcp_connectors": [ + "gitea", + "browser", + "observability", + "telegram_gateway", + "package_registry", + "database_readonly", + "backup_status", + "argocd_readonly", + "http_probe", + "fixture_store" + ], + "rag_sources": [ + "LOGBOOK chunks", + "HARD_RULES / ADR chunks", + "runbook chunks", + "incident / alert traces", + "agent market evidence", + "owner response templates" + ], + "growth_metrics": [ + "km_entries_created", + "playbook_updates", + "recommendations_created", + "replay_score_delta", + "blocked_action_prevented_count", + "receipt_missing_count" + ] + }, + "risk_policy": { + "low": "文件、報表、UI smoke 與只讀摘要可在 guard 後自動處理並進每日 digest。", + "medium": "中風險只能先產 proposal、no-write dry-run 或 PR 草案；涉及設定、外部服務、成本、路由或未知狀態時升級審核。", + "high": "高風險必須 OpenClaw 仲裁、QA / Security gate 與統帥批准；不得自動執行。", + "critical": "production write、kubectl、ArgoCD sync、Telegram 實發、secret、付費 API、restore、rollback、host write 均維持 blocked until explicit approval。" + }, + "reporting_contract": { + "daily": { + "required": true, + "owner_agent": "hermes", + "telegram_stage": "tg_bridge_04_daily_weekly_monthly_digest" + }, + "weekly": { + "required": true, + "owner_agent": "hermes", + "telegram_stage": "tg_bridge_04_daily_weekly_monthly_digest" + }, + "monthly": { + "required": true, + "owner_agent": "hermes", + "telegram_stage": "tg_bridge_04_daily_weekly_monthly_digest" + }, + "action_required": { + "required": true, + "owner_agent": "openclaw", + "telegram_stage": "tg_bridge_05_action_required_digest" + } + }, + "redaction_contract": { + "redaction_required": true, + "conversation_transcript_display_allowed": false, + "raw_prompt_display_allowed": false, + "private_reasoning_display_allowed": false, + "secret_value_display_allowed": false, + "raw_runtime_payload_display_allowed": false, + "telegram_message_must_be_sanitized": true, + "forbidden_terms": [ + "work_window_transcript", + "raw prompt", + "private reasoning", + "chain-of-thought", + "telegram token", + "authorization header", + "secret value" + ], + "frontend_display_policy": "前端與 Telegram preview 只顯示脫敏摘要、env ref、dedup key、preview hash 與審核狀態；不得顯示工作視窗對話、未遮罩提示、私有推理、機密明文或未遮罩 runtime payload。", + "message_preview_redaction_checks": [ + "conversation_context_redacted", + "prompt_surface_redacted", + "runtime_payload_redacted", + "credential_material_absent", + "action_button_absent", + "execution_claim_absent" + ] + }, + "rollups": { + "professional_task_count": 24, + "domain_count": 8, + "telegram_stage_count": 5, + "telegram_message_type_count": 6, + "approval_required_count": 19, + "low_risk_task_count": 3, + "medium_risk_task_count": 10, + "high_risk_task_count": 6, + "critical_risk_task_count": 5, + "current_live_count": 0, + "gateway_queue_write_count": 0, + "telegram_send_count": 0, + "bot_api_call_count": 0, + "delivery_receipt_write_count": 0, + "production_write_count": 0, + "secret_read_count": 0, + "paid_api_call_count": 0, + "host_write_count": 0, + "kubectl_action_count": 0, + "no_send_preview_count": 6, + "dedup_key_count": 6, + "receipt_expectation_count": 6, + "canary_approval_package_count": 1, + "queue_preview_envelope_field_count": 6, + "canary_required_preview_count": 6, + "approval_required_preview_count": 4, + "preview_send_enabled_count": 0, + "preview_queue_write_enabled_count": 0, + "preview_bot_api_call_enabled_count": 0, + "receipt_live_write_enabled_count": 0, + "canary_live_send_enabled_count": 0 + } +} diff --git a/docs/schemas/ai_agent_professional_task_expansion_v1.schema.json b/docs/schemas/ai_agent_professional_task_expansion_v1.schema.json index c7dedfb34..eaa547a75 100644 --- a/docs/schemas/ai_agent_professional_task_expansion_v1.schema.json +++ b/docs/schemas/ai_agent_professional_task_expansion_v1.schema.json @@ -41,10 +41,10 @@ "enum": ["P0", "P1", "P2", "P3"] }, "current_task_id": { - "const": "P2-405A" + "const": "P2-405B" }, "next_task_id": { - "const": "P2-405B" + "const": "P2-405C" }, "overall_completion_percent": { "type": "integer", @@ -86,7 +86,12 @@ "telegram_send_enabled", "delivery_receipt_write_enabled", "stages", - "message_types" + "message_types", + "no_send_message_previews", + "dedup_policy", + "queue_preview_readback", + "receipt_expectations", + "canary_approval_package" ], "properties": { "canonical_room": { @@ -132,9 +137,189 @@ "message_types": { "type": "array", "minItems": 6, + "maxItems": 6, "items": { "type": "object" } + }, + "no_send_message_previews": { + "type": "array", + "minItems": 6, + "maxItems": 6, + "items": { + "type": "object", + "required": [ + "preview_id", + "message_type", + "title", + "target_room_alias", + "target_room_env", + "owner_agent", + "risk_tier", + "status", + "dedup_key", + "receipt_expectation_id", + "approval_required_before_send", + "canary_required_before_live", + "send_enabled", + "gateway_queue_write_enabled", + "bot_api_call_enabled", + "delivery_receipt_write_enabled", + "sanitized_body_lines", + "redaction_checks" + ], + "properties": { + "target_room_env": { + "const": "SRE_GROUP_CHAT_ID" + }, + "status": { + "const": "preview_ready_no_send" + }, + "send_enabled": { + "const": false + }, + "gateway_queue_write_enabled": { + "const": false + }, + "bot_api_call_enabled": { + "const": false + }, + "delivery_receipt_write_enabled": { + "const": false + }, + "sanitized_body_lines": { + "type": "array", + "minItems": 1, + "items": { + "type": "string" + } + } + }, + "additionalProperties": true + } + }, + "dedup_policy": { + "type": "object", + "required": ["required", "key_count", "live_cache_write_enabled", "keys"], + "properties": { + "required": { + "const": true + }, + "key_count": { + "const": 6 + }, + "live_cache_write_enabled": { + "const": false + }, + "keys": { + "type": "array", + "minItems": 6, + "maxItems": 6 + } + }, + "additionalProperties": true + }, + "queue_preview_readback": { + "type": "object", + "required": ["enabled", "preview_only", "write_enabled", "readback_enabled"], + "properties": { + "enabled": { + "const": true + }, + "preview_only": { + "const": true + }, + "write_enabled": { + "const": false + }, + "readback_enabled": { + "const": true + } + }, + "additionalProperties": true + }, + "receipt_expectations": { + "type": "array", + "minItems": 6, + "maxItems": 6, + "items": { + "type": "object", + "required": [ + "receipt_id", + "message_type", + "preview_id", + "expected_signal", + "required_evidence_refs", + "receipt_write_enabled", + "production_receipt_readback_enabled", + "missing_receipt_policy" + ], + "properties": { + "required_evidence_refs": { + "type": "array", + "minItems": 1 + }, + "receipt_write_enabled": { + "const": false + }, + "production_receipt_readback_enabled": { + "const": false + } + }, + "additionalProperties": true + } + }, + "canary_approval_package": { + "type": "object", + "required": [ + "package_id", + "status", + "package_ready", + "approval_required", + "target_room_alias", + "target_room_env", + "arbiter", + "owner_agent", + "reviewers", + "message_type_scope", + "required_approval_inputs", + "approval_checklist", + "live_send_enabled", + "gateway_queue_write_enabled", + "bot_api_call_enabled", + "delivery_receipt_write_enabled", + "production_write_enabled" + ], + "properties": { + "target_room_env": { + "const": "SRE_GROUP_CHAT_ID" + }, + "status": { + "const": "blocked_until_explicit_approval" + }, + "package_ready": { + "const": true + }, + "approval_required": { + "const": true + }, + "live_send_enabled": { + "const": false + }, + "gateway_queue_write_enabled": { + "const": false + }, + "bot_api_call_enabled": { + "const": false + }, + "delivery_receipt_write_enabled": { + "const": false + }, + "production_write_enabled": { + "const": false + } + }, + "additionalProperties": true } }, "additionalProperties": true @@ -261,7 +446,16 @@ "secret_read_count", "paid_api_call_count", "host_write_count", - "kubectl_action_count" + "kubectl_action_count", + "no_send_preview_count", + "dedup_key_count", + "receipt_expectation_count", + "canary_approval_package_count", + "preview_send_enabled_count", + "preview_queue_write_enabled_count", + "preview_bot_api_call_enabled_count", + "receipt_live_write_enabled_count", + "canary_live_send_enabled_count" ], "properties": { "professional_task_count": { @@ -320,6 +514,33 @@ }, "kubectl_action_count": { "const": 0 + }, + "no_send_preview_count": { + "const": 6 + }, + "dedup_key_count": { + "const": 6 + }, + "receipt_expectation_count": { + "const": 6 + }, + "canary_approval_package_count": { + "const": 1 + }, + "preview_send_enabled_count": { + "const": 0 + }, + "preview_queue_write_enabled_count": { + "const": 0 + }, + "preview_bot_api_call_enabled_count": { + "const": 0 + }, + "receipt_live_write_enabled_count": { + "const": 0 + }, + "canary_live_send_enabled_count": { + "const": 0 } }, "additionalProperties": true diff --git a/docs/superpowers/specs/2026-04-15-MASTER-ai-autonomous-flywheel-v2.md b/docs/superpowers/specs/2026-04-15-MASTER-ai-autonomous-flywheel-v2.md index cb65970fe..b79e74704 100644 --- a/docs/superpowers/specs/2026-04-15-MASTER-ai-autonomous-flywheel-v2.md +++ b/docs/superpowers/specs/2026-04-15-MASTER-ai-autonomous-flywheel-v2.md @@ -638,7 +638,7 @@ Alert / Sentry / SigNoz / Gitea / Market Watch / Operator | `docs/evaluations/ai_agent_report_runtime_dry_run_2026-06-12.json` + `GET /api/v1/agents/agent-report-runtime-dry-run` | P2-403M 報表 runtime no-write dry-run 證據包；建立 5 個 dry-run artifact、3 個 SRE 戰情室 queue digest 草案、4 個 readback verifier case、3 個 Agent dry-run role 與 6 個 operator checkpoint；不寫 Gateway queue、不送 Telegram、不呼叫 Bot API、不寫 delivery receipt、不啟動 worker、不跑 verifier live readback、不讀 secret，已由 P2-403N fixture readback 承接 | | `docs/evaluations/ai_agent_report_runtime_fixture_readback_2026-06-12.json` + `GET /api/v1/agents/agent-report-runtime-fixture-readback` | P2-403N fixture smoke / queue preview readback / verifier dry-run 證據包；建立 5 個 fixture smoke、3 個 SRE 戰情室 queue preview readback、4 個 verifier dry-run case、3 個 Agent fixture role 與 5 個 operator checkpoint；不寫 Gateway queue、不送 Telegram、不呼叫 Bot API、不寫 delivery receipt、不啟動 worker、不跑 verifier live readback、不讀 secret，下一步 P2-404 | | `docs/evaluations/ai_agent_runtime_worker_shadow_gate_2026-06-12.json` + `GET /api/v1/agents/agent-runtime-worker-shadow-gate` | P2-404 runtime worker shadow / no-write execution evidence gate；建立 5 個 shadow candidate、4 個 no-write replay、4 個 verifier shadow case、3 個 Agent shadow role 與 6 個 operator checkpoint；shadow live worker、Gateway queue write、Telegram send、Bot API、delivery receipt、auto worker、verifier live readback、production write 與 secret read 全部 `0 / false`，下一步 P2-101 | -| `docs/evaluations/ai_agent_professional_task_expansion_2026-06-15.json` + `GET /api/v1/agents/agent-professional-task-expansion` | P2-405A AI Agent 專業任務擴展與 Telegram Runtime Bridge；承接 12-Agent War Room、P2-403 report/runtime 鏈與 monitoring owner response acceptance，固定 24 類專業任務、8 個任務領域、5 段 Telegram bridge、6 種訊息類型、MCP/RAG stack、日報 / 週報 / 月報 / action-required 報告契約，並接入 governance automation inventory 卡片；需批准任務 `19`、low / medium / high / critical = `3 / 10 / 6 / 5`；Gateway queue write、Telegram send、Bot API、delivery receipt、production write、secret read、paid API、host write 與 kubectl action 全部 `0 / false`，下一步 P2-405B 完整 no-send message preview / canary approval package | +| `docs/evaluations/ai_agent_professional_task_expansion_2026-06-15_1445_p2_405b.json` + `GET /api/v1/agents/agent-professional-task-expansion` | P2-405B AI Agent 專業任務擴展與 Telegram Runtime Bridge；承接 12-Agent War Room、P2-403 report/runtime 鏈與 monitoring owner response acceptance，固定 24 類專業任務、8 個任務領域、5 段 Telegram bridge、6 種訊息類型、6 個 no-send preview、6 個 dedup key、6 組 receipt expectation、1 份 canary approval package、MCP/RAG stack、日報 / 週報 / 月報 / action-required 報告契約，並接入 governance automation inventory 卡片；需批准任務 `19`、preview live write `0`；Gateway queue write、Telegram send、Bot API、delivery receipt production write、production write、secret read、paid API、host write 與 kubectl action 全部 `0 / false`，下一步 P2-405C approved canary send approval packet | | `docs/evaluations/ai_agent_operation_permission_model_2026-06-12.json` + `GET /api/v1/agents/agent-operation-permission-model` | P2-101 操作類別權限模型；建立 5 條 permission lane、13 類 operation category、3 個 Agent permission role、8 個 gate transition 與 5 個 operator decision template；runtime execution、Gateway queue write、Telegram send、Bot API、delivery receipt、auto worker、verifier live readback、production write、secret / paid provider、host command 與 destructive action 全部 `0 / false`，已由 P2-102 承接 | | `docs/evaluations/ai_agent_candidate_operation_dry_run_evidence_2026-06-12.json` + `GET /api/v1/agents/agent-candidate-operation-dry-run-evidence` | P2-102 候選操作 dry-run 證據；13 類候選操作全部具備 input / output evidence hash、side-effect count、verifier plan、rollback/no-op plan 與人工 handoff；6 個 verifier plan、7 個 gate evidence requirement、5 個 operator handoff；runtime、Gateway queue、Telegram、production write、secret / paid provider 與 destructive action 全部 `0 / false`，已由 P2-103 承接 | | `docs/evaluations/ai_agent_task_result_audit_trail_2026-06-13.json` + `GET /api/v1/agents/agent-task-result-audit-trail` | P2-103 任務結果稽核軌跡；8 條 result route、6 個 writeback contract、7 個 audit checkpoint、5 個 operator handoff；把 diagnostic-only、repair candidate、execution failed、provider unmatched、report zero-signal 等結果固定到 KM 草稿、LOGBOOK 證據、audit trail、timeline 與人工下一步；KM / LOGBOOK / audit DB / timeline / PlayBook trust / Gateway queue / Telegram 寫入全為 `0 / false`，已由 P2-104 承接 | @@ -841,7 +841,7 @@ Repo / registry / release notes / K8s / host / observability / backup evidence 64. 建立 owner response preflight 與拒收邊界。✅ P2-143 已完成正式驗證；承接 P2-141 decision input prep 與 P2-142 War Room 基線，固定 response intake lane `5`、required owner field `18`、intake validation check `6`、rejection guard `6`、operator action `5`、waiting external response `5`；owner response received / accepted / rejected、redacted payload ingested、reviewer queue write、Gateway queue write、Telegram send、Bot API、result capture、learning、PlayBook trust、production write、secret read、destructive operation 仍為 `0 / false`；feature commit `755b0a8d`、deploy marker `667d6329`、Gitea code-review `2961` / CD `2960` success、本地 P2-142 War Room + P2-139 至 P2-143 regression `37 passed`、JSON parse、Python compile、Web typecheck、guard、doc secret sanity、禁用外露值掃描、production API readback 與 desktop / mobile smoke 通過。下一步 P2-144 owner response readback。 65. 建立 owner response readback。✅ P2-144 已完成正式驗證；承接 P2-143 preflight，固定 response readback lane `5`、required owner field `18`、readback validation check `6`、readback rejection guard `6`、operator action `5`、waiting external response `5`、no external response received lane `5`；owner response received / accepted / rejected、redacted payload ingested、reviewer queue write、Gateway queue write、Telegram send、Bot API、result capture、learning、PlayBook trust、production write、secret read、destructive operation 仍為 `0 / false`；feature commit `8795f100`、deploy marker `ac938037`、Gitea code-review `2965` / CD `2964` success；本地 P2-139 至 P2-144 regression `45 passed`，rebase 後含 tenants regression 的推送前回歸 `47 passed`、JSON parse、Python compile、Web typecheck、guard、doc secret sanity、value-only 禁用外露值掃描、i18n key parity 與 diff check 通過；production API readback、desktop / mobile smoke、水平溢位 `0`、危險控制 `0`、工作溝通片語命中 `0` 已完成。下一步 P2-145 owner response acceptance gate。 66. 建立 owner response acceptance gate。✅ P2-145 已完成並正式驗證；承接 P2-144 readback，固定 acceptance gate lane `5`、required owner field `18`、acceptance validation check `6`、acceptance rejection guard `6`、operator action `5`、blocked no external response `5`、no acceptable external response `5`；owner response received / accepted / rejected、redacted payload ingested、reviewer queue write、Gateway queue write、Telegram send、Bot API、result capture、learning、PlayBook trust、production write、secret read、destructive operation 仍為 `0 / false`；feature commit `386dbd07`、deploy marker `36fbfc6b`、Gitea code-review `2969` / CD `2968` success；P2-144 + P2-145 regression `16 passed`、JSON parse、Python compile、Web typecheck、guard、doc secret sanity、顯示值工作視窗污染掃描、i18n key parity、diff check、production API readback、in-app browser smoke、desktop / mobile smoke、水平溢位 `0`、P2-145 卡片操作控制 `0`、工作溝通片語命中 `0` 已完成。下一步 P2-146 acceptance receipt preview，且必須等合格、遮罩、欄位完整、可驗證來源的外部正式回覆後才能建立 receipt preview。 -66a. 建立 AI Agent 專業任務擴展與 Telegram Runtime Bridge。✅ P2-405A 本地完成；承接 12-Agent War Room、P2-403 report/runtime 鏈與 monitoring owner response acceptance，固定 24 類專業任務、8 個領域、5 段 Telegram bridge、6 種訊息類型與 MCP/RAG stack，並接入 governance automation inventory 卡片；需批准任務 `19`、low / medium / high / critical = `3 / 10 / 6 / 5`；Gateway queue write、Telegram send、Bot API、delivery receipt、production write、secret read、paid API、host write、kubectl action 仍為 `0 / false`。下一步 P2-405B 完整 no-send message preview / canary approval package；這不占用 P2-146 owner response receipt preview。 +66a. 建立 AI Agent 專業任務擴展與 Telegram Runtime Bridge。✅ P2-405B 本地完成；承接 12-Agent War Room、P2-403 report/runtime 鏈與 monitoring owner response acceptance，固定 24 類專業任務、8 個領域、5 段 Telegram bridge、6 種訊息類型、6 個 no-send preview、6 個 dedup key、6 組 receipt expectation、1 份 canary approval package 與 MCP/RAG stack，並接入 governance automation inventory 卡片；需批准任務 `19`、low / medium / high / critical = `3 / 10 / 6 / 5`；Gateway queue write、Telegram send、Bot API、delivery receipt production write、production write、secret read、paid API、host write、kubectl action 仍為 `0 / false`。下一步 P2-405C approved canary send approval packet；這不占用 P2-146 owner response receipt preview。 67. 新增 P0 配置控管優先序前台可視化。✅ 正式驗證完成；`/zh-TW/iwooos` 已集中顯示 Nginx public gateway、DNS / TLS / certbot、K8s / ArgoCD / production manifests、Workflow / runner / secret metadata、Public / admin / API runtime config、agent-bounty runtime / treasury 六類即時風險配置；owner response `0 / 0`、live evidence `0`、執行期 `0`、操作按鈕 `0`；feature commit `e992af89`、deploy marker `ed651a98`、Gitea code-review `2971` / CD `2970` success；本地與正式 in-app browser、desktop `1440x1100`、mobile `390x844` smoke 通過。這不是 Nginx live conf 讀取、`nginx -t`、reload、DNS / TLS probe、certbot renew、ArgoCD sync、kubectl、workflow / secret 修改、public route change、agent-bounty runtime、payout / withdrawal、production write 或 runtime gate。 68. 補強 P0 高價值配置 Gate path pattern、工作樹 preflight、owner packet 與 coverage snapshot。✅ 本地完成；`k8s/nginx/**`、`scripts/ops/**/*cert*`、`scripts/ops/**/*tls*` 已納入 high-value config classification，Nginx public gateway 與 DNS / TLS / certbot sample 從 `matched=0 / C0=0` 收斂到 `matched=3 / C0=2`；gate snapshot 顯示 `changed_files=6`、`matched=6`、`categories=3`、`c0=2`、`c1=0`；預設模式已可讀取 staged / unstaged / untracked，臨時 `k8s/nginx/*` smoke 命中 C0；owner packet snapshot `packets=3 / c0=2 / runtime_gate=0`；coverage snapshot `categories=14 / c0=8 / avg=67 / runtime_gate=0`；owner evidence 仍 `provided=false / complete=false`，runtime execution 仍 `false`。這不是 live config read、`nginx -t`、reload、certbot renew、DNS / TLS probe、host write、active scan、workflow 修改、secret 收集、production write 或 runtime gate。 69. 同步高價值配置 Owner Packet 前台 projection。✅ 已完成並正式驗證；`/zh-TW/iwooos` 與 `/zh-TW/awooop` 已顯示 owner packet snapshot `packet=3 / c0=2`、最高命中 `C0 / P0`、Nginx public gateway、DNS / TLS / certbot 與 security tooling 影響範圍；feature commit `e999c16b`、deploy marker `16c6b983`、Gitea code-review `2973` / CD `2972` success；request sent、received、accepted、runtime gate 與 action buttons 仍為 `0`；本地與正式 desktop / mobile / in-app browser smoke 已通過，水平溢位 `0`、卡片內操作控制 `0`、危險連結 `0`、工作溝通片語命中 `0`。不得因此調高 IwoooS headline。 @@ -894,7 +894,7 @@ Repo / registry / release notes / K8s / host / observability / backup evidence | `docs/evaluations/ai_agent_report_runtime_dry_run_2026-06-12.json` + `GET /api/v1/agents/agent-report-runtime-dry-run` | P2-403M 報表 runtime no-write dry-run 證據包；5 個 dry-run artifact、3 個 queue digest 草案、4 個 readback verifier case、3 個 Agent dry-run role、6 個 operator checkpoint；不寫 Gateway queue、不送 Telegram、不呼叫 Bot API、不寫 delivery receipt、不啟動 worker、不跑 verifier live readback | | `docs/evaluations/ai_agent_report_runtime_fixture_readback_2026-06-12.json` + `GET /api/v1/agents/agent-report-runtime-fixture-readback` | P2-403N fixture smoke / queue preview readback / verifier dry-run 證據包；5 個 fixture smoke、3 個 queue preview readback、4 個 verifier dry-run case、3 個 Agent fixture role、5 個 operator checkpoint；不寫 Gateway queue、不送 Telegram、不呼叫 Bot API、不寫 delivery receipt、不啟動 worker、不跑 verifier live readback | | `docs/evaluations/ai_agent_runtime_worker_shadow_gate_2026-06-12.json` + `GET /api/v1/agents/agent-runtime-worker-shadow-gate` | P2-404 runtime worker shadow / no-write execution evidence gate；5 個 shadow candidate、4 個 no-write replay、4 個 verifier shadow case、3 個 Agent shadow role、6 個 operator checkpoint；不啟動 live worker、不寫 Gateway queue、不送 Telegram、不呼叫 Bot API、不寫 production target | -| `docs/evaluations/ai_agent_professional_task_expansion_2026-06-15.json` + `GET /api/v1/agents/agent-professional-task-expansion` | P2-405A AI Agent 專業任務擴展與 Telegram Runtime Bridge；24 類專業任務、8 個領域、5 段 Telegram bridge、6 種訊息類型、MCP/RAG stack、治理頁 P2-405A 卡片；不寫 Gateway queue、不送 Telegram、不呼叫 Bot API、不寫 production target、不讀 secret、不用 paid API | +| `docs/evaluations/ai_agent_professional_task_expansion_2026-06-15_1445_p2_405b.json` + `GET /api/v1/agents/agent-professional-task-expansion` | P2-405B AI Agent 專業任務擴展與 Telegram Runtime Bridge；24 類專業任務、8 個領域、5 段 Telegram bridge、6 種訊息類型、6 個 no-send preview、6 個 dedup key、6 組 receipt expectation、1 份 canary approval package、MCP/RAG stack、治理頁 P2-405B 卡片；不寫 Gateway queue、不送 Telegram、不呼叫 Bot API、不寫 production target、不讀 secret、不用 paid API | | `docs/evaluations/ai_agent_operation_permission_model_2026-06-12.json` + `GET /api/v1/agents/agent-operation-permission-model` | P2-101 操作類別權限模型；5 條 permission lane、13 類 operation category、3 個 Agent permission role、8 個 gate transition、5 個 operator decision template；不啟動 runtime worker、不寫 Gateway queue、不送 Telegram、不呼叫 Bot API、不寫 production target、不讀 secret | | `docs/evaluations/ai_agent_candidate_operation_dry_run_evidence_2026-06-12.json` + `GET /api/v1/agents/agent-candidate-operation-dry-run-evidence` | P2-102 候選操作 dry-run 證據；13 類候選操作、13 組 dry-run evidence、6 個 verifier plan、7 個 gate evidence requirement、5 個 operator handoff；不啟動 runtime worker、不寫 Gateway queue、不送 Telegram、不呼叫 Bot API、不寫 production target、不讀 secret、不執行 destructive action | | `docs/evaluations/ai_agent_task_result_audit_trail_2026-06-13.json` + `GET /api/v1/agents/agent-task-result-audit-trail` | P2-103 任務結果稽核軌跡；8 條 result route、6 個 writeback contract、7 個 audit checkpoint、5 個 operator handoff；不寫 KM、不 runtime append LOGBOOK、不寫 audit DB、不寫 timeline、不更新 PlayBook trust、不寫 Gateway queue、不送 Telegram | @@ -4993,3 +4993,15 @@ Trigger commit `f5cd37b7` 與 deploy marker `0ba92357` 已把 governance UI 的 - 任務風險 rollup 固定：總任務 `24`、需批准 `19`、low / medium / high / critical = `3 / 10 / 6 / 5`；高風險與 critical 任務必須保留 approval gate。 **裁決：** 這是專業任務與 Telegram 整合的只讀擴展契約，不是 Telegram send、Gateway queue write、Bot API call、delivery receipt write、production write、host write、kubectl action、secret read、paid API call 或 runtime authorization；下一步 P2-405B 只能做完整 no-send message preview 與 canary approval package，可讓統帥看到將來 TG 訊息長相，但仍不得實發。 + +### 2026-06-15 14:45 (台北) — §3.2 / §5 — 新增 P2-405B Telegram no-send preview / dedup / receipt / canary approval package — 讓 TG 訊息長相可審核但不實發 + +**觸發**：P2-405A 已部署到正式 API，統帥要求繼續推進 Telegram 群組 / TG Bot 整合。下一步必須讓治理頁看見未來 Telegram 訊息長相、dedup key、receipt expectation 與 canary approval package，同時保留 no-send guard，避免把 UI 可見誤判成實發授權。 + +**已推進：** +- 新增 `docs/evaluations/ai_agent_professional_task_expansion_2026-06-15_1445_p2_405b.json`，`current_task_id=P2-405B`、`next_task_id=P2-405C`、overall `88%`。 +- `ai_agent_professional_task_expansion_v1` schema 與 API loader 已要求 6 個 no-send preview、6 個 dedup key、6 組 receipt expectation、1 份 canary approval package。 +- `/zh-TW/governance?tab=automation-inventory` P2-405B 卡片顯示 Telegram no-send 訊息預覽、dedup、receipt、canary approval checklist 與 preview live write `0`。 +- 新增 / 更新測試，明確拒絕 preview send enabled、duplicate dedup key、canary live send enabled、queue write、Bot API call 或 receipt production write。 + +**裁決：** 這是 Telegram no-send preview 與 canary approval package，不是 Telegram send、Gateway queue write、Bot API call、delivery receipt production write、approved canary、production write、secret read、host write、kubectl action 或 runtime authorization；下一步 P2-405C 只能產生 approved canary send approval packet 與人工批准欄位，未批准前不得實發。