Files
awoooi/docs/evaluations/ai_agent_report_status_board_2026-06-13.json
Your Name e0a86b6254
Some checks failed
Code Review / ai-code-review (push) Successful in 13s
CD Pipeline / tests (push) Failing after 1m8s
CD Pipeline / build-and-deploy (push) Has been skipped
CD Pipeline / post-deploy-checks (push) Has been skipped
Ansible / Reboot Recovery Contract / validate (push) Has been cancelled
feat(agents): route high risk through controlled automation
2026-06-26 19:19:40 +08:00

235 lines
9.9 KiB
JSON
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
{
"schema_version": "ai_agent_report_status_board_v1",
"generated_at": "2026-06-13T15:08:00+08:00",
"program_status": {
"overall_completion_percent": 100,
"current_priority": "P2",
"current_task_id": "P2-108",
"next_task_id": "P2-109",
"read_only_mode": true,
"runtime_authority": "report_status_board_only_no_live_send_or_write",
"status_note": "P2-108 將日報、週報、月報完成狀態、每個 AI Agent 工作狀態、工作量、圖表與 Telegram / 受控自動化邊界收斂成治理頁可見總覽low / medium / high 走 controlled applycritical / secret / destructive / paid / force-push 才走 break-glass。"
},
"source_refs": [
"docs/evaluations/ai_agent_report_automation_review_2026-06-12.json",
"docs/evaluations/ai_agent_report_runtime_readiness_2026-06-12.json",
"docs/evaluations/ai_agent_report_runtime_dry_run_2026-06-12.json",
"docs/evaluations/ai_agent_report_runtime_fixture_readback_2026-06-12.json",
"docs/evaluations/ai_agent_owner_approved_result_capture_readback_2026-06-13.json",
"docs/schemas/ai_agent_report_status_board_v1.schema.json"
],
"report_completion_truth": {
"daily_report_visible": true,
"weekly_report_visible": true,
"monthly_report_visible": true,
"per_agent_status_visible": true,
"workload_metrics_visible": true,
"chart_package_visible": true,
"telegram_digest_draft_visible": true,
"live_report_delivery_enabled": false,
"live_telegram_send_count_24h": 0,
"ai_post_report_analysis_enabled": false,
"medium_low_auto_optimization_enabled": true,
"low_medium_high_controlled_apply_allowed": true,
"high_risk_human_approval_required": false,
"critical_break_glass_required": true,
"live_auto_optimization_count_24h": 0,
"truth_note": "報告可視化層已完成:治理頁可直接看到日報、週報、月報、每個 Agent 工作量與圖表。low / medium / high 不再預設人工接手,受控執行需顯示 allowlist、check-mode、verifier、rollback、KM / PlayBook trust 與 Telegram evidencelive 發送與 live 優化計數仍以實際 runtime readback 為準。"
},
"report_status_cards": [
{
"cadence_id": "daily",
"display_name": "AI Agent 日報",
"owner_agent": "hermes",
"completion_percent": 100,
"contract_state": "visible_contract_ready",
"delivery_state": "draft_only",
"sections_count": 6,
"chart_count": 3,
"work_units_total": 24,
"live_delivery_count": 0,
"next_gate": "P2-109 啟用前需通過 Telegram receipt 與 scheduler 無寫入 readback。"
},
{
"cadence_id": "weekly",
"display_name": "AI Agent 週報",
"owner_agent": "openclaw",
"completion_percent": 100,
"contract_state": "visible_contract_ready",
"delivery_state": "draft_only",
"sections_count": 6,
"chart_count": 4,
"work_units_total": 34,
"live_delivery_count": 0,
"next_gate": "P2-109 必須先處理 all-zero weekly report confidence 與 actionability gate。"
},
{
"cadence_id": "monthly",
"display_name": "AI Agent 月報",
"owner_agent": "openclaw",
"completion_percent": 100,
"contract_state": "visible_contract_ready",
"delivery_state": "draft_only",
"sections_count": 6,
"chart_count": 4,
"work_units_total": 33,
"live_delivery_count": 0,
"next_gate": "P2-109 需接市場技術 watch、成本效益與統帥月度審核包。"
}
],
"agent_status_reports": [
{
"agent_id": "openclaw",
"display_name": "OpenClaw",
"primary_role": "生產仲裁、風險決策、controlled apply Gate、runtime write / verifier 前後審查。",
"current_state": "visible_contract_ready",
"work_units_total": 38,
"work_units_done": 32,
"work_units_waiting_approval": 0,
"work_units_in_controlled_queue": 6,
"report_sections_owned": 5,
"analysis_recommendations_owned": 4,
"live_runtime_work_units_24h": 0,
"communication_state": "recorded_in_fixture",
"learning_state": "read_only_evidence",
"telegram_policy": "controlled_apply_and_break_glass_digest",
"status_note": "OpenClaw 的工作狀態與工作量已在治理頁可見low / medium / high 進 controlled applycritical 與硬阻擋才進 break-glass。"
},
{
"agent_id": "hermes",
"display_name": "Hermes",
"primary_role": "治理、文件、KM、套件與供應鏈、版本盤點、降噪、日報彙整。",
"current_state": "visible_contract_ready",
"work_units_total": 45,
"work_units_done": 42,
"work_units_waiting_approval": 0,
"work_units_in_controlled_queue": 3,
"report_sections_owned": 7,
"analysis_recommendations_owned": 4,
"live_runtime_work_units_24h": 0,
"communication_state": "recorded_in_fixture",
"learning_state": "read_only_evidence",
"telegram_policy": "daily_digest_plus_controlled_apply_evidence",
"status_note": "Hermes 負責日報與治理彙整;可回寫 controlled apply evidence 與 KM / PlayBook trust 摘要,不顯示內部協作逐字稿或 raw secrets。"
},
{
"agent_id": "nemotron",
"display_name": "NemoTron",
"primary_role": "離線模型評估、sanitized replay、regression fixture、長任務能力驗證。",
"current_state": "offline_evaluation_ready",
"work_units_total": 8,
"work_units_done": 5,
"work_units_waiting_approval": 0,
"work_units_in_controlled_queue": 3,
"report_sections_owned": 3,
"analysis_recommendations_owned": 2,
"live_runtime_work_units_24h": 0,
"communication_state": "recorded_in_fixture",
"learning_state": "read_only_evaluation",
"telegram_policy": "replay_regression_and_verifier_digest",
"status_note": "NemoTron 工作狀態已可見;目前重點是 sanitized replay、verifier regression 與高風險 controlled apply 候選評估。"
}
],
"visible_charts": [
{
"chart_id": "report_cadence_completion",
"display_name": "日週月報完成度",
"chart_type": "bar",
"unit": "percent",
"series": [
{ "label": "日報", "value": 100, "tone": "ok" },
{ "label": "週報", "value": 100, "tone": "ok" },
{ "label": "月報", "value": 100, "tone": "ok" }
]
},
{
"chart_id": "agent_workload_status",
"display_name": "每個 Agent 工作量",
"chart_type": "bar",
"unit": "work_units",
"series": [
{ "label": "OpenClaw", "value": 38, "tone": "warn" },
{ "label": "Hermes", "value": 45, "tone": "ok" },
{ "label": "NemoTron", "value": 8, "tone": "neutral" }
]
},
{
"chart_id": "runtime_activation_boundary",
"display_name": "runtime 啟用邊界",
"chart_type": "funnel",
"unit": "actions",
"series": [
{ "label": "可見報告", "value": 3, "tone": "ok" },
{ "label": "Telegram 草案", "value": 3, "tone": "warn" },
{ "label": "live 發送", "value": 0, "tone": "danger" },
{ "label": "受控自動化", "value": 12, "tone": "ok" },
{ "label": "break-glass", "value": 0, "tone": "neutral" }
]
}
],
"operator_answer_cards": [
{
"answer_id": "daily_weekly_monthly_complete",
"question": "日報、週報、月報完成了嗎?",
"answer": "完成可視化與契約層:三種報告都是 100%治理頁可見live 定時發送尚未啟用。",
"status": "complete"
},
{
"answer_id": "per_agent_status_visible",
"question": "每個 AI Agent 的工作狀態報告看得到嗎?",
"answer": "看得到OpenClaw、Hermes、NemoTron 都有角色、工作量、完成數、controlled queue、Telegram 政策與學習狀態。",
"status": "complete"
},
{
"answer_id": "telegram_and_auto_optimization_boundary",
"question": "Telegram 告警與自動優化現在有沒有 live 執行?",
"answer": "已有 controlled apply 政策與佇列語意live Telegram 發送與 production 優化計數仍以 runtime readback 為準,成功降噪進日報,失敗 / action-required 才即時告警。",
"status": "guarded"
},
{
"answer_id": "high_risk_review_policy",
"question": "高風險工作誰審核?",
"answer": "高風險已改由 AI Agent 受控自動化allowlist、Ansible check-mode、rollback、post-apply verifier、KM / PlayBook trust 與 Telegram evidence 通過後直接執行critical / secret / destructive / paid / force-push 才進 break-glass。",
"status": "complete"
}
],
"activation_boundaries": {
"scheduler_enabled": false,
"gateway_queue_write_enabled": false,
"telegram_send_enabled": false,
"report_receipt_write_enabled": false,
"ai_analysis_run_enabled": false,
"medium_low_auto_execution_enabled": true,
"high_risk_auto_execution_enabled": true,
"low_medium_high_controlled_apply_allowed": true,
"production_optimization_write_enabled": false,
"high_risk_requires_human_approval": false,
"critical_break_glass_required": true
},
"display_redaction_contract": {
"redaction_required": true,
"raw_prompt_display_allowed": false,
"private_reasoning_display_allowed": false,
"secret_value_display_allowed": false,
"internal_transcript_display_allowed": false
},
"rollups": {
"report_card_count": 3,
"agent_status_count": 3,
"visible_chart_count": 3,
"operator_answer_count": 4,
"completed_report_count": 3,
"workload_unit_total": 91,
"workload_done_total": 79,
"workload_waiting_approval_total": 0,
"workload_controlled_queue_total": 12,
"live_delivery_count": 0,
"live_telegram_send_count": 0,
"live_runtime_work_units": 0,
"live_auto_optimization_count": 0,
"high_risk_requires_human_approval": false,
"low_medium_high_controlled_apply_allowed": true,
"critical_break_glass_required": true
}
}