@tea-agent/loop-agent 0.2.1 → 0.4.0

package/docs/templates/agent-dag-decision-gate-dogfood-report.md

@@ -1,117 +1,117 @@
-# Agent DAG Decision Gate Dogfood Report Template
-
-> Copy to `docs/reports/YYYY-MM-DD-agent-dag-decision-gate-dogfood.md` after each dogfood run.
-
-## Run Metadata
-
-| Field | Value |
-|---|---|
-| Date | YYYY-MM-DD |
-| Topic |  |
-| DAG input | `<temp-dir>/<topic>-decision-gate-dag.json` |
-| Run ID |  |
-| Run facts | `.harness/dag-runs/completed/<run-id>/` |
-| Gate type | `acceptance-gate` |
-| Decision node | `decision-pi` |
-| Model | `gpt-5.5` via `executorModels.pi.HIGH` |
-
-## Work Type
-
-Choose one:
-
-- [ ] Low-risk docs/template change — expected `auto-approve` or `approve-with-constraints`
-- [ ] Runtime/loop-agent change — expected `request-revision`, `run-more-verification`, or `auto-approve`
-- [ ] Contract/security/high-risk simulation — expected `escalate-to-human`
-
-## Deterministic Evidence
-
-| Evidence | Status | Notes |
-|---|---|---|
-| `verify-shell/result.summary.md` | verified / partial / missing |  |
-| `state.json` | verified / partial / missing |  |
-| git diff / changed file list | verified / partial / missing |  |
-| progress/report/artifacts | verified / partial / missing |  |
-
-## Decision Envelope Summary
-
-Paste the **exact** ` ```DECISION_ENVELOPE_JSON ` fenced block from `decision-pi/result.summary.md` (info string must be `DECISION_ENVELOPE_JSON`, not `json`). Also record `decision`, `requiresHuman`, `nextAction`, and verified evidence count.
-
-```DECISION_ENVELOPE_JSON
-{
-  "schemaVersion": 1,
-  "gateType": "acceptance-gate",
-  "decisionScope": "dag-run",
-  "decision": "auto-approve",
-  "confidence": 0.88,
-  "riskLevel": "low",
-  "requiresHuman": false,
-  "nextAction": "continue",
-  "policyVersion": "agent-dag-decision-gate-v1",
-  "policyChecks": {
-    "mustEscalateFlags": [],
-    "evidenceComplete": true,
-    "allowedAutoApprove": true
-  },
-  "rationale": ["..."],
-  "evidence": [
-    {
-      "path": ".harness/dag-runs/completed/<run-id>/verify-shell/result.summary.md",
-      "kind": "shell-output",
-      "status": "verified",
-      "summary": "verification commands passed"
-    }
-  ],
-  "blockingFindings": [],
-  "requiredRevisions": [],
-  "riskFlags": [],
-  "humanEscalation": null,
-  "audit": {
-    "runId": "<run-id>",
-    "nodeId": "decision-pi",
-    "model": "gpt-5.5"
-  }
-}
-```
-
-## Quality Metrics
-
-| Metric | Value | Notes |
-|---|---:|---|
-| `decisionLatencyMs` |  | From node duration |
-| `tokenCostEstimate` |  | If available |
-| `humanInterruptCount` |  | Should be 0 for low-risk auto decisions |
-| `falseEscalation` | yes / no | Escalated when policy allowed auto decision |
-| `missedEscalation` | yes / no | Auto-approved when policy required human escalation |
-| `revisionCaughtBeforeHuman` | yes / no | Gate requested revision before human involvement |
-| `verifyPassAfterGate` | yes / no | Later verification passed after gate action |
-| `evidenceCompleteness` | complete / partial / missing | Based on verified evidence count |
-
-## Evaluation
-
-### What the gate got right
-
-- 
-
-### What the gate got wrong or over/under-weighted
-
-- 
-
-### Prompt / policy adjustments needed
-
-- 
-
-## Recommendation
-
-Choose one:
-
-- [ ] `proceed` — enough evidence to move toward parser/runtime work
-- [ ] `needs-more-dogfood` — continue M1/M2 advisory runs
-- [ ] `defer` — gate quality not yet good enough
-
-Rationale:
-
-- 
-
-## Follow-ups
-
-- 
+# Agent DAG Decision Gate Dogfood Report Template
+
+> Copy to `docs/reports/YYYY-MM-DD-agent-dag-decision-gate-dogfood.md` after each dogfood run.
+
+## Run Metadata
+
+| Field | Value |
+|---|---|
+| Date | YYYY-MM-DD |
+| Topic |  |
+| DAG input | `<temp-dir>/<topic>-decision-gate-dag.json` |
+| Run ID |  |
+| Run facts | `.harness/dag-runs/completed/<run-id>/` |
+| Gate type | `acceptance-gate` |
+| Decision node | `decision-pi` |
+| Model | `gpt-5.5` via `executorModels.pi.HIGH` |
+
+## Work Type
+
+Choose one:
+
+- [ ] Low-risk docs/template change — expected `auto-approve` or `approve-with-constraints`
+- [ ] Runtime/loop-agent change — expected `request-revision`, `run-more-verification`, or `auto-approve`
+- [ ] Contract/security/high-risk simulation — expected `escalate-to-human`
+
+## Deterministic Evidence
+
+| Evidence | Status | Notes |
+|---|---|---|
+| `verify-shell/result.summary.md` | verified / partial / missing |  |
+| `state.json` | verified / partial / missing |  |
+| git diff / changed file list | verified / partial / missing |  |
+| progress/report/artifacts | verified / partial / missing |  |
+
+## Decision Envelope Summary
+
+Paste the **exact** ` ```DECISION_ENVELOPE_JSON ` fenced block from `decision-pi/result.summary.md` (info string must be `DECISION_ENVELOPE_JSON`, not `json`). Also record `decision`, `requiresHuman`, `nextAction`, and verified evidence count.
+
+```DECISION_ENVELOPE_JSON
+{
+  "schemaVersion": 1,
+  "gateType": "acceptance-gate",
+  "decisionScope": "dag-run",
+  "decision": "auto-approve",
+  "confidence": 0.88,
+  "riskLevel": "low",
+  "requiresHuman": false,
+  "nextAction": "continue",
+  "policyVersion": "agent-dag-decision-gate-v1",
+  "policyChecks": {
+    "mustEscalateFlags": [],
+    "evidenceComplete": true,
+    "allowedAutoApprove": true
+  },
+  "rationale": ["..."],
+  "evidence": [
+    {
+      "path": ".harness/dag-runs/completed/<run-id>/verify-shell/result.summary.md",
+      "kind": "shell-output",
+      "status": "verified",
+      "summary": "verification commands passed"
+    }
+  ],
+  "blockingFindings": [],
+  "requiredRevisions": [],
+  "riskFlags": [],
+  "humanEscalation": null,
+  "audit": {
+    "runId": "<run-id>",
+    "nodeId": "decision-pi",
+    "model": "gpt-5.5"
+  }
+}
+```
+
+## Quality Metrics
+
+| Metric | Value | Notes |
+|---|---:|---|
+| `decisionLatencyMs` |  | From node duration |
+| `tokenCostEstimate` |  | If available |
+| `humanInterruptCount` |  | Should be 0 for low-risk auto decisions |
+| `falseEscalation` | yes / no | Escalated when policy allowed auto decision |
+| `missedEscalation` | yes / no | Auto-approved when policy required human escalation |
+| `revisionCaughtBeforeHuman` | yes / no | Gate requested revision before human involvement |
+| `verifyPassAfterGate` | yes / no | Later verification passed after gate action |
+| `evidenceCompleteness` | complete / partial / missing | Based on verified evidence count |
+
+## Evaluation
+
+### What the gate got right
+
+- 
+
+### What the gate got wrong or over/under-weighted
+
+- 
+
+### Prompt / policy adjustments needed
+
+- 
+
+## Recommendation
+
+Choose one:
+
+- [ ] `proceed` — enough evidence to move toward parser/runtime work
+- [ ] `needs-more-dogfood` — continue M1/M2 advisory runs
+- [ ] `defer` — gate quality not yet good enough
+
+Rationale:
+
+- 
+
+## Follow-ups
+
+-