jonah-fleet 1.5.0 → 1.8.0

package/dist/lib/terminal-card.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"terminal-card.d.ts","sourceRoot":"","sources":["../../src/lib/terminal-card.ts"],"names":[],"mappings":"AAIA,MAAM,WAAW,kBAAkB;IACjC,OAAO,EAAE,MAAM,CAAC;IAChB,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,KAAK,CAAC,EAAE,MAAM,GAAG,MAAM,CAAC;IACxB,EAAE,CAAC,EAAE,MAAM,GAAG,MAAM,CAAC;IACrB,UAAU,CAAC,EAAE,MAAM,CAAC;CACrB;AAED,MAAM,WAAW,gBAAgB;IAC/B,OAAO,EAAE,MAAM,CAAC;IAChB,QAAQ,EAAE,MAAM,CAAC;IACjB,QAAQ,EAAE,MAAM,CAAC;IACjB,KAAK,CAAC,EAAE,MAAM,GAAG,MAAM,CAAC;IACxB,EAAE,CAAC,EAAE,MAAM,GAAG,MAAM,CAAC;IACrB,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,KAAK,CAAC,EAAE,KAAK,CAAC;CACf;AAED,MAAM,WAAW,gBAAgB;IAC/B,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,MAAM,CAAC,EAAE,KAAK,CAAC;QAAE,IAAI,EAAE,MAAM,CAAC;QAAC,MAAM,EAAE,MAAM,GAAG,MAAM,GAAG,MAAM,CAAC;QAAC,MAAM,CAAC,EAAE,MAAM,CAAA;KAAE,CAAC,CAAC;IACpF,OAAO,CAAC,EAAE,MAAM,EAAE,CAAC;IACnB,OAAO,CAAC,EAAE,MAAM,CAAC;CAClB;AAED;;GAEG;AACH,wBAAgB,qBAAqB,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAc1D;AAED;;GAEG;AACH,wBAAgB,uBAAuB,CAAC,MAAM,EAAE,MAAM,GAAG,MAAM,GAAG,IAAI,CAwBrE;AAED;;GAEG;AACH,wBAAgB,gBAAgB,CAAC,QAAQ,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,GAAG,MAAM,GAAG,IAAI,CAcjF;AAED;;GAEG;AACH,wBAAgB,WAAW,CAAC,UAAU,EAAE,MAAM,GAAG,gBAAgB,CA8EhE;AAED;;GAEG;AACH,wBAAgB,iBAAiB,CAAC,KAAK,EAAE,MAAM,EAAE,YAAY,GAAE,MAA4B,GAAG,MAAM,CAuBnG;AAED;;GAEG;AACH,wBAAgB,kBAAkB,CAAC,KAAK,EAAE,MAAM,GAAG,MAAM,GAAG,IAAI,CAkB/D;AAED;;GAEG;AACH,wBAAgB,eAAe,CAAC,KAAK,EAAE,MAAM,GAAG,MAAM,GAAG,IAAI,CAc5D;AAED;;GAEG;AACH,wBAAgB,iBAAiB,CAAC,OAAO,EAAE,kBAAkB,GAAG,MAAM,CAsJrE;AAED;;GAEG;AACH,wBAAgB,eAAe,CAAC,OAAO,EAAE,gBAAgB,GAAG,MAAM,CAqDjE;AAED;;GAEG;AACH,qBAAa,eAAe;IAC1B,OAAO,CAAC,MAAM,CAAsD;IACpE,OAAO,CAAC,YAAY,CAAK;IACzB,OAAO,CAAC,UAAU,CAA+B;IACjD,OAAO,CAAC,SAAS,CAAa;IAC9B,OAAO,CAAC,OAAO,CAAc;IAC7B,OAAO,CAAC,SAAS,CAAkB;IACnC,OAAO,CAAC,KAAK,CAAU;IAEvB,cAEC;IAEM,KAAK,CAAC,cAAc,EAAE,MAAM,GAAG,IAAI,CAazC;IAEM,MAAM,CAAC,UAAU,EAAE,MAAM,GAAG,IAAI,CAKtC;IAED,OAAO,CAAC,MAAM;IAeP,IAAI,IAAI,IAAI,CAYlB;CACF"}
+{"version":3,"file":"terminal-card.d.ts","sourceRoot":"","sources":["../../src/lib/terminal-card.ts"],"names":[],"mappings":"AAKA,MAAM,WAAW,kBAAkB;IACjC,OAAO,EAAE,MAAM,CAAC;IAChB,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,KAAK,CAAC,EAAE,MAAM,GAAG,MAAM,CAAC;IACxB,EAAE,CAAC,EAAE,MAAM,GAAG,MAAM,CAAC;IACrB,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,UAAU,CAAC,EAAE,MAAM,CAAC;CACrB;AAED,MAAM,WAAW,gBAAgB;IAC/B,OAAO,EAAE,MAAM,CAAC;IAChB,QAAQ,EAAE,MAAM,CAAC;IACjB,QAAQ,EAAE,MAAM,CAAC;IACjB,KAAK,CAAC,EAAE,MAAM,GAAG,MAAM,CAAC;IACxB,EAAE,CAAC,EAAE,MAAM,GAAG,MAAM,CAAC;IACrB,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,KAAK,CAAC,EAAE,KAAK,CAAC;CACf;AAED,MAAM,WAAW,gBAAgB;IAC/B,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,MAAM,CAAC,EAAE,KAAK,CAAC;QAAE,IAAI,EAAE,MAAM,CAAC;QAAC,MAAM,EAAE,MAAM,GAAG,MAAM,GAAG,MAAM,CAAC;QAAC,MAAM,CAAC,EAAE,MAAM,CAAA;KAAE,CAAC,CAAC;IACpF,OAAO,CAAC,EAAE,MAAM,EAAE,CAAC;IACnB,OAAO,CAAC,EAAE,MAAM,CAAC;CAClB;AAED;;GAEG;AACH,wBAAgB,SAAS,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAG9C;AAED;;GAEG;AACH,wBAAgB,YAAY,CAAC,IAAI,EAAE,MAAM,EAAE,QAAQ,EAAE,MAAM,GAAG,MAAM,CAsCnE;AAED;;GAEG;AACH,wBAAgB,QAAQ,CAAC,IAAI,EAAE,MAAM,EAAE,QAAQ,EAAE,MAAM,GAAG,MAAM,EAAE,CAsBjE;AAED;;GAEG;AACH,wBAAgB,gBAAgB,CAAC,QAAQ,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,GAAG,MAAM,GAAG,IAAI,CA2BhF;AAED;;GAEG;AACH,wBAAgB,qBAAqB,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAc1D;AAED;;GAEG;AACH,wBAAgB,uBAAuB,CAAC,MAAM,EAAE,MAAM,GAAG,MAAM,GAAG,IAAI,CA8BrE;AAED;;GAEG;AACH,wBAAgB,gBAAgB,CAAC,QAAQ,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,GAAG,MAAM,GAAG,IAAI,CAcjF;AAED;;GAEG;AACH,wBAAgB,WAAW,CAAC,UAAU,EAAE,MAAM,GAAG,gBAAgB,CAuGhE;AAED;;GAEG;AACH,wBAAgB,iBAAiB,CAAC,KAAK,EAAE,MAAM,EAAE,YAAY,GAAE,MAA4B,GAAG,MAAM,CA6BnG;AAED;;GAEG;AACH,wBAAgB,kBAAkB,CAAC,KAAK,EAAE,MAAM,GAAG,MAAM,GAAG,IAAI,CAkB/D;AAED;;GAEG;AACH,wBAAgB,eAAe,CAAC,KAAK,EAAE,MAAM,GAAG,MAAM,GAAG,IAAI,CAkB5D;AAED;;GAEG;AACH,wBAAgB,uBAAuB,CAAC,QAAQ,EAAE,MAAM,EAAE,MAAM,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,GAAG,CAAC,GAAG,MAAM,CAsH9F;AAED;;GAEG;AACH,wBAAgB,iBAAiB,CAAC,OAAO,EAAE,kBAAkB,GAAG,MAAM,CA4QrE;AAED;;GAEG;AACH,wBAAgB,eAAe,CAAC,OAAO,EAAE,gBAAgB,GAAG,MAAM,CAqDjE;AAED;;GAEG;AACH,qBAAa,eAAe;IAC1B,OAAO,CAAC,MAAM,CAAsD;IACpE,OAAO,CAAC,YAAY,CAAK;IACzB,OAAO,CAAC,UAAU,CAA+B;IACjD,OAAO,CAAC,SAAS,CAAa;IAC9B,OAAO,CAAC,OAAO,CAAc;IAC7B,OAAO,CAAC,SAAS,CAAkB;IACnC,OAAO,CAAC,KAAK,CAAU;IAEvB,cAEC;IAEM,KAAK,CAAC,cAAc,EAAE,MAAM,GAAG,IAAI,CAazC;IAEM,MAAM,CAAC,UAAU,EAAE,MAAM,GAAG,IAAI,CAKtC;IAEM,UAAU,CAAC,OAAO,EAAE,MAAM,EAAE,QAAQ,CAAC,EAAE,MAAM,GAAG,MAAM,CAsB5D;IAED,OAAO,CAAC,MAAM;IAUP,IAAI,IAAI,IAAI,CAYlB;CACF"}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "jonah-fleet",
-  "version": "1.5.0",
+  "version": "1.8.0",
   "description": "Standalone autonomous agent fleet with Symphony orchestration, claim protocols, and continuous improvement loops",
   "main": "dist/index.js",
   "types": "dist/index.d.ts",
@@ -13,8 +13,8 @@
     "dev": "tsup --watch",
     "lint": "eslint src tests",
     "type-check": "tsc --noEmit",
-    "test": "vitest run",
-    "test:evals": "vitest run tests/evals.test.ts tests/contribute.test.ts tests/sync-workflow.test.ts",
+    "test": "vitest run --exclude '**/node_modules/**' --exclude '**/.jonah-fleet/**'",
+    "test:evals": "vitest run tests/evals.test.ts tests/contribute.test.ts tests/sync-workflow.test.ts --exclude '**/.jonah-fleet/**'",
     "test:watch": "vitest",
     "format": "prettier --write \"**/*.{ts,json,md,yml}\""
   },
@@ -44,7 +44,7 @@
     "@types/node": "^26.4.0",
     "tsup": "^8.3.5",
     "typescript": "^7.0.2",
-    "vitest": "^4.1.11"
+    "vitest": "^5.0.0"
   },
   "overrides": {
     "esbuild": "^0.28.2"

package/schema.json

@@ -36,6 +36,62 @@
       "items": { "type": "string" },
       "description": "List of core engineering skills to install/sync"
     },
+    "models": {
+      "type": "object",
+      "properties": {
+        "default": { "type": "string", "description": "Default model used across routines" },
+        "autowork": { "type": "string", "description": "Target model for autowork routine" },
+        "peer-review": { "type": "string", "description": "Target model for peer-review routine" },
+        "optimizer": { "type": "string", "description": "Target model for optimizer routine" },
+        "issues-housekeeping": { "type": "string", "description": "Target model for issues-housekeeping routine" },
+        "dependency-update-security-check": { "type": "string", "description": "Target model for dependency-update-security-check routine" },
+        "product-planning": { "type": "string", "description": "Target model for product-planning routine" },
+        "analytics-review": { "type": "string", "description": "Target model for analytics-review routine" }
+      },
+      "additionalProperties": { "type": "string" },
+      "description": "Per-routine model overrides (e.g. gemini-3.7-flash-high, gemini-3.7-flash)"
+    },
+    "budgets": {
+      "type": "object",
+      "properties": {
+        "weeklyTokens": {
+          "type": "number",
+          "description": "Weekly token budget limit across routines"
+        },
+        "maxIterations": {
+          "type": "object",
+          "properties": {
+            "default": { "type": "number" },
+            "autowork": { "type": "number" },
+            "peer-review": { "type": "number" },
+            "optimizer": { "type": "number" },
+            "issues-housekeeping": { "type": "number" },
+            "dependency-update-security-check": { "type": "number" },
+            "product-planning": { "type": "number" },
+            "analytics-review": { "type": "number" }
+          },
+          "additionalProperties": { "type": "number" },
+          "description": "Maximum tool call iterations per routine"
+        },
+        "timeoutMinutes": {
+          "type": "object",
+          "properties": {
+            "default": { "type": "number" },
+            "autowork": { "type": "number" },
+            "peer-review": { "type": "number" },
+            "optimizer": { "type": "number" },
+            "issues-housekeeping": { "type": "number" },
+            "dependency-update-security-check": { "type": "number" },
+            "product-planning": { "type": "number" },
+            "analytics-review": { "type": "number" }
+          },
+          "additionalProperties": { "type": "number" },
+          "description": "Workflow runner timeout in minutes per routine"
+        }
+      },
+      "additionalProperties": false,
+      "description": "Per-routine token, iteration, and timeout budgets"
+    },
     "schedules": {
       "type": "object",
       "properties": {
@@ -90,6 +146,18 @@
       },
       "additionalProperties": false,
       "description": "Configuration for priority-driven dual local/cloud agent execution"
+    },
+    "labels": {
+      "type": "object",
+      "properties": {
+        "protected": {
+          "type": "array",
+          "items": { "type": "string" },
+          "description": "User-defined list or wildcard patterns of protected labels retained during label pruning"
+        }
+      },
+      "additionalProperties": false,
+      "description": "Configuration for label management and protection shields"
     }
   },
   "required": ["version", "preset", "routines", "skills"],

package/templates/evals/ambiguity-benchmark.json

@@ -0,0 +1,92 @@
+[
+  {
+    "id": "eval-ambiguous-01",
+    "title": "Improve telemetry export speed",
+    "body": "The telemetry export is feeling sluggish lately. Please optimize it to make it faster.",
+    "labels": ["enhancement"],
+    "expectedAction": "ASK_QUESTIONS",
+    "rationale": "Lacks baseline metrics, target latency thresholds, profiling data, or specific code bottlenecks.",
+    "origin": "synthetic"
+  },
+  {
+    "id": "eval-ambiguous-02",
+    "title": "Add role-based permissions",
+    "body": "We need role-based permissions in the system so only authorized users can access certain routes.",
+    "labels": ["feat"],
+    "expectedAction": "ASK_QUESTIONS",
+    "rationale": "Does not specify what roles exist, which routes/mutations require auth, or how permissions are stored/checked.",
+    "origin": "synthetic"
+  },
+  {
+    "id": "eval-ambiguous-03",
+    "title": "Fix the layout on mobile devices",
+    "body": "The UI looks broken on phones. Please clean up the styles.",
+    "labels": ["bug"],
+    "expectedAction": "ASK_QUESTIONS",
+    "rationale": "No specific viewport sizes, reproduction steps, screenshots, or target components identified.",
+    "origin": "synthetic"
+  },
+  {
+    "id": "eval-ambiguous-04",
+    "title": "Refactor database queries",
+    "body": "Database queries need to be cleaned up and refactored for better architecture.",
+    "labels": ["refactor"],
+    "expectedAction": "ASK_QUESTIONS",
+    "rationale": "Unbounded scope with no target tables, architectural patterns, or verifiable acceptance criteria.",
+    "origin": "synthetic"
+  },
+  {
+    "id": "eval-ambiguous-05",
+    "title": "Clean up stale code",
+    "body": "There are some old unused functions in the codebase. Let's delete them.",
+    "labels": ["chore"],
+    "expectedAction": "ASK_QUESTIONS",
+    "rationale": "Vague premise without naming files, modules, or defining what criteria determines staleness.",
+    "origin": "synthetic"
+  },
+  {
+    "id": "eval-specified-01",
+    "title": "Fix CLI flag parsing for --budget in telemetry command",
+    "body": "When running `jonah-fleet telemetry --budget 5000000`, the budget is ignored because the option parser expects a number.\n\n## Tasks\n- [ ] Parse `options.budget` string to integer in `src/commands/telemetry.ts`\n- [ ] Fall back to manifest `weeklyTokenBudget` if omitted\n- [ ] Add unit test in `tests/telemetry-cli.test.ts` verifying flag override\n\nVerification: `npm test tests/telemetry-cli.test.ts`",
+    "labels": ["bug", "priority/P1"],
+    "expectedAction": "PROCEED_TO_IMPLEMENT",
+    "rationale": "Carries reproduction path, explicit file locations, task checklist, and test verification command.",
+    "origin": "synthetic"
+  },
+  {
+    "id": "eval-specified-02",
+    "title": "Add --json output flag to jonah-fleet status command",
+    "body": "Support structured machine-readable output for `jonah-fleet status`.\n\n## Tasks\n- [ ] Add `-j, --json` option to `src/index.ts` status command definition\n- [ ] Update `src/commands/status.ts` to output `JSON.stringify(statusReport, null, 2)` when `options.json` is true\n- [ ] Verify unit test passes via `npm test tests/status.test.ts`",
+    "labels": ["enhancement", "priority/P2"],
+    "expectedAction": "PROCEED_TO_IMPLEMENT",
+    "rationale": "Defines explicit CLI interface, component files, output contract, and unit test command.",
+    "origin": "synthetic"
+  },
+  {
+    "id": "eval-specified-03",
+    "title": "Strict TypeScript compilation failure in src/lib/worktree.ts on Node 22",
+    "body": "`tsc --noEmit` fails on `src/lib/worktree.ts:42` with `TS2345: Argument of type 'string | undefined' is not assignable to parameter of type 'string'`.\n\n## Tasks\n- [ ] Add null check for `process.env.GIT_WORK_TREE` before path resolution\n- [ ] Run `npm run type-check` to verify 0 errors",
+    "labels": ["bug", "priority/P0"],
+    "expectedAction": "PROCEED_TO_IMPLEMENT",
+    "rationale": "Exact compiler error code, line number, fix directive, and verification command provided.",
+    "origin": "synthetic"
+  },
+  {
+    "id": "eval-specified-04",
+    "title": "Add missing grill-me skill validation assertion to tests/installer.test.ts",
+    "body": "Ensure `installFleet` verifies `.agents/skills/grill-me/SKILL.md` is installed in standard preset.\n\n## Tasks\n- [ ] Add `expect(fs.existsSync(path.join(tempDir, '.agents/skills/grill-me/SKILL.md'))).toBe(true)` to `tests/installer.test.ts`\n- [ ] Verify test passes with `npm test tests/installer.test.ts`",
+    "labels": ["test", "size/S"],
+    "expectedAction": "PROCEED_TO_IMPLEMENT",
+    "rationale": "Exact code assertion, file path, and verification command specified.",
+    "origin": "synthetic"
+  },
+  {
+    "id": "eval-specified-05",
+    "title": "Export AmbiguityMetrics interface from src/lib/telemetry.ts",
+    "body": "External consumers and tests need direct access to `AmbiguityMetrics`.\n\n## Tasks\n- [ ] Add `export` keyword to `AmbiguityMetrics` interface in `src/lib/telemetry.ts`\n- [ ] Export from `src/index.ts`\n- [ ] Run `npm run type-check` to verify export integrity",
+    "labels": ["enhancement", "size/S"],
+    "expectedAction": "PROCEED_TO_IMPLEMENT",
+    "rationale": "Fully specified type export with exact interface name, target files, and verification steps.",
+    "origin": "synthetic"
+  }
+]

package/templates/prompts/ORCHESTRATION.md

@@ -6,16 +6,16 @@ How agent routines in this repository are dispatched, claimed, and reconciled
 
 This project's automation is a GitHub-native implementation of the orchestration pattern formalized by OpenAI's [Symphony specification](https://github.com/openai/symphony/blob/main/SPEC.md) for orchestrating autonomous coding agents against an issue tracker. There is **no long-running orchestrator daemon**; the roles map onto GitHub primitives:
 
-| Symphony Concept | Implementation in this repo |
-|---|---|
-| `WORKFLOW.md` (repo-owned config + prompt templates) | `AGENTS.md` (aliased as `GEMINI.md`/`CLAUDE.md`) + `.github/prompts/*.md` |
-| Orchestrator (poll, dispatch, reconcile) | GitHub Actions triggers + scheduled routine sessions |
-| Issue tracker (Linear in Symphony) | GitHub Issues |
-| Agent runner (Codex app-server in per-issue workspace) | An ephemeral agent session (Antigravity CLI `agy`) in an isolated fresh clone |
-| Tracker is reader/scheduler; mutations happen via agent tools | Routines only schedule; the agent session makes every GitHub write |
-
+| Symphony Concept                                              | Implementation in this repo                                                   |
+| ------------------------------------------------------------- | ----------------------------------------------------------------------------- |
+| `WORKFLOW.md` (repo-owned config + prompt templates)          | `AGENTS.md` (aliased as `GEMINI.md`/`CLAUDE.md`) + `.github/prompts/*.md`     |
+| Orchestrator (poll, dispatch, reconcile)                      | GitHub Actions triggers + scheduled routine sessions                          |
+| Issue tracker (Linear in Symphony)                            | GitHub Issues                                                                 |
+| Agent runner (Codex app-server in per-issue workspace)        | An ephemeral agent session (Antigravity CLI `agy`) in an isolated fresh clone |
+| Tracker is reader/scheduler; mutations happen via agent tools | Routines only schedule; the agent session makes every GitHub write            |
 
 Dispatch is both **scheduled** and **event-driven**. All routines run as ephemeral agent sessions via **Antigravity CLI (`agy`)** powered by **Gemini 3.7 Flash (High reasoning)**. The routine suite is calibrated to operate within a **strict 70% weekly token ceiling across all routines combined**, supervised by `optimizer.md`:
+
 - **Scheduled cron sweeps**: Autowork runs periodically (`autowork-cron.yml`), complemented by prompt optimization (`prompt-optimizer-cron.yml`), issues housekeeping (`issues-housekeeping-cron.yml`), and dependency security checks (`dependency-check-cron.yml`).
 - **Event-driven & manual triggers**: GitHub Actions workflows fire routines on events and interactive commands so work starts within seconds instead of waiting for scheduled ticks:
   - `trigger-review-routine.yml` fires Peer Review automatically when a PR is marked ready for review, updated, or review is requested (`ready_for_review`, `opened`, `reopened`, `synchronize`, `review_requested`). It can also be manually (re)triggered via `workflow_dispatch` (with optional `pr_number` for Targeted mode or blank for Scan mode) or by commenting `/review`, `/peer-review`, `/retrigger`, or `/re-review` on any open pull request.
@@ -27,8 +27,8 @@ Autowork triggers pass the target issue via environment variables (`TARGET_ISSUE
 
 Invariants deliberately upheld from this spec:
 
-- **Single-flight per issue** — at most one run works an issue at a time, enforced by the autowork claim protocol (assign → read-back → earliest-timestamp tiebreak). For **umbrella** issues, single-flight is maintained at the *child-issue* level so slices progress cleanly.
-- **Recover dead-run claims** — a crashed run's orphaned claim is released back to the pool rather than starving the issue, both opportunistically during candidate selection and periodically via issues housekeeping.
+- **Single-flight per issue and PR convergence** — at most one run works an issue or pull request at a time, enforced by the autowork claim protocol (assign → read-back → earliest-timestamp tiebreak). For **umbrella** issues, single-flight is maintained at the _child-issue_ level so slices progress cleanly.
+- **Recover dead-run claims** — a crashed run's orphaned claim is released back to the pool rather than starving the issue or PR, both opportunistically during candidate selection and periodically via issues housekeeping.
 - **Reader/writer separation** — the routine that authors a PR never merges it; the Peer Review routine is the sole merge authority for **product** PRs. (Operational log-only PRs are exempt; see Log delivery fallback.)
 - **Warm-Context Review Synchronization** — Autowork maintains an active warm session during implementation, polling for Peer Review's verdict. When Peer Review bounces a PR to draft with findings, Autowork immediately detects the draft state in-session, applies fixes directly to its warm working tree, and re-marks the PR ready—re-firing Peer Review for Round N+1 without cold-start overhead.
 
@@ -36,18 +36,29 @@ Invariants deliberately upheld from this spec:
 
 ## Stale-Claim Definition
 
-Single source of truth for both autowork candidate reclamation and housekeeping sweeps. An assigned issue is a *stale claim* (a dead autowork run's orphaned reservation, safe to release) only when **all** of these hold:
+Single source of truth for both autowork candidate reclamation and housekeeping sweeps. An assigned issue is a _stale claim_ (a dead autowork run's orphaned reservation, safe to release) only when **all** of these hold:
 
 1. **It is an autowork claim, not a manual one.** The issue carries a `🔒 Claimed by autowork run …` comment. An assigned issue with **no** such comment is never stale; leave it alone (it may be a person working manually).
 2. **No live work exists.** There is **no open PR** referencing the issue (`Closes #N`). An open PR is live, recoverable work that autowork Phase 1 owns — never reclaim it, at any age.
 3. **The claim is old.** The most recent `🔒 Claimed by autowork run …` comment's GitHub creation time (`created_at`) is **more than 6 hours** ago. Measure age from that `created_at` only — never the issue's `updated_at`.
 
 **Releasing a stale claim is a destructive write and MUST be guarded:**
+
 - **Re-read immediately before writing.** Re-read the issue (`issue_read`) right before the unassign and re-confirm conditions 1–3 still hold. If any no longer holds, abort the release and move on.
 - **Remove only the named dead owner.** Unassign that specific login; never blindly clear all assignees.
 
 ---
 
+## PR Stale-Claim Definition (Phase 1 Convergence)
+
+Single source of truth for Autowork Phase 1 pull request convergence. An assigned pull request or draft PR with unaddressed review comments is a _stale claim_ (safe to reclaim and reassign by another runner) only when **all** of these hold:
+
+1. **It carries an autowork claim comment**: The PR thread contains `🔒 Addressing review findings by autowork run …` or `🔒 Addressing review findings by local autowork session …`.
+2. **The claim is old**: The most recent claim comment's `created_at` is **more than 2 hours** ago. (2 hours instead of 6 hours because PR review convergence is a rapid turnaround loop).
+3. **No active commits or review activity**: No new commit has been pushed to the PR branch within the last 2 hours.
+
+---
+
 ## Routine Matching & Invocation
 
 **Identify the applicable routine at the start of every conversation, before doing any work:**
@@ -56,15 +67,15 @@ Single source of truth for both autowork candidate reclamation and housekeeping
 2. **Content match** — compare the task against the routine table below. When matching an interactive request from a human, name the matched routine and confirm before proceeding.
 3. **No match** — follow the general Working Practices, PR Workflow, and documentation rules with no routine-specific constraints.
 
-| Routine | File | Applies when the conversation is about... |
-|---|---|---|
-| Autowork | `.github/prompts/autowork.md` | Converging on open work: addressing PR review comments, closing issues whose PRs merged, then claiming and implementing the highest-priority unclaimed issue |
-| Peer Review | `.github/prompts/peer-review.md` | Reviewing a pull request (a named PR or scan mode) and merging it or leaving findings and bouncing to draft |
-| Prompt Optimizer | `.github/prompts/optimizer.md` | Diagnosing failures, inefficiency, token anomalies, and analyzing resolved bugs to propose prompt/test/workflow fixes and upstream contributions |
-| Issues Housekeeping | `.github/prompts/issues-housekeeping.md` | Sweeping open issues for staleness, duplicates, label drift, priority accuracy, and orphaned claims |
-| Dependency Update & Security Check | `.github/prompts/dependency-update-security-check.md` | Checking dependencies for updates and known vulnerabilities, opening actionable PRs |
-| Product Planning | `.github/prompts/product-planning.md` | Turning roadmap priorities into staged issues (`/to-tickets`) and formal PRDs (`/to-spec`) |
-| Analytics Review | `.github/prompts/analytics-review.md` | Evaluating telemetry & measurement trackers against success metrics, emitting action directives (PIVOT/DEPRECATE/ITERATE), and bridging to product planning |
+| Routine                            | File                                                  | Applies when the conversation is about...                                                                                                                    |
+| ---------------------------------- | ----------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------ |
+| Autowork                           | `.github/prompts/autowork.md`                         | Converging on open work: addressing PR review comments, closing issues whose PRs merged, then claiming and implementing the highest-priority unclaimed issue |
+| Peer Review                        | `.github/prompts/peer-review.md`                      | Reviewing a pull request (a named PR or scan mode) and merging it or leaving findings and bouncing to draft                                                  |
+| Prompt Optimizer                   | `.github/prompts/optimizer.md`                        | Diagnosing failures, inefficiency, token anomalies, and analyzing resolved bugs to propose prompt/test/workflow fixes and upstream contributions             |
+| Issues Housekeeping                | `.github/prompts/issues-housekeeping.md`              | Sweeping open issues for staleness, duplicates, label drift, priority accuracy, and orphaned claims                                                          |
+| Dependency Update & Security Check | `.github/prompts/dependency-update-security-check.md` | Checking dependencies for updates and known vulnerabilities, opening actionable PRs                                                                          |
+| Product Planning                   | `.github/prompts/product-planning.md`                 | Turning roadmap priorities into staged issues (`/to-tickets`) and formal PRDs (`/to-spec`)                                                                   |
+| Analytics Review                   | `.github/prompts/analytics-review.md`                 | Evaluating telemetry & measurement trackers against success metrics, emitting action directives (PIVOT/DEPRECATE/ITERATE), and bridging to product planning  |
 
 ---
 
@@ -98,9 +109,11 @@ The routines invoke specialized engineering skills at key workflow checkpoints:
 ## Log Delivery Fallback
 
 Single source of truth for every routine's Logging section:
-1. **Direct commit to `main` is default**: For operational run logs under `.github/prompts/logs/**`, commit directly to `main` via GitHub API or git push.
-2. **Draft PR fallback**: If direct push fails, commit the log to a dedicated, fresh branch and open a draft PR carrying only the log files.
-3. **Automated landing**: `auto-merge-log-prs.yml` or `issues-housekeeping.md` lands accumulated log PRs. Draft log PRs are never reviewed by Peer Review and do not count toward Autowork's backpressure limits.
+
+1. **Direct commit to `main` is default**: For operational run logs under `.github/prompts/logs/**`, commit directly to `main` via GitHub API or git push. Never commit logs to an active feature branch.
+2. **Mandatory `[skip ci]`**: All log commits MUST include `[skip ci]` in the commit message (e.g. `docs(log): record {routine-name} run {timestamp} [skip ci]`) to prevent unnecessary CI executions or approval blocks.
+3. **Draft PR fallback**: If direct push fails, commit the log to a dedicated, fresh branch and open a draft PR carrying only the log files.
+4. **Automated landing**: `auto-merge-log-prs.yml` or `issues-housekeeping.md` lands accumulated log PRs. Draft log PRs are never reviewed by Peer Review and do not count toward Autowork's backpressure limits.
 
 ---
 
@@ -115,7 +128,9 @@ How token spend, runaway loops, and budget anomalies are detected, triaged, and
    - **Iteration Ceiling Exhaustion**: >20% of runs in a routine terminate at the `token_limit` / max iteration cap. Trigger: tasks too complex for single-flight execution or unbounded looping. Remediation: enforce vertical slicing / umbrella decomposition, tighten pre-ready self-audits, or refine termination bounds.
    - **Review Loop Burn**: Pull requests experiencing $\ge 3$ bounce rounds between autowork and peer-review. Trigger: ambiguous reviewer feedback, pedantic non-blocking findings, or brittle test assertions. Remediation: tighten reviewer trust/noise rules, calibrate reviewer severity thresholds, and engage human escalation via ping-pong caps.
    - **Feedback Loop Stagnation**: A downstream processing routine (e.g. measurement loop, verification, triage) records 0 intake (`filed: 0`) across $\ge 2$ consecutive runs while upstream PRs merge or roadmap/feature issues close in the same window. Trigger: overly broad or qualitative discovery sweeps that falsely claim all items are covered. Remediation: mandate deterministic per-issue matching tables and itemized reconciliation against upstream closed issues/PRs.
-3. **Automated Remediation PRs**: The optimizer automatically drafts targeted PRs—locally for repo-specific rules/configs, or upstream via `npx jonah-fleet contribute` for fleet-wide prompt/workflow improvements.
+   - **Passive Order-Taking Anomaly ("Yes-Man Blindspot")**: The Ambiguity Gate trigger rate across intake runs in `autowork` or `triage` is <5% despite elevated PR review bounces ($\ge 2$) or high iteration usage ($\ge 35$), indicating agents are silently guessing requirements and building flawed implementations rather than interrogating underspecified issues. Trigger: agents defaulting to compliance over inquiry. Remediation: tighten Step 12 criteria in `autowork.md` and `triage.md` to force explicit questions on ambiguous terms, and feed problem cases into the automated ambiguity benchmark eval suite.
+   - **Speculative Runaway Waste**: An agent run consumed >50k tokens on an underspecified issue with 0 clarifying questions asked, and subsequently failed, bounced, or required post-merge rework. Remediation: tighten ambiguity gate checks and export the failing issue into the automated benchmark eval dataset.
+3. **Automated Remediation PRs & Benchmark Evals**: The optimizer automatically drafts targeted PRs—locally for repo-specific rules/configs, or upstream via `npx jonah-fleet contribute` for fleet-wide prompt/workflow improvements—and appends failure fixtures to the automated benchmark eval suite.
 
 ---
 
@@ -127,13 +142,14 @@ How external and human contributor pull requests are reconciled into the issue t
 2. **Autonomous Synthesis on Merge**: When `peer-review.md` approves a pull request lacking a `Closes #N` link, the review routine automatically synthesizes a tracking issue before merging:
    - Creates a tracked issue via `gh issue create` capturing the PR title, body, and deliverables.
    - Appends `Closes #<synthesized_issue_id>` to the PR description via `gh pr edit`.
-3. **Audit & Single-Flight Lineage**: When the PR is squash-merged, GitHub's native issue closure kicks in and automatically closes the synthesized issue. This maintains 100% issue auditability, project board tracking, telemetry metrics, and release changelogs without imposing any friction on human contributors.
+3. **Audit & Single-Flight Lineage**: When the PR is squash-merged, `peer-review` explicitly closes the tracking issue (`gh issue close <ISSUE_NUMBER>`) to guarantee tracking closure, even when bot credentials or draft PR states bypass GitHub's native issue auto-close. This maintains 100% issue auditability, project board tracking, telemetry metrics, and release changelogs without leaving stray open issues.
 
 ---
 
 ## Fleet Telemetry & Weekly Token Economics
 
 Cross-repository telemetry aggregation and token tracking protocol:
+
 1. **Lightweight Routine Telemetry**: Every autonomous routine log (`.github/prompts/logs/*/*.md`) emits a structured `RoutineTelemetrySummary` capturing routine identity, duration, iterations, result, failure category, cost, and tokens.
 2. **Opt-in Emission Step**: GitHub Actions workflows (`autowork-cron.yml`, `trigger-review-routine.yml`, `prompt-optimizer-cron.yml`) run `jonah-fleet telemetry --emit` using optional `JONAH_FLEET_TELEMETRY_ENDPOINT` secrets.
 3. **Global 70% Budget Ceiling**: Tracks rolling 7-day spend across all fleet repositories against the global ceiling (~8.75M tokens/week).
@@ -152,24 +168,29 @@ How the fleet guarantees continuous review throughput, recovers from transient A
 1. **Failure Trapping & Transparency**: When a peer-review workflow session encounters an execution failure (e.g. LLM API rate limit / quota exhaustion or runner timeout), `trigger-review-routine.yml` traps the failure and posts an informational status notice on the PR. Failures are made immediately visible on the PR timeline rather than silently failing in the background.
 2. **Periodic Scan Sweep (Watchdog)**: `trigger-review-routine.yml` runs a scheduled 2-hour cron sweep (`cron: '45 */2 * * *'`) in Scan mode. Any open PR in `ready_for_review` state that was orphaned due to a transient API rate limit or missed event trigger is automatically picked up, evaluated, and resolved.
 3. **Interactive Re-triggering**: Any team member or author can immediately re-dispatch review by commenting `/review`, `/peer-review`, `/retrigger`, or `/re-review` on any open pull request, or manually triggering `trigger-review-routine.yml` via `workflow_dispatch`.
-4. **Autowork Phase 1 Watchdog**: During Phase 1 convergence, Autowork actively identifies open ready PRs that have received no review feedback for $>2$ hours, re-triggering review via draft toggle or `/review` comment before picking up new work.
+4. **Autowork Phase 1 Watchdog**: During Phase 1 convergence, Autowork actively identifies open ready PRs that have received no review feedback for $>2$ hours, re-triggering review via draft toggle or `/review` comment before picking up new work. Review re-triggering is strictly conditioned on all CI checks having passed (`conclusion: SUCCESS`, `mergeStateStatus: CLEAN`) and is strictly prohibited if checks are in-progress or awaiting approval (`ACTION_REQUIRED`).
 
 ---
 
-## Upstream Symphony Intel & Architectural Evaluation Framework
+## Upstream Symphony & Funes Intel & Architectural Evaluation Framework
 
-How changes and innovations from [openai/symphony](https://github.com/openai/symphony) are systematically audited and evaluated for incorporation into Jonah Fleet:
+How changes and innovations from upstream ecosystems—[openai/symphony](https://github.com/openai/symphony) for issue-tracker orchestration and [huggingface/funes](https://github.com/huggingface/funes) for agent memory & session indexing—are systematically audited and evaluated for incorporation into Jonah Fleet:
 
-1. **Automated Radar (`symphony-radar.yml`)**: A weekly scheduled workflow runs `.github/scripts/fetch-symphony-radar.js` to inspect upstream commits, specification updates (`SPEC.md`), and releases, generating an actionable digest issue in Jonah Fleet.
+1. **Automated Ecosystem Radar (`symphony-radar.yml`)**: A weekly scheduled workflow runs `.github/scripts/fetch-symphony-radar.js` to inspect upstream commits, specification updates (`SPEC.md`), and releases across `openai/symphony` (orchestration) and `huggingface/funes` (memory tooling), generating an actionable digest issue in Jonah Fleet.
 2. **The 4 Evaluation Layers**:
    - **Layer 1 (Zero-Daemon Invariant)**: Can the enhancement execute in ephemeral GitHub Actions and `agy` CLI sessions without requiring a 24/7 background server or persistent WebSocket?
    - **Layer 2 (Issue Tracker Abstraction)**: Does the pattern map cleanly to native GitHub Issues, labels, and PR checks without proprietary tracker dependencies?
    - **Layer 3 (Token & Cost Economy)**: Does the change optimize LLM spend within Jonah Fleet's 70% weekly token ceiling (~8.75M tokens)?
    - **Layer 4 (Multi-Repo Portability)**: Can the routine or skill be distributed via `agents-manifest.json` and `jonah-fleet sync` across any consumer repository?
-3. **Classification & Action Protocol**:
-   - **🟢 Category A (Adopt Directly)**: Security guardrails, claim lock invariants, reader/writer rules, prompt engineering optimizations.
-   - **🟡 Category B (Adapt to Actions/CLI)**: Dynamic orchestrator pacing, backpressure controls, multi-stage review checks.
-   - **🔴 Category C (Skip)**: Elixir/OTP supervision trees, BEAM memory tuning, proprietary runtime internals.
+3. **Agent Memory & Session Indexing Evaluation Dimensions (Funes Watch)**:
+   - **Zero-LLM Ingestion**: Deterministic parsing of agent session traces (`.jsonl`/Parquet) into LanceDB without spending LLM tokens from the weekly budget.
+   - **Pull-Based Memory Delivery**: Memory served strictly on demand via MCP (`recall`, `get`) to prevent prompt context bloat.
+   - **Cross-Session Provenance**: Verbatim turns and provenance retention instead of lossy summary drift.
+   - **Multi-Agent Portability**: Standardized trace ingestion across Antigravity CLI (`agy`), Claude Code, and Codex.
+4. **Classification & Action Protocol**:
+   - **🟢 Category A (Adopt Directly)**: Security guardrails, claim lock invariants, reader/writer rules, prompt engineering optimizations, deterministic zero-LLM indexing.
+   - **🟡 Category B (Adapt to Actions/CLI)**: Dynamic orchestrator pacing, backpressure controls, multi-stage review checks, pull-based memory MCP integrations.
+   - **🔴 Category C (Skip)**: Elixir/OTP supervision trees, BEAM memory tuning, proprietary runtime internals, always-loaded memory context dumps.
 
 ---
 
@@ -194,7 +215,7 @@ How agent routines are partitioned between cloud GitHub Actions (24/7 cloud runn
 2. **PR Priority Mirroring**: When Autowork opens a PR, it automatically mirrors the parent issue's priority labels (`priority/P0`..`P3`) onto the pull request so downstream review workflows can filter triggers without extra API overhead.
 3. **Workspace Isolation via Git Worktrees**: Local agents execute inside ephemeral worktrees (`.jonah-fleet/worktrees/<routine>-<issue>`) off `origin/main`, ensuring active editor sessions, uncommitted changes, and local branches are never modified or disturbed.
 4. **Local Claim Protocol**:
-   - Local agent claims post: `🔒 Claimed by local autowork session (host: <hostname>) <timestamp>`.
+   - Local agent issue claims post: `🔒 Claimed by local autowork session (host: <hostname>) <timestamp>`.
+   - Local agent PR convergence claims post: `🔒 Addressing review findings by local autowork session (host: <hostname>) <timestamp>`.
    - Local processes trap `SIGINT`/`SIGTERM` to unassign claims and remove worktrees cleanly on exit.
-   - Standard 6-hour stale-claim rule safely reclaims orphaned local claims if a machine powers down unexpectedly.
-
+   - Standard stale-claim rules (6h for issues, 2h for PRs) safely reclaim orphaned local claims if a machine powers down unexpectedly.

package/templates/prompts/_prompt-template.md

@@ -40,4 +40,16 @@ After completing (SUCCESS or FAILURE), write a log file to `.github/prompts/logs
 - Full execution trace with tool calls
 - If FAILURE: root cause, category, and suggested fix
 
-**Important**: Commit the log file directly to `main` and push — explicitly permitted for files under `.github/prompts/logs/**`. Follow the Log delivery fallback in `ORCHESTRATION.md` if direct push fails.
+**Log Delivery Protocol & Invariants**:
+- **Negative Rule**: NEVER commit or push run logs to a feature branch or open PR branch. Doing so emits a `pull_request: synchronize` event under bot credentials, triggering GitHub Actions workflow approval gates (`action_required`) that stall CI.
+- **Mandatory `[skip ci]`**: Always append `[skip ci]` to any log commit message (e.g. `docs(log): record {routine-name} run {timestamp} [skip ci]`).
+- **Direct Push to `main`**: Commit the log file directly to `main` and push — explicitly permitted for files under `.github/prompts/logs/**`:
+  ```bash
+  git checkout main
+  git pull origin main
+  git add .github/prompts/logs/{routine-name}/{timestamp}.md
+  git commit -m "docs(log): record {routine-name} run {timestamp} [skip ci]"
+  git push origin main
+  ```
+  Follow the Log delivery fallback in `ORCHESTRATION.md` if direct push fails.
+