jonah-fleet 1.4.1 → 1.4.2

package/CHANGELOG.md

@@ -5,16 +5,35 @@ All notable changes to `jonah-fleet` will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [1.4.2] - 2026-09-03
+
+### Changed
+- Strict Allowlist for Cloud Review Triggers:
+  - Updated `trigger-review-routine.yml` to strictly require `priority/P0` or `priority/P1` labels to fire immediate cloud reviews on `pull_request` events.
+  - Untagged PRs and lower-priority PRs (P2/P3) skip immediate cloud review, routing review execution exclusively to local peer-review daemons (with 48h scheduled cloud catchup fallback).
+
 ## [1.4.1] - 2026-09-03
 
 ### Added
 - Decoupled Multi-Cadence Daemon & Zero-Cost PR Preflight:
   - Added independent scheduling in `jonah-fleet daemon` via `--review-interval` (default: 3m) and `--autowork-interval` (default: 30m).
   - Added ultra-fast (~100ms) local PR preflight check (`countOpenReadyPRs`) in `src/lib/daemon.ts` that queries `gh pr list` and skips agent invocations with 0 token spend when 0 ready PRs are open.
+- User-Friendly Terminal Cards & Daemon Ticker:
+  - Replaced noisy raw markdown streaming in `jonah-fleet daemon --foreground` and `jonah-fleet run` with a dynamic single-line status spinner showing target and active phase heuristics.
+  - Added real-time token stream redirection to `.jonah-fleet/daemon.log` with `-v, --verbose` escape hatch for debugging.
+  - Added formatted Unicode summary and error cards with ephemeral worktree path sanitization and disk log fallbacks (`src/lib/terminal-card.ts`).
+  - Added in-place countdown ticker (`\r`) during daemon watchdog idle states to avoid cluttering terminal scrollback.
 
 ## [1.4.0] - 2026-09-03
 
 ### Added
+- Inquisitive Agent Fleet & Grilling Protocol:
+  - Added new `grill-me` engineering skill (`templates/skills/grill-me/SKILL.md` and `.agents/skills/grill-me/SKILL.md`) to stress-test proposals, requirements, and designs through a targeted, sequential interview to eliminate assumptions before implementation (#54, #55).
+  - Included `grill-me` in `standard` and `full` presets in `src/lib/presets.ts` and `agents-manifest.json` (#54, #55).
+  - Added Requirements Discovery & Inquisitive Stance rules to `templates/docs/AGENTS.template.md` and `AGENTS.md`, establishing default skepticism on necessity, zero-guesswork ambiguity gates, and high-leverage question structuring (#54, #55).
+  - Added Ambiguity & Missing Acceptance Criteria Gate to Step 12 in `autowork.md` (`templates/prompts/autowork.md` & `.github/prompts/autowork.md`) to post clarifying questions, label `needs-info`, and release the claim instead of making silent assumptions (#54, #55).
+  - Added Selective Server-Side Issue Querying in Step 8 of `autowork.md` using `gh --search` with exclusion filters (`no:assignee -label:measurement -label:needs-human -label:needs-design -label:wontfix -label:needs-info`) to avoid tool buffer truncation and token waste (#54, #55).
+  - Updated `triage.md` and `ORCHESTRATION.md` to integrate `/grill-me` into issue evaluation and planning (#54, #55).
 - Priority-Driven Dual Agent Execution (Local + GitHub Actions):
   - Added CLI `jonah-fleet run <routine>` command to execute prompt routines locally with automatic Git worktree isolation (`.jonah-fleet/worktrees/`), protecting active working copies and uncommitted editor files.
   - Added CLI `jonah-fleet daemon [start|stop|status]` background worker to continuously poll open issues and PRs for local execution.

package/dist/index.js

@@ -49,7 +49,8 @@ var PRESET_CONFIGS = {
       "diagnosing-bugs",
       "resolving-merge-conflicts",
       "writing-for-agents",
-      "triage"
+      "triage",
+      "grill-me"
     ]
   },
   full: {
@@ -71,6 +72,7 @@ var PRESET_CONFIGS = {
       "resolving-merge-conflicts",
       "writing-for-agents",
       "triage",
+      "grill-me",
       "to-spec",
       "to-tickets"
     ]
@@ -90,7 +92,7 @@ var ROUTINE_TO_WORKFLOW_MAP = {
   "product-planning": [],
   "analytics-review": []
 };
-var FLEET_VERSION = "1.4.1";
+var FLEET_VERSION = "1.4.2";
 var SCHEMA_URL = "https://raw.githubusercontent.com/juliendurandeu/jonah-fleet/main/schema.json";
 
 // src/lib/manifest.ts
@@ -2161,7 +2163,7 @@ import pc9 from "picocolors";
 import fs11 from "fs";
 import path11 from "path";
 import os2 from "os";
-import { spawn } from "child_process";
+import { spawn, execSync as execSync2 } from "child_process";
 
 // src/lib/worktree.ts
 import fs10 from "fs";

package/dist/lib/presets.d.ts

@@ -47,6 +47,6 @@ export declare const PRESET_CONFIGS: Record<Exclude<PresetName, 'custom'>, {
 }>;
 export declare const ROUTINE_TO_WORKFLOW_MAP: Record<keyof FleetManifest['routines'], string[]>;
 export declare const WORKFLOW_TO_ROUTINE_MAP: Record<string, keyof FleetManifest['routines'] | 'sync-fleet'>;
-export declare const FLEET_VERSION = "1.4.1";
+export declare const FLEET_VERSION = "1.4.2";
 export declare const SCHEMA_URL = "https://raw.githubusercontent.com/juliendurandeu/jonah-fleet/main/schema.json";
 //# sourceMappingURL=presets.d.ts.map

package/dist/lib/presets.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"presets.d.ts","sourceRoot":"","sources":["../../src/lib/presets.ts"],"names":[],"mappings":"AAAA,MAAM,MAAM,UAAU,GAAG,SAAS,GAAG,UAAU,GAAG,MAAM,GAAG,QAAQ,CAAC;AAEpE,MAAM,WAAW,aAAa;IAC5B,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,OAAO,EAAE,MAAM,CAAC;IAChB,MAAM,EAAE,UAAU,CAAC;IACnB,QAAQ,EAAE;QACR,QAAQ,EAAE,OAAO,CAAC;QAClB,aAAa,EAAE,OAAO,CAAC;QACvB,SAAS,EAAE,OAAO,CAAC;QACnB,qBAAqB,EAAE,OAAO,CAAC;QAC/B,kCAAkC,EAAE,OAAO,CAAC;QAC5C,kBAAkB,EAAE,OAAO,CAAC;QAC5B,kBAAkB,EAAE,OAAO,CAAC;KAC7B,CAAC;IACF,SAAS,CAAC,EAAE;QACV,QAAQ,CAAC,EAAE,MAAM,CAAC;QAClB,aAAa,CAAC,EAAE,MAAM,CAAC;QACvB,SAAS,CAAC,EAAE,MAAM,CAAC;QACnB,qBAAqB,CAAC,EAAE,MAAM,CAAC;QAC/B,kCAAkC,CAAC,EAAE,MAAM,CAAC;QAC5C,kBAAkB,CAAC,EAAE,MAAM,CAAC;QAC5B,YAAY,CAAC,EAAE,MAAM,CAAC;QACtB,CAAC,GAAG,EAAE,MAAM,GAAG,MAAM,GAAG,SAAS,CAAC;KACnC,CAAC;IACF,MAAM,EAAE,MAAM,EAAE,CAAC;IACjB,YAAY,CAAC,EAAE,MAAM,EAAE,CAAC;IACxB,UAAU,CAAC,EAAE;QACX,OAAO,EAAE,OAAO,CAAC;QACjB,OAAO,EAAE,QAAQ,GAAG,QAAQ,CAAC;KAC9B,CAAC;IACF,SAAS,CAAC,EAAE;QACV,OAAO,CAAC,EAAE,OAAO,CAAC;QAClB,QAAQ,CAAC,EAAE,MAAM,CAAC;QAClB,iBAAiB,CAAC,EAAE,MAAM,CAAC;KAC5B,CAAC;IACF,aAAa,CAAC,EAAE,mBAAmB,CAAC;CACrC;AAED,MAAM,WAAW,mBAAmB;IAClC,OAAO,CAAC,EAAE,OAAO,CAAC;IAClB,eAAe,CAAC,EAAE,MAAM,EAAE,CAAC;IAC3B,iBAAiB,CAAC,EAAE,MAAM,CAAC;CAC5B;AAED,eAAO,MAAM,6BAA6B,EAAE,mBAI3C,CAAC;AAEF,eAAO,MAAM,cAAc,EAAE,MAAM,CAAC,OAAO,CAAC,UAAU,EAAE,QAAQ,CAAC,EAAE;IAAE,QAAQ,EAAE,aAAa,CAAC,UAAU,CAAC,CAAC;IAAC,MAAM,EAAE,MAAM,EAAE,CAAA;CAAE,CA+D3H,CAAC;AAEF,eAAO,MAAM,uBAAuB,EAAE,MAAM,CAAC,MAAM,aAAa,CAAC,UAAU,CAAC,EAAE,MAAM,EAAE,CAarF,CAAC;AAEF,eAAO,MAAM,uBAAuB,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,aAAa,CAAC,UAAU,CAAC,GAAG,YAAY,CAOlG,CAAC;AAEF,eAAO,MAAM,aAAa,UAAU,CAAC;AACrC,eAAO,MAAM,UAAU,kFAAkF,CAAC"}
+{"version":3,"file":"presets.d.ts","sourceRoot":"","sources":["../../src/lib/presets.ts"],"names":[],"mappings":"AAAA,MAAM,MAAM,UAAU,GAAG,SAAS,GAAG,UAAU,GAAG,MAAM,GAAG,QAAQ,CAAC;AAEpE,MAAM,WAAW,aAAa;IAC5B,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,OAAO,EAAE,MAAM,CAAC;IAChB,MAAM,EAAE,UAAU,CAAC;IACnB,QAAQ,EAAE;QACR,QAAQ,EAAE,OAAO,CAAC;QAClB,aAAa,EAAE,OAAO,CAAC;QACvB,SAAS,EAAE,OAAO,CAAC;QACnB,qBAAqB,EAAE,OAAO,CAAC;QAC/B,kCAAkC,EAAE,OAAO,CAAC;QAC5C,kBAAkB,EAAE,OAAO,CAAC;QAC5B,kBAAkB,EAAE,OAAO,CAAC;KAC7B,CAAC;IACF,SAAS,CAAC,EAAE;QACV,QAAQ,CAAC,EAAE,MAAM,CAAC;QAClB,aAAa,CAAC,EAAE,MAAM,CAAC;QACvB,SAAS,CAAC,EAAE,MAAM,CAAC;QACnB,qBAAqB,CAAC,EAAE,MAAM,CAAC;QAC/B,kCAAkC,CAAC,EAAE,MAAM,CAAC;QAC5C,kBAAkB,CAAC,EAAE,MAAM,CAAC;QAC5B,YAAY,CAAC,EAAE,MAAM,CAAC;QACtB,CAAC,GAAG,EAAE,MAAM,GAAG,MAAM,GAAG,SAAS,CAAC;KACnC,CAAC;IACF,MAAM,EAAE,MAAM,EAAE,CAAC;IACjB,YAAY,CAAC,EAAE,MAAM,EAAE,CAAC;IACxB,UAAU,CAAC,EAAE;QACX,OAAO,EAAE,OAAO,CAAC;QACjB,OAAO,EAAE,QAAQ,GAAG,QAAQ,CAAC;KAC9B,CAAC;IACF,SAAS,CAAC,EAAE;QACV,OAAO,CAAC,EAAE,OAAO,CAAC;QAClB,QAAQ,CAAC,EAAE,MAAM,CAAC;QAClB,iBAAiB,CAAC,EAAE,MAAM,CAAC;KAC5B,CAAC;IACF,aAAa,CAAC,EAAE,mBAAmB,CAAC;CACrC;AAED,MAAM,WAAW,mBAAmB;IAClC,OAAO,CAAC,EAAE,OAAO,CAAC;IAClB,eAAe,CAAC,EAAE,MAAM,EAAE,CAAC;IAC3B,iBAAiB,CAAC,EAAE,MAAM,CAAC;CAC5B;AAED,eAAO,MAAM,6BAA6B,EAAE,mBAI3C,CAAC;AAEF,eAAO,MAAM,cAAc,EAAE,MAAM,CAAC,OAAO,CAAC,UAAU,EAAE,QAAQ,CAAC,EAAE;IAAE,QAAQ,EAAE,aAAa,CAAC,UAAU,CAAC,CAAC;IAAC,MAAM,EAAE,MAAM,EAAE,CAAA;CAAE,CAiE3H,CAAC;AAEF,eAAO,MAAM,uBAAuB,EAAE,MAAM,CAAC,MAAM,aAAa,CAAC,UAAU,CAAC,EAAE,MAAM,EAAE,CAarF,CAAC;AAEF,eAAO,MAAM,uBAAuB,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,aAAa,CAAC,UAAU,CAAC,GAAG,YAAY,CAOlG,CAAC;AAEF,eAAO,MAAM,aAAa,UAAU,CAAC;AACrC,eAAO,MAAM,UAAU,kFAAkF,CAAC"}

package/dist/lib/runner.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"runner.d.ts","sourceRoot":"","sources":["../../src/lib/runner.ts"],"names":[],"mappings":"AAOA,MAAM,WAAW,sBAAsB;IACrC,SAAS,EAAE,MAAM,CAAC;IAClB,OAAO,EAAE,MAAM,CAAC;IAChB,KAAK,CAAC,EAAE,MAAM,GAAG,MAAM,CAAC;IACxB,EAAE,CAAC,EAAE,MAAM,GAAG,MAAM,CAAC;IACrB,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,UAAU,CAAC,EAAE,OAAO,CAAC;IACrB,YAAY,CAAC,EAAE,OAAO,CAAC;IACvB,MAAM,CAAC,EAAE,OAAO,CAAC;IACjB,GAAG,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IAC7B,KAAK,CAAC,EAAE,CAAC,KAAK,EAAE,MAAM,KAAK,IAAI,CAAC;CACjC;AAED,MAAM,WAAW,qBAAqB;IACpC,OAAO,EAAE,OAAO,CAAC;IACjB,QAAQ,EAAE,MAAM,CAAC;IACjB,MAAM,EAAE,MAAM,CAAC;IACf,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,UAAU,CAAC,EAAE,MAAM,CAAC;CACrB;AAED;;GAEG;AACH,wBAAgB,oBAAoB,CAAC,SAAS,EAAE,MAAM,GAAG,MAAM,CAoB9D;AAED;;GAEG;AACH,wBAAgB,kBAAkB,CAChC,SAAS,EAAE,MAAM,EACjB,OAAO,EAAE,MAAM,EACf,OAAO,GAAE;IAAE,KAAK,CAAC,EAAE,MAAM,GAAG,MAAM,CAAC;IAAC,EAAE,CAAC,EAAE,MAAM,GAAG,MAAM,CAAA;CAAO,GAC9D,MAAM,CAmBR;AAED;;GAEG;AACH,wBAAgB,cAAc,IAAI,OAAO,CAQxC;AAED;;GAEG;AACH,wBAAsB,eAAe,CAAC,OAAO,EAAE,sBAAsB,GAAG,OAAO,CAAC,qBAAqB,CAAC,CAwIrG"}
+{"version":3,"file":"runner.d.ts","sourceRoot":"","sources":["../../src/lib/runner.ts"],"names":[],"mappings":"AAOA,MAAM,WAAW,sBAAsB;IACrC,SAAS,EAAE,MAAM,CAAC;IAClB,OAAO,EAAE,MAAM,CAAC;IAChB,KAAK,CAAC,EAAE,MAAM,GAAG,MAAM,CAAC;IACxB,EAAE,CAAC,EAAE,MAAM,GAAG,MAAM,CAAC;IACrB,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,UAAU,CAAC,EAAE,OAAO,CAAC;IACrB,YAAY,CAAC,EAAE,OAAO,CAAC;IACvB,MAAM,CAAC,EAAE,OAAO,CAAC;IACjB,GAAG,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IAC7B,KAAK,CAAC,EAAE,CAAC,KAAK,EAAE,MAAM,KAAK,IAAI,CAAC;CACjC;AAED,MAAM,WAAW,qBAAqB;IACpC,OAAO,EAAE,OAAO,CAAC;IACjB,QAAQ,EAAE,MAAM,CAAC;IACjB,MAAM,EAAE,MAAM,CAAC;IACf,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,UAAU,CAAC,EAAE,MAAM,CAAC;CACrB;AAED;;GAEG;AACH,wBAAgB,oBAAoB,CAAC,SAAS,EAAE,MAAM,GAAG,MAAM,CAoB9D;AAED;;GAEG;AACH,wBAAgB,kBAAkB,CAChC,SAAS,EAAE,MAAM,EACjB,OAAO,EAAE,MAAM,EACf,OAAO,GAAE;IAAE,KAAK,CAAC,EAAE,MAAM,GAAG,MAAM,CAAC;IAAC,EAAE,CAAC,EAAE,MAAM,GAAG,MAAM,CAAA;CAAO,GAC9D,MAAM,CAmBR;AAED;;GAEG;AACH,wBAAgB,cAAc,IAAI,OAAO,CAOxC;AAED;;GAEG;AACH,wBAAsB,eAAe,CAAC,OAAO,EAAE,sBAAsB,GAAG,OAAO,CAAC,qBAAqB,CAAC,CAwIrG"}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "jonah-fleet",
-  "version": "1.4.1",
+  "version": "1.4.2",
   "description": "Standalone autonomous agent fleet with Symphony orchestration, claim protocols, and continuous improvement loops",
   "main": "dist/index.js",
   "types": "dist/index.d.ts",

package/templates/docs/AGENTS.template.md

@@ -41,6 +41,12 @@ tests/            # Test suites
 - **Error Handling**: Validate inputs at server and module boundaries. Return explicit structured errors.
 - **Testing Philosophy**: Write regression unit tests for every bug fix and integration tests for key workflows.
 
+## Requirements Discovery & Inquisitive Stance
+
+- **Challenge Premise First**: Do not be a passive order-taker. Before writing code or drafting large plans, verify whether the requested change is necessary, whether it solves a root cause rather than a symptom, or if existing primitives in the codebase already handle it.
+- **Zero-Guesswork Ambiguity Gate**: Never make silent assumptions about underspecified requirements, schemas, edge cases, or UX flows. Stop and ask clarifying questions before implementation.
+- **Structured Clarifications & Grilling**: When asking questions, propose 1–2 high-leverage questions at a time with concrete options or trade-offs. For complex features or fuzzy requirements, invoke `/grill-me` to stress-test the design before writing code.
+
 ## Conventions
 
 - **Language**: All code, comments, commit messages, and PR titles must be written in **English**.

package/templates/prompts/ORCHESTRATION.md

@@ -83,8 +83,11 @@ The routines invoke specialized engineering skills at key workflow checkpoints:
   - `/resolving-merge-conflicts`: Resolves merge conflicts mechanically before squash-merging.
 - **Issues Housekeeping (`issues-housekeeping.md`)**:
   - `/triage`: Evaluates incoming issues into canonical roles.
+- **Triage & Interactive Planning**:
+  - `/grill-me`: Interrogates proposals and requirements to uncover edge cases and force explicit trade-offs before implementation.
 - **Product Planning (`product-planning.md`)**:
   - `/domain-modeling`: Pressure-tests proposals and records domain terms / ADRs.
+  - `/grill-me`: Stress-tests proposed roadmaps and user stories.
   - `/to-spec`: Authors formal PRDs for larger proposals.
   - `/to-tickets`: Decomposes approved epics/proposals into dependency-linked issues.
 - **Prompt Optimizer (`optimizer.md`)**:

package/templates/prompts/autowork.md

@@ -43,6 +43,7 @@ If any criterion cannot be met, stop immediately and log FAILURE with the reason
 - Do not start implementing an issue before claiming it (both assignment AND claim comment).
 - Do not mark a PR ready while its `mergeable_state` is `dirty` — resolve merge conflicts first.
 - Do not fall into the **Telemetry Rabbit Hole**: do not spend cycles instrumenting elaborate fallback telemetry or defensive error handling for features that suffer from lack of user intent rather than software bugs.
+- Do not guess or invent arbitrary specifications for ambiguous issues — post clarifying questions, label `needs-info`, and release the claim instead of blindly writing code.
 
 ## Instructions
 
@@ -85,10 +86,13 @@ b. **Eligible → claim, then implement.** Call `get_me` once to learn your own
 6. Count open PRs authored by automated sessions representing live reviewable work (excluding log-only PRs). If 3 or more non-log PRs are open, STOP — run is SUCCESS with "Too many open PRs, converging first".
 7. Call `get_me` once to learn your own GitHub account (`login`).
 8. List open issues sorted by priority labels (P1 > P2 > P3). Within the same priority tier, order by type (`bug`/`security` before others), then oldest first.
+   - **Selective Server-Side Querying**: To prevent tool buffer truncation and avoid ingesting un-actionable issues, query GitHub with server-side exclusion filters rather than dumping all issues. E.g.:
+     `gh issue list --state open --search "no:assignee -label:measurement -label:needs-human -label:needs-design -label:wontfix -label:needs-info" --limit 30`
+     (or filter by priority tier: `--search "label:priority/P0,priority/P1 no:assignee -label:measurement -label:needs-human -label:needs-design -label:wontfix -label:needs-info"`).
    - **Dual Execution Priority Routing**:
      - If running in **Cloud Actions** (`$GITHUB_ACTIONS` / `$CI`): Scan mode selects `priority/P0` or `priority/P1` issues first. Lower-priority `priority/P2` or `priority/P3` issues are eligible in cloud Scan mode ONLY if the issue has remained unclaimed for more than 48 hours (`created_at` older than 48h, acting as a cloud catchup sweep).
      - If running in **Local Agent** (`$LOCAL_AGENT`): Scan mode selects across all priorities (P0 → P1 → P2 → P3) without time gating, prioritizing active local backlog consumption.
-   - Skip assigned issues (unless stale claim per `ORCHESTRATION.md`), issues with open PRs, issues with unclosed blocking dependencies (`Blocked by #N` / `Depends on #N`), issues labeled `needs-human` or `needs-design`, and issues under cross-run cooldown.
+   - Skip assigned issues (unless stale claim per `ORCHESTRATION.md`), issues with open PRs, issues with unclosed blocking dependencies (`Blocked by #N` / `Depends on #N`), issues labeled `needs-human`, `needs-design`, or `needs-info`, and issues under cross-run cooldown.
 9. If no eligible candidate exists, STOP — run is SUCCESS with "No unclaimed work available".
 10. If the candidate should be closed already (work done, PRs merged), close it and return to step 8.
 10a. **Stale-claim reclamation:** If candidate is a stale claim per `ORCHESTRATION.md`, re-read immediately before writing, unassign the dead owner, post reclamation comment, and proceed to claim.
@@ -103,6 +107,11 @@ b. **Eligible → claim, then implement.** Call `get_me` once to learn your own
     - Read the issue description, linked code, and comment thread.
     - If bug: use `/diagnosing-bugs` to establish reproduction test before fixing.
     - If large/complex: use `/domain-modeling` and `/codebase-design`.
+    - **Ambiguity & Missing Acceptance Criteria Gate**: Challenge underspecified or incomplete requests before writing any code. If the issue lacks observable acceptance criteria, relies on unverified assumptions, or leaves critical technical/UX decisions ambiguous:
+      - Do NOT guess or invent arbitrary requirements to force completion.
+      - Post a comment on the issue posing 1–3 focused clarifying questions that identify the exact decisions or trade-offs needed.
+      - Apply the `needs-info` label and release the claim (unassign).
+      - Select the next candidate (evaluating ambiguous issues counts toward step 12's infeasible-continuation cap).
     - **Intent vs. Defect Guardrail**: When investigating issues related to low conversion, zero-click events, or underperforming features: verify whether the issue is a software defect or a lack of user intent. If data indicates the root cause is **lack of user intent** (e.g. button is rendered above fold and functions correctly when clicked, but user interaction rate is <2%) rather than a software defect, do NOT fall into the **telemetry rabbit hole** (adding elaborate fallback telemetry, downstream error handling, or defensive rendering). Categorize the issue as a **product/UX question** (`needs-design` / `roadmap/*`), comment explaining the lack of user intent, release the claim (unassign), and select the next candidate.
     - If infeasible: comment explaining blocker, release claim (unassign), and select next candidate (up to 3 infeasible evaluations per run). If permanent blocker on 2nd strike, apply `needs-human` label and tag repo owner.
 12a. **Umbrella-issue handoff + batching:** If candidate is an umbrella epic:

package/templates/skills/grill-me/SKILL.md

@@ -0,0 +1,64 @@
+---
+name: grill-me
+description: Stress-test proposals, requirements, and designs through a targeted, sequential interview to eliminate assumptions before implementation.
+---
+
+# Grill Me
+
+Stress-test plans, requirements, and designs through an active, sequential interview. The agent acts as an inquisitor—a cross between a principal engineer and a rigorous technical product manager.
+
+Your goal is **not** to validate or agree with the user. Your goal is to **uncover gaps, challenge necessity, expose hidden edge cases, and force clear decisions** before any code is written or formal spec drafted.
+
+## The Grilling Mindset
+
+- **Default to skepticism**: Ask why this needs to be built at all. Can this be solved with existing codebase primitives, configuration, or by deleting code?
+- **Zero-guesswork discipline**: Do not guess what the user meant when requirements are ambiguous. Force explicit choices.
+- **One to two questions per turn**: Never dump a questionnaire or a wall of questions. Ask at most 1–2 high-leverage questions at a time, provide concrete trade-offs or recommended options, and wait for the response before proceeding.
+
+## Process
+
+Execute these phases in sequence:
+
+### Phase 1: Challenge the Premise & Scope
+
+1. **Problem vs. Symptom**: Is the user solving the real root problem or patching a surface symptom?
+2. **Alternative & Simplicity**: Could this be handled without new abstractions? What is the simplest thing that could possibly work?
+3. **Necessity & ROI**: What happens if we do nothing? Who specifically benefits from this change?
+
+Probe the premise:
+- "What exact problem does this solve, and why can't we solve it with [existing module/primitive]?"
+- "What is explicitly *out of scope* for this change?"
+
+### Phase 2: Stress-Test Technical Seams & Failure Modes
+
+Once the premise is solid, probe failure modes and system boundaries:
+
+1. **Failure & Degraded States**: What happens if network calls fail, services time out, or dependencies crash?
+2. **Boundary & Empty States**: What happens with empty data, maximum limits, concurrent writes, or conflicting inputs?
+3. **Security & Auth**: Are there permission gates, auth boundaries, or data leakage risks?
+4. **Data Model & Schema**: Does this change existing contracts? Is it backward-compatible?
+
+### Phase 3: Establish Concrete Acceptance Criteria
+
+Force observable, testable definitions of done:
+
+1. **Test Seams**: At which seam (unit, integration, end-to-end) will this be verified?
+2. **Observable Outcomes**: What is the exact input and expected output? How will a test or user prove this works?
+3. **Negative Paths**: What negative test cases must fail safely?
+
+### Phase 4: Hand-off & Synthesis
+
+When all critical ambiguities are resolved:
+1. Summarize the agreed scope, explicit non-goals, architectural decisions, and acceptance criteria in a concise brief.
+2. Direct the next action:
+   - For formal requirements: run `/to-spec` or author an agent brief.
+   - For domain terminology: run `/domain-modeling`.
+   - For immediate implementation: proceed test-first via `/tdd`.
+
+## Completion Criteria
+
+The grilling session is complete only when:
+- [ ] Core motivation and explicit non-goals are established
+- [ ] Edge cases and failure modes have explicit decisions recorded
+- [ ] Acceptance criteria and verification seams are concrete and testable
+- [ ] The user confirms alignment with the summarized decisions

package/templates/skills/grill-me/agents/openai.yaml

@@ -0,0 +1,5 @@
+interface:
+  display_name: "Grill Me"
+  short_description: "Stress-test proposals, requirements, and designs"
+policy:
+  allow_implicit_invocation: false

package/templates/skills/triage/SKILL.md

@@ -73,7 +73,7 @@ Show counts and a one-line summary per item. Let the maintainer pick.
 
 3. **Verify the claim.** Before any grilling, check that the claim holds up. For a bug, reproduce it from the reporter's steps. For a PR, confirm the diff does what it claims: check it out, run the relevant tests or commands. Report what happened: confirmed (with code path), failed, or insufficient detail (a strong `needs-info` signal). A confirmed verification makes a much stronger agent brief.
 
-4. **Grill (if needed).** If the request needs fleshing out, call the Skill tool twice, for "grilling" and "domain-modeling", and grill it into shape a round of questions at a time, sharpening domain terms and updating `CONTEXT.md`/ADRs inline as decisions land.
+4. **Grill (if needed).** If the request needs fleshing out, call the Skill tool twice, for "grill-me" and "domain-modeling", and grill it into shape a round of questions at a time, sharpening domain terms and updating `CONTEXT.md`/ADRs inline as decisions land.
 
 5. **Apply the outcome:**
    - `ready-for-agent`: post an agent brief comment ([AGENT-BRIEF.md](AGENT-BRIEF.md)).

package/templates/workflows/trigger-review-routine.yml

@@ -107,12 +107,13 @@ jobs:
           fi
 
           # Priority-driven dual execution filter on direct pull_request triggers:
-          # If PR is labeled priority/P2 or priority/P3, skip immediate cloud review to let local agent daemons review it
+          # Strict allowlist: only trigger immediate cloud review if explicitly labeled priority/P0 or priority/P1.
+          # Untagged PRs and lower-priority PRs (P2/P3) skip immediate cloud review and are handled by local peer-review daemons
           # (cloud scheduled watchdog sweep will catch up if unreviewed > 48h).
           if [ "${{ github.event_name }}" = "pull_request" ]; then
             labels=$(echo "$pr" | jq -r '.labels[].name' 2>/dev/null || echo "")
-            if echo "$labels" | grep -qE '^priority/P[23]$'; then
-              echo "PR is labeled as lower priority (P2/P3) — skipping immediate cloud review for local peer-review daemon."
+            if ! echo "$labels" | grep -qE '^priority/P[01]$'; then
+              echo "PR is not explicitly labeled priority/P0 or priority/P1 — skipping immediate cloud review for local peer-review daemon."
               echo "skip=true" >> "$GITHUB_OUTPUT"
               exit 0
             fi