git:20261013.4a43eb8 to git:20261013.66fea96

12 added, 10 removed. Audit A to A.

---

name: muse-review

- description: Launch an independent Meta Muse Wonder High code review through Pi and Vercel AI Gateway, and return it verbatim. Use only when the user explicitly invokes /muse-review.

+ description: Launch an independent Meta Muse Wonder Max code review through Pi and OpenRouter, and return it verbatim. Use only when the user explicitly invokes /muse-review.

disable-model-invocation: true

triggers: [user, model]

---

# Muse Review

- Always use **Meta Muse Wonder at High reasoning** through the Pi harness and the `vercel-ai-gateway` provider, orchestrated by the orchestration tool. No other model, route, effort, or harness, ever. High is the highest effort this route offers.

+ Always use **Meta Muse Wonder at Max reasoning** through the Pi harness and the `openrouter` provider, orchestrated by the orchestration service. No other model, route, effort, or harness, ever. Only OpenRouter offers `max` for this model; Pi sends it through unchanged.

## Run

- Read the orchestration CLI instructions and follow its launch checks. Reuse the current environment, including uncommitted work, and coordinate through the parent thread.

+ Read the orchestration CLI's documentation and follow its launch checks. Reuse the current environment, including uncommitted work, and use its parent-session coordination option when coordinating.

- Always spawn with exactly these flags:

+ Always select exactly:

- ```bash

- --provider pi --model vercel-ai-gateway/meta/muse-wonder-1.0 --reasoning-level high

+ ```text

+ provider: pi

+ model: openrouter/meta/muse-wonder

+ reasoning level: max

```

- - Before spawning, confirm the orchestration CLI's model listing includes `vercel-ai-gateway/meta/muse-wonder-1.0` with `high` in its supported reasoning efforts.

- - After spawning, confirm in the thread details that it runs `vercel-ai-gateway/meta/muse-wonder-1.0` at `high`.

- - If either check fails, stop and report the blocker. Never fall back to another model, route (such as `openrouter/meta/muse-wonder`), effort, or harness.

+ - Before spawning, confirm the orchestration CLI's provider model listing includes `openrouter/meta/muse-wonder` with `max` among its supported reasoning efforts.

+ - After spawning, confirm through the orchestration CLI's thread details that the thread runs `openrouter/meta/muse-wonder` at `max`.

+ - If either check fails, stop and report the blocker. Never fall back to another model, route (such as `vercel-ai-gateway/meta/muse-wonder-1.0`), effort, or harness.

## Review brief

Give neutral context: scope, paths, intended behavior, and diff/base revision. Ask for a thorough senior-developer review, including related code and tests, without steering toward suspected bugs, solutions, or verdicts. Review only; do not change files. Request a concise plain-English report of serious or critical issues, fixes, and production merge readiness. Distinguish verified findings from uncertainty and identify validation gaps.

## Result

- Wait for the thread to finish and retrieve its output using the orchestration CLI. Verify this review actually completed; idle status, timeouts, gateway errors, and queued retries are not completion. Report blockers rather than presenting partial output as a finished review.

+ Wait for the review thread to finish and retrieve its output using the orchestration CLI. Verify this review actually completed; idle status, timeouts, gateway errors, and queued retries are not completion. Report blockers rather than presenting partial output as a finished review.

Return the completed reviewer's full final response verbatim.