Menu
Multi-agent deep review for code PRs in any repo. Use when asked to "deep review this PR," "multi-agent review," "review
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Generate a new Claude Code skill from a plain-language description — decides invocation control (disable-model-invocation vs user-invocable), arguments (argument-hint, $ARGUMENTS), and context cost, then scaffolds and tests it.
Audit an LLM evaluation or benchmark repo for integrity and credibility practices. Use when asked to "audit my benchmark," "is my eval trustworthy," "check my leaderboard for contamination," "review this benchmark's methodology," or "what would a reviewer attack in my eval." Greps the target repo for evidence across seven dimensions (pre-registration, contamination, holdout hygiene, judge validity, statistical honesty, reproducibility, leaderboard exclusions) and emits a scored report with file:line evidence, severity, and concrete fixes.
Tripwire check for multi-session drift. Scans state files, recent commits, and file conflicts caused by parallel Claude Code sessions.
Scan for orphaned worktrees and stale branches after crashes or abandoned sessions. Offers safe cleanup options.
Structured session lifecycle for Claude Code — start, checkpoint, end, and daily heartbeat commands that maintain project state across conversations.
Single-source-of-truth drift auditor for documentation-heavy repos. Use when asked to "check for drift," "find copies of this number," "audit the docs for stale facts," or "set up an SSOT manifest." Finds facts hand-copied across files, builds a manifest of canonical locations, and verifies every copy still matches.
| name | code-review |
| description | Multi-agent deep review for code PRs in any repo. Use when asked to "deep review this PR," "multi-agent review," "review |
| argument-hint | [pr-number-or-branch] |
User-invokable orchestrator for code-PR review across any repo: Copilot for line-level findings, parallel subagents for the architectural and operational misses line-level review can't see. (Validation evidence and origin story: README.)
Fallbacks: no gh or not a GitHub repo → review the local diff (git diff <base>...HEAD) and skip the Copilot lane entirely; no subagent tool → run the review angles sequentially in the main thread. Detect what's available before assuming Copilot access.
Invocation: deliberately model-invocable — "review this PR" phrasing is the trigger. Pushes and issue-creation inside the flow are confirmation-gated.
The review target is $ARGUMENTS (a PR number, branch, or URL). If empty, infer it from the conversation or ask.
gh pr view N --json files,additions,deletions,title
Categorize:
| Tier | Examples | Treatment |
|---|---|---|
| Trivial | Typo, copy edit, single-line deps bump, README polish | Skip multi-agent. Copilot alone is enough. |
| Standard | Component refactor, single-feature bug fix, contained logic change | Copilot + 1 subagent (pick the angle that matches the risk shape) |
| High-stakes | Payment, auth, crypto/wallet, RPC, deployment configs, scale-relevant changes, new external SDK integrations | Copilot + 2 parallel subagents minimum (always include adversarial + operational) |
If trivial → stop reading this skill, just request Copilot review and ship.
Three angles that empirically work. Pick by risk shape:
Spawn all chosen subagents in a single tool-call batch so they run in parallel. Prompt templates live in patterns/subagent-prompts.md.
Every subagent prompt must specify:
gh pr edit N --add-reviewer copilot-pull-request-reviewer
Don't wait for Copilot to start before spawning subagents. They run in parallel.
Three buckets:
Real findings — fix now. Push fix to the PR branch. Pushing changes external state: unless the user already asked you to fix or babysit this PR, confirm before the first push of the session.
Stale re-flags — Copilot will sometimes re-flag findings against unmoved diff lines after fixes ship at HEAD. Always read the file at HEAD before re-fixing. If the fix is already there, drop a one-line reply on the comment ("Fixed at ") and move on. Do not re-fix.
Deferred (filed) — tests, refactors, or work outside scope. File as a GitHub issue (confirm with the user before creating it), link the issue number in the merge commit message, move on. Don't expand the PR.
Maximum 2 Copilot review rounds. Empirically, by round 3 the signal-to-noise ratio collapses (~50% stale re-flags). Ship after round 2. If subagents flag genuinely new issues after round 2, those are issue-tracker items, not PR-blockers.
/security-review if you suspect a problem.Validation case (5 rounds on a high-stakes PR):
maxDuration not set — function would time out under realistic load. Architectural miss invisible at the line level.bodyParser size limit (DoS surface).Subagents are not redundant with Copilot. They cover a different layer.
patterns/subagent-prompts.md — paste-ready prompt templates for the three angles.