| name | review-code |
| description | Run a multi-agent code review on recent changes. Spawns a team of specialized reviewers (bugs, security, perf, tests, usability, etc.), collects findings, triages into fix-now vs defer, and optionally applies fixes. Use when asked to "review code", "review changes", "review this PR", or "get a second opinion". |
Multi-Agent Code Review
Spawn parallel reviewer agents over recent changes; synthesize one triaged report.
Original skill by Bjorn Tipling.
Arguments
--apply — apply the "fix now" batch after presenting findings, without asking.--defer-file <path> — append deferred items to this file. Default: deferred items are not written to any file.
Scope
If the branch is ahead of main, review the full branch diff (git diff origin/main...HEAD); otherwise review uncommitted changes. Exclude vendor/, generated files, go.sum.
Select Reviewers
First check the project for domain-specific reviewer skills: skills/CATALOG.md and skills/*-review/SKILL.md (convention: names ending in -review are reviewer personas). Spawn any whose domain the changeset touches. Prefer these over generic defaults — they encode conventions generic reviewers miss. Their criteria prompt is: "Read {skill-path}/SKILL.md and apply its checklist verbatim. Flag any violations."
Generic roster — pick per what the code touches:
| Reviewer | Focus | Include when |
|---|
| bugs-reviewer | Logic bugs, edge cases, error handling, concurrency | Always — baseline |
| security-reviewer | Injection, path traversal, credentials, DoS |
