Rigorous code-review discipline — evaluate feedback technically instead of agreeing performatively, request reviews via a code-reviewer subagent, and gate any completion claim behind fresh verification evidence.
---
name: code-review
description: Use when receiving code-review feedback (especially if unclear or technically questionable), when completing tasks or major features that need review, or before making any completion/success claims. Covers receiving feedback with technical rigor, requesting reviews via a code-reviewer subagent, and verification gates requiring evidence before status claims.
---
Three code-review practices anchored on one principle: technical correctness over social comfort — verify before implementing, ask before assuming, evidence before claims.
## How to use
1. Receiving feedback: READ -> UNDERSTAND -> VERIFY -> EVALUATE -> RESPOND -> IMPLEMENT; stop and ask when items are unclear, and verify external-reviewer suggestions technically before applying.
2. Requesting review: invoke the code-reviewer subagent after each task in subagent-driven work, on finishing a major feature/refactor, and before merging to main.
3. Verification gates: before claiming tests pass, the build succeeds, or work is done, run the verification command and state the claim WITH evidence — no fresh evidence means no claim.
4. Use the quick decision tree to route each situation; consult the matching `references/` file (reception, requesting, verification-before-completion) for the full protocol.
Full skill & source: https://github.com/mrgoonie/claudekit-skills/tree/80113d86bc4407f105af40a2c4ea58194f7c370a/.claude/skills/code-reviewSign in to rate and review this skill.
No reviews yet. Be the first to review this skill.