Review

This is a code review skill that acts like a senior engineer looking over your changes. It checks for correctness, security issues, broken contracts, and whether your tests actually exercise the new code. Point it at a file, diff, or commit and it'll categorize findings as blockers, important issues, or nits. The instruction to be harder on AI-generated code is smart, since LLMs can sound confident while being subtly wrong. It won't auto-fix anything, which is the right call for a review tool. One notable detail: it'll call out when tests mock the function under test or just assert what the code did rather than what it should do, treating those as merge blockers. Supports custom review guidelines via REVIEW.md if your team has specific standards.

Install to Claude Code

npx -y skills add owainlewis/blueprint --skill review --agent claude-code