Runs a structured five-axis code review covering correctness, readability, architecture, security, and performance before any merge. The approval bar is pragmatic: code should improve overall health, not be perfect. It enforces change sizing (aim for 100 lines, split anything over 1000), requires useful commit messages beyond "fix bug," and categorizes feedback by severity so you know what's blocking versus what's a nit. The multi-model review pattern is clever: one model writes, another reviews, first model addresses feedback, then human decides. Best used before merging PRs, after feature work, or when evaluating code from another agent. Treat this as your pre-merge checklist, not a style guide.
npx skills add https://github.com/addyosmani/agent-skills --skill code-review-and-quality