This teaches Claude how to handle code review feedback without the performative "You're absolutely right!" responses that violate its guidelines. The core workflow is verify before implementing, ask before assuming, and push back with technical reasoning when reviewers are wrong. It distinguishes between trusted feedback from your partner (implement after understanding) and external reviewers (check against the actual codebase first, apply YAGNI, verify it won't break things). The enforcement is strict on banning gratitude theater and requiring clarification of all unclear items before touching any code. Honestly most useful for the specific anti-patterns it calls out, like implementing suggestions that break existing functionality or add unused features just because a reviewer suggested them.
npx skills add https://github.com/sickn33/antigravity-awesome-skills --skill receiving-code-review