This one's about handling code review feedback without the performative "you're absolutely right!" responses that creep into Claude's default behavior. It pushes you to verify suggestions against the actual codebase before implementing, ask clarifying questions when feedback is unclear, and push back with technical reasoning when reviewers are wrong. The YAGNI check is smart: if someone suggests "implementing properly," grep for actual usage first. Works best when you're dealing with external reviewers who might lack full context, or when your human partner drops six fixes at once and two of them don't make sense yet. The core idea is technical correctness over social comfort.
npx skills add https://github.com/davila7/claude-code-templates --skill receiving-code-review