概要
This skill transforms how Claude handles code reviews by enforcing a strict 'verify before implementation' protocol. It prevents the common AI pitfall of blind agreement, ensuring that suggestions are technically sound, necessary for the specific codebase, and free of unnecessary 'YAGNI' features. By providing structured response patterns for both human partners and external reviewers, it shifts the focus from social pleasantries to technical acknowledgment and reasoned pushback when necessary.