Code Review Checklist

This is a structured code review framework that gives Claude a practical checklist covering correctness, security, performance, and code quality. It includes AI-specific concerns like prompt injection protection and output sanitization, which is smart given how many codebases now integrate LLMs. The anti-pattern examples are concrete (magic numbers, deep nesting, any types) and the emoji-based review comment system (🔴 for blocking, 🟡 for suggestions) is a nice touch for consistent PR feedback. Use it when reviewing pull requests or doing security audits. It won't catch everything a human would, but it enforces a consistent baseline and the LLM-specific sections are genuinely useful for modern stacks.