This skill provides an automated, deep-dive code review of your current working branch compared to your base branch (main/master). It systematically analyzes your diff across six critical dimensions: correctness, security, performance, code quality, test coverage, and documentation.
Manual code reviews are time-consuming and prone to human oversight. While standard LLM prompting might catch surface-level issues, this skill follows a rigorous multi-step engineering process: it identifies the merge-base, reads full file context (not just the diff), analyzes commit history quality, and categorizes findings by technical severity. It replaces vague feedback with structured, actionable PR_REVIEW.md files that include specific code suggestions.
You receive a professional-grade PR_REVIEW.md file categorized by severity (CRITICAL to INFO). It identifies specific line numbers, explains the underlying risk (e.g., N+1 queries, XSS vulnerabilities, or race conditions), and provides logic-ready code blocks for quick fixes.