What problem does it solve?
Manual code reviews miss structural issues like SQL injection, race conditions, and LLM trust boundary violations that automated tests don't catch. This Skill runs a systematic pre-landing review so critical bugs are caught before merge, not in production.
Core Features & Use Cases
- Two-Pass Static Analysis: Scans diffs for SQL safety, shell injection, race conditions, enum completeness, and LLM output trust boundaries in a structured critical-then-informational workflow.
- Specialist Subagents: Spawns parallel security, performance, testing, maintainability, and data migration specialists for deep-dive analysis on large diffs.
- Scope Drift Detection: Compares the actual diff against TODOS.md and PR descriptions to catch scope creep and missing requirements before the review begins.
- Use Case: You're about to merge a 300-line branch that touches payments. Run this Skill to automatically flag N+1 queries, missing parameterized SQL, and incomplete error handling before the code reaches production.
Quick Start
Use the review skill to analyze the current branch diff against the base branch and report any critical or informational issues found.