What problem does it solve? Manual code reviews are inconsistent and often miss architectural drift, performance traps like N+1 queries, and untested edge cases before code reaches production. This Skill performs a structured, adversarial review across the entire codebase and produces severity-ranked findings with concrete fixes. ## Core Features & Use Cases - Architecture Conformance Audit: Compares implementation against every ADR, API contract, and data ownership boundary, flagging violations with severity levels. - Multi-Dimensional Review: Runs parallel phases covering SOLID/DRY code quality, cyclomatic complexity, performance bottlenecks (N+1, unbounded queries, missing caching), and test quality (coverage gaps, weak assertions). - Actionable Reports: Writes findings to critical/high/medium/low files, metrics to JSON, and mechanical auto-fix suggestions as before/after patch files — all read-only, never modifying source code. - Use Case: After backend, frontend, and tests are complete, run this Skill as a quality gate to catch a race condition causing double charges or an unindexed query on a high-traffic endpoint before deployment. ## Quick Start Review the services, frontend, and tests directories against the architecture docs and produce a severity-ranked review report with findings and auto-fix suggestions.
