review-hog-perspective-performance-reliability

Review the changed code for N+1 query patterns, missing database indexes, unnecessary re-renders, and synchronous operations that should be async. This Skill automates that pass using ripgrep searches over the PR chunk and reports findings with severity levels.

It checks resource efficiency, error handling and recovery, scalability patterns like caching and pagination, and operational readiness such as logging, metrics, timeouts, and health checks. Logic, security, and style concerns are handled by separate review perspectives.

N+1 problems are detected by searching for database queries or fetches inside loops, for example with ripgrep patterns matching loop constructs near query calls. The Skill flags these as performance findings, typically at must-fix or should-fix severity.

No. Findings are reported only for non-test application code. Vendor, third-party, and generated files are skipped except when needed for context, keeping the review focused on production-relevant changes.

Findings use three levels: must fix for issues causing outages or severe degradation, should fix for noticeable performance impact or reliability risk, and consider for minor optimizations or nice-to-have improvements.