What problem does it solve?
This Skill solves performance regressions and slow user experiences by turning optimization into a disciplined MEASURE → IDENTIFY → FIX → VERIFY → GUARD workflow across frontend and backend.
Core Features & Use Cases
- Performance measurement & baselining: Establish real-world baselines using synthetic testing (e.g., Lighthouse) and real user monitoring (RUM), plus APM and DB timing to quantify p50/p95/p99.
- Bottleneck identification: Pinpoint whether slowness comes from frontend Core Web Vitals issues (LCP/INP/CLS) or backend latency causes such as N+1 queries, missing indexes, connection-pool saturation, or async/CPU bottlenecks.
- Targeted fixes with verification: Use
EXPLAIN ANALYZE (Postgres) / EXPLAIN FORMAT=JSON (MySQL) and profiling tools like py-spy / cProfile to change only what measurements prove matters, then re-measure to confirm the improvement.
- Regression guardrails: Add monitoring and CI checks (bundle budgets, Lighthouse-CI, alerting at ~80% of budgets) so improvements don’t decay after the next release.
Use case example: You notice p95 API latency spiking after a release; use this Skill to measure the specific endpoint and trace it to N+1 query patterns, apply a batching/join/eager-loading fix, then validate the new p95 drops into the agreed latency budget.
Quick Start
Use the performance-optimization skill when you have a slow endpoint or Core Web Vitals regression and want a MEASURE → IDENTIFY → FIX → VERIFY → GUARD plan that includes profiling, targeted fixes (e.g., N+1 and indexing), and a regression guard.
