This one walks you through finding performance bottlenecks using CPU profiling, flame graphs, and benchmarking across multiple languages. You get working examples for Node.js perf_hooks, Chrome DevTools profiling, Python's cProfile, and database query analysis. The approach is sensible: profile first, focus on hot paths, measure everything. What I like is it includes both the instrumentation code and the tooling to actually visualize what's slow, not just raw timing data. The best practices are refreshingly pragmatic about not optimizing cold paths or sacrificing readability for negligible gains. Solid reference if you're investigating slow requests or need to justify optimization work with real data instead of hunches.
npx skills add https://github.com/aj-geddes/useful-ai-prompts --skill profiling-optimization