Performance Optimization Skill

Entry-point guard: This is a user-invoked command skill (/perf-optimize). If you are a spawned subagent, do NOT execute this skill — return to your assigned scope card.

Trigger: User provides profiling data, asks to optimize performance, or benchmarks show regression.

Skill prerequisite: Read .agents/skills/perf-optimization/SKILL.md before starting.

Purpose

Profile-driven performance optimization using multi-dimensional parallel analysis. Dispatches subagents across orthogonal performance dimensions to analyze profiling data and scan code, then implements fixes sequentially with TDD and isolated benchmarks.

When to Use

• User provides profiling data (pprof, flamegraph, py-spy, Chrome DevTools)
• User asks to optimize performance of a specific component
• A benchmark regression is detected
• After deploying a new feature that touches a hot path