perf-testlisted

<!-- Generated from https://github.com/nielsmadan/agentic-coding — edits here are overwritten. --> # Performance Test Set up, run, and analyze performance tests with an iterative improvement cycle. For occasional manual runs, use the shared [manual test record convention](../doc/references/manual-tests.md) under `docs/tests/`. Routine automated benchmark runs keep their normal runner/CI output. If the request is only to save a completed run, use `doc --session` with that convention; capture the available evidence without starting a new test or optimization cycle. ## Usage ``` /perf-test src/utils/parser.ts # Profile a file/module /perf-test calculateTotal # Profile a specific function /perf-test /api/users # Load test an endpoint /perf-test checkout-flow # E2E scenario test /perf-test # Ask what to test ``` ## Gotchas - First runs include JIT/cold-start overhead. Always configure warmup passes in the proposal — without them, baseline measurements are inflated and improvement percentages are misleading. - Before/after comparisons are invalid if run on different machines, different concurrency levels, or with different iteration counts. Lock test configuration between runs. ## Workflow ### Phase 1: Understanding Look for an existing procedure and baseline under `docs/tests/` (or legacy `docs/perf/`). Reuse its scripts, configuration, and comparable baseline when they cover the