Perf benchmarker
/SKILLUse when running performance benchmarks, establishing baselines, or validating regressions with sequential runs. Enforces 60s minimum runs (30s only for binary search) and no parallel benchmarks.
--- name: perf-benchmarker description: "Use when running performance benchmarks, establishing baselines, or validating regressions with sequential runs. Enforces a minimum run duration of 60 seconds (30 seconds only for binary search) and no parallel benchmarks." version: 1.0.0 argument-hint: "<command> [duration]" --- # perf-benchmarker Run sequential benchmarks with strict duration rules. Follow docs/perf-requirements.md as the canonical contract. ## Required Rules - Benchmarks MUST run sequentially (never in parallel). - Minimum duration: 60s per run (30s only for binary search). - Warmup: At least 10s before measurement. - Re-run anomalies. ## Output Format `` command: <benchmark command> duration: <seconds> warmup: <seconds> results: <metrics summary> notes: <anomalies or reruns> ` ## Output Contract Benchmarks MUST emit a JSON metrics block between markers: ` PERF_METRICS_START {"scenarios":{"low":{"latency_ms":120},"high":{"latency_ms":450}}} PERF_METRICS_END `` ## Constraints - No short runs unless in the binary-search phase. - Do not change the code while benchmarking.
