--- description: Structured performance investigation with baselines, profiling, and evidence-backed decisions agent: general --- # /perf - Performance Investigation Workflow Run a rigorous, evidence-driven performance investigation with strict rules, baselines, and reproducible benchmarks. ## Canonical Requirements All behavior must follow: - `docs/perf-requirements.md` (source of truth) - `docs/perf-research-methodology.md` ## Arguments - `--resume`: Continue the latest investigation from `{state-dir}/perf/investigation.json` - `--phase `: Force starting phase (use only when resuming) - `--id `: Set investigation id (new only) - `--scenario `: Short scenario description - `--command `: Benchmark command (prints PERF_METRICS markers) - `--version `: Baseline version label - `--duration `: Benchmark duration override (default 60s; use smaller values for micro-benchmarks) - `--runs `: Number of runs for start-to-end benchmarks (use with median aggregation) - `--aggregate `: Aggregation method for multi-run benchmarks (default median) - `--quote `: User quote to record in logs - `--hypotheses-file `: JSON file with hypothesis list (for hypotheses phase) - `--param-env `: Env var for breaking-point value (default PERF_PARAM_VALUE) - `--param-min `: Breaking-point min value (default 1) - `--param-max `: Breaking-point max value (default 500) - `--cpu `: Constraint CPU limit (default 1) - `--memory `: Constraint memory limit (default 1GB) - `--change

`: Optimization change summary - `--verdict `: Decision verdict - `--rationale `: Decision rationale ## Phase 1: Initialize Investigation State *(JavaScript reference - not executable in OpenCode)* ## Output - Updated `{state-dir}/perf/investigation.json` - Investigation log at `{state-dir}/perf/investigations/.md` - Baseline files at `{state-dir}/perf/baselines/.json` Begin the performance investigation now.