perflisted

# Perf Skill > Quick Ref: `$perf profile <target>` | `$perf bench <target>` | `$perf compare <baseline> <candidate>` | `$perf optimize <target>` **YOU MUST EXECUTE THIS WORKFLOW. Do not just describe it.** Performance profiling, benchmarking, regression detection, and optimization recommendations for any language runtime. Produces actionable metrics, not vague advice. ## Modes | Mode | Command | Purpose | |------|---------|---------| | **Profile** | `$perf profile <target>` | Profile execution, find hotspots | | **Benchmark** | `$perf bench <target>` | Create or run benchmarks | | **Compare** | `$perf compare <baseline> <candidate>` | Compare two runs for regression | | **Optimize** | `$perf optimize <target>` | Analyze and apply optimizations | If no mode is specified, default to **profile**. --- ## Step 0: Detect Language and Tooling Identify the language/runtime from file extensions, `go.mod`, `package.json`, `pyproject.toml`, `Cargo.toml`, or explicit user input. Select the profiling stack: | Language | Benchmarking | CPU Profile | Memory Profile | Comparison | |----------|-------------|-------------|----------------|------------| | **Go** | `go test -bench` | `go tool pprof` (cpu) | `go tool pprof` (alloc) | `benchstat` | | **Python** | `pytest-benchmark`, `timeit` | `cProfile`, `py-spy` | `memory_profiler`, `tracemalloc` | manual diff | | **Node** | `benchmark.js`, `vitest bench` | `--prof`, `clinic.js` | `--heap-prof`, `0x` | manual diff | | **Rust** | `crite