About perf-profiler
A skill for running profiling tools and recording performance hotspots with supporting evidence. It is intended for developers investigating a defined performance scenario and keeps results concise.
Key features
- 01Runs a selected profiling tool and command
- 02Requires debug-symbol verification
- 03Records hotspot file and line numbers
- 04Provides a flame graph or equivalent when possible
Use cases
- 01Profile a defined performance scenario
- 02Locate and document code hotspots
- 03Capture evidence for performance investigation
Install
Copy the command and run it in a terminal.
$
npx skills add https:/ / github.com/ agent-sh/ agentsys --skill perf-profilerAfter installing, type/perf-profilerin Claude Code to confirm it loaded.
Details
From the project READMEperf-profiler
Run profiling tools and capture hotspots with evidence.
Follow docs/perf-requirements.md as the canonical contract.
Parse Arguments
const args = '$ARGUMENTS'.split(' ').filter(Boolean);
const tool = args[0] || '';
const command = args.slice(1).join(' ');
Required Rules
- Verify debug symbols before profiling.
- Capture file:line for hotspots.
- Provide flame graph or equivalent output when possible.
Output Format
tool: <profiler>
command: <command>
hotspots:
- file:line - reason
artifacts:
- <path to flame graph or profile>
Constraints
- No profiling without a clear scenario.
- Keep outputs minimal and evidence-backed.
Comments
Markdown supported