performance-profiler
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides legitimate instructions and scripts for performance engineering, adhering to a 'measure-first' methodology.
- [SAFE]: Python utility scripts (
benchmark_reporter.py,bottleneck_detector.py,resource_analyzer.py) perform standard data processing and statistical regression on JSON inputs without dangerous execution patterns. - [SAFE]: References to elevated privileges (e.g.,
sudo py-spy) are documented specifically for low-level process attachment on restricted environments, which is standard behavior for profiling tools. - [SAFE]: External dependencies and tools recommended for installation (Clinic.js, memray, k6) are well-known, trusted utilities from established organizations.
- [SAFE]: Prompt instructions are strictly diagnostic and do not attempt to bypass safety guardrails or persist across sessions.
Audit Metadata