perf-benchmarker
Pass
Audited by Gen Agent Trust Hub on Mar 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns, obfuscation, or sensitive data exposure were detected. The skill consists entirely of markdown instructions.
- [COMMAND_EXECUTION]: The skill is designed to guide an agent through executing performance benchmark commands. This behavior is consistent with the stated purpose of establishing baselines and validating regressions.
- [INDIRECT_PROMPT_INJECTION]: The skill identifies an attack surface where it processes external data from benchmark outputs. Evidence chain:
- Ingestion points: The agent reads the standard output/error from executed benchmark commands.
- Boundary markers: The skill explicitly defines 'PERF_METRICS_START' and 'PERF_METRICS_END' markers to isolate structured data, which serves as a security best practice for parsing external content.
- Capability inventory: The skill facilitates the execution of arbitrary commands as defined in the 'argument-hint'.
- Sanitization: No explicit sanitization or filtering of the command output is described beyond the use of boundary markers.
Audit Metadata