performance-proof
Pass
Audited by Gen Agent Trust Hub on Jul 25, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: No attempts to override agent behavior or bypass safety guidelines were found. The instructions are focused on performance benchmarking and algorithmic selection.
- [DATA_EXFILTRATION]: No sensitive file access or unauthorized network operations were detected. The skill operates within the context of the user's repository for measurement and profiling.
- [COMMAND_EXECUTION]: The skill does not contain any arbitrary command execution patterns. It explicitly states that source edits require user authorization.
- [REMOTE_CODE_EXECUTION]: There are no patterns involving downloading and executing remote scripts or installing untrusted packages.
- [EXTERNAL_DOWNLOADS]: No external URLs or remote resources are referenced for download or execution.
Audit Metadata