scoring
Pass
Audited by Gen Agent Trust Hub on May 8, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The content of this skill is purely instructional and descriptive, consisting of scoring criteria and evaluation examples. There is no executable code or logic that would pose a security threat.
- [DATA_EXPOSURE]: While the rubric contains a penalty rule for 'hardcoded secrets' in settings files, the skill itself does not contain or expose any actual credentials, API keys, or sensitive configuration data.
- [COMMAND_EXECUTION]: The file mentions dangerous command patterns such as 'rm -rf' and 'DROP TABLE' within the context of scoring criteria for 'Hooks', but these are provided as static examples of unsafe behavior to be penalized rather than instructions for the agent to perform.
Audit Metadata