The Agent Skills Directory

[SAFE]: The skill reads local metadata files (.maestro/audit.jsonl and .maestro/decisions.jsonl) created by the Maestro tool. This is the intended behavior for performance analysis and does not involve accessing sensitive user files like SSH keys or environment variables.
[SAFE]: No network activity, external downloads, or remote code execution patterns were identified in the analysis.
[SAFE]: The skill's instructions focus on statistical analysis of logged data (completion rates, duration, cost), which poses no threat to the underlying system or user data privacy.

reflect