The Agent Skills Directory

[SAFE]: The skill facilitates performance investigations through a well-defined 'Optimization Contract' that prevents arbitrary repository modifications and enforces evidence-based decision making.
[COMMAND_EXECUTION]: The skill runtime utilizes subprocess.run to call internal validation scripts. This is managed via a static manifest that uses secure path expansion, effectively neutralizing command injection risks.
[SAFE]: Repository file access is performed to verify evidence anchors. The implementation includes strict path validation to prevent access to files outside of the intended repository root.
[SAFE]: The skill does not include external dependencies or remote script execution, relying entirely on the provided package scripts and the Python standard library.

optimizations