The Agent Skills Directory

[SAFE]: The skill is designed as an auditing utility for research rigor. All included scripts (Python and Bash) are deterministic, use standard libraries for JSON processing and file comparison, and do not exhibit malicious behavior.
[COMMAND_EXECUTION]: The skill uses local shell and Python scripts to validate report manifests. These scripts operate within a local scope and do not execute payloads or interact with the network.
[EXTERNAL_DOWNLOADS]: The skill references established open-source libraries such as Captum and SHAP for researchers to use, but it does not perform automated remote downloads or execution of external code.

explainability