The Agent Skills Directory

[SAFE]: The skill primarily serves as a knowledge base and a local audit tool for model metadata. It provides detailed reference cards for medical imaging architectures and a Python utility for provenance checking.
[DYNAMIC_EXECUTION]: The test suite (tests/test_model_provenance.sh) utilizes exec() within Python one-liners to programmatically generate JSON dossier variants. This usage is restricted to the local testing environment for regression testing purposes and does not process external or untrusted input during the skill's normal operation.
[COMMAND_EXECUTION]: Shell scripts (verify.sh and test_model_provenance.sh) are included to validate the audit logic. These scripts execute the local Python audit tool against static fixtures and do not include patterns for remote code execution or exfiltration.

model-selection