The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes specialized Python scripts (e.g., model_buildvsbuy_calculator.py, ai_risk_classifier.py, ai_cost_economics.py) located within the vendor's directory structure. These scripts are used to process user-provided data files such as use_case.json and workload.json as part of the intended project review workflow.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data via user-supplied JSON files. While these files could theoretically contain malicious instructions designed to influence the agent's review verdict, this is a standard risk surface for diagnostic tools and the skill does not grant excessive privileges to the processed data.

caio-review