The Agent Skills Directory

[SAFE]: The skill uses a local script for all processing tasks. It correctly identifies the script path and enforces its use, reducing the risk of arbitrary code generation by the agent. No network operations or sensitive file accesses were detected.\n- [REMOTE_CODE_EXECUTION]: Static analysis identified the use of importlib.import_module in scripts/analyze_patents.py. This is used to load standard libraries like matplotlib, plotly, and seaborn using static strings, which is a safe practice and does not allow for arbitrary code execution.\n- [PROMPT_INJECTION]: The skill processes untrusted patent data from Excel files. The script implements proper sanitization using html.escape for the HTML dashboard output, addressing potential indirect injection vectors in the generated content.

patent-analysis-viz