alterlab-figure-qa
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is a specialized utility for scientific figure QA. It operates locally on provided data and figures without requiring network access, adhering to the principle of least privilege.
- [COMMAND_EXECUTION]: The skill uses Python and the
uvpackage manager to execute inspection scripts. This is the intended functionality for programmatically measuring Matplotlib objects, bounding boxes, and image resolutions. - [INDIRECT_PROMPT_INJECTION]: The skill processes external source data to verify plotting fidelity, creating a theoretical ingestion surface.
- Ingestion points: Source data (e.g., CSV/DataFrames) and rendered figure objects.
- Boundary markers: None explicitly defined in the prompt text.
- Capability inventory: Python execution (matplotlib/pillow) and file system read/write access.
- Sanitization: Relies on the standard behavior of the plotting libraries and the agent's runtime environment.
Audit Metadata