histolab
Pass
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: LOW
Full Analysis
- SAFE (SAFE): The provided documentation describes legitimate data science workflows for processing whole slide images (WSI).
- Category 2 (Data Exposure): No sensitive file access or network exfiltration patterns were found. The use of local paths for slide images and output directories is standard for image processing tasks.
- Category 4 (RCE): All imports refer to well-known libraries (histolab, matplotlib, numpy, opencv). There are no instances of remote script execution or arbitrary command invocation.
- Category 8 (Indirect Prompt Injection): While the skill involves processing external data (image files), it is for deterministic image analysis and visualization. There is no natural language processing of untrusted data that could lead to agent instruction override.
Audit Metadata