alterlab-histolab
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external Whole Slide Image (WSI) files, which acts as a potential entry point for untrusted data.\n
- Ingestion points: The
Slideclass inSKILL.mdandreferences/slide_management.mdloads external files (e.g., SVS, TIFF, NDPI) into the agent's context for processing.\n - Boundary markers: The instructions do not define specific boundary markers or safety guidelines to mitigate the risk of malicious content or metadata embedded within the binary slide files.\n
- Capability inventory: The skill utilizes
Write,Edit, andBashcapabilities, which provide a significant surface for potential exploitation if an injection occurs through a compromised slide file.\n - Sanitization: The skill relies on the
histolabandOpenSlidelibraries for data parsing; no additional sanitization or metadata validation is implemented within the agent's instructions.
Audit Metadata