omero-integration
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external OMERO servers, which creates a potential surface for indirect prompt injection via image metadata and annotations.
- Ingestion points: The skill retrieves image data and associated metadata, including tags and key-value pair annotations, through the
BlitzGatewayinSKILL.md(e.g.,image.listAnnotations(),conn.getObject(), andconn.getRoiService()). - Boundary markers: There are no explicit delimiters or instructions provided to the agent to treat retrieved metadata as untrusted or to ignore embedded instructions within these data fields.
- Capability inventory: The skill has the capability to perform network operations (connecting to OMERO servers) and write to the file system (e.g., using
tifffile.imwriteto save processed images as described in the workflows inSKILL.md). - Sanitization: No sanitization, validation, or escaping of the retrieved metadata or annotation values is performed before the data is potentially processed or presented to the agent.
Audit Metadata