The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from an external OMERO.server which could potentially contain malicious instructions intended to influence the agent's behavior.
Ingestion points: Data is retrieved from user-defined OMERO_HOST endpoints using the omero-py BlitzGateway and OMERO CLI (e.g., in scripts/inventory.py and scripts/export_image_metadata.py).
Boundary markers: The instructions emphasize using the "smallest explicit data scope" and "bounded every list, page, ROI, shape, annotation, table row, pixel plane, and local file scan." The json_safe utility in scripts/omero_common.py enforces string limits.
Capability inventory: The skill possesses network connectivity to the OMERO server, local file system write access (used for JSON reports), and the ability to execute OMERO CLI commands for data transfers.
Sanitization: The json_safe function in scripts/omero_common.py truncates strings to a default of 512 characters and limits collection sizes to 1,000 items. Error messages are scrubbed of potentially sensitive information using scrubbed_error.

omero-integration