protocolsio-integration
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill retrieves protocol titles, descriptions, and step-by-step instructions from the external protocols.io API. This data is processed by the agent and can be exported to local files.
- Ingestion points: Protocol data is fetched via GET requests to
https://www.protocols.io/api/v4/protocolsas defined inSKILL.md. - Boundary markers: The skill does not implement specific boundary markers or instructions to treat the retrieved protocol text as untrusted content.
- Capability inventory: The skill includes functionality to write data to the local file system using
pandas.to_csvfor storing protocol steps and materials, as demonstrated in the 'Workflow 2' section ofSKILL.md. - Sanitization: No explicit sanitization or validation of the protocol content (e.g., stripping HTML or checking for embedded instructions) is performed before the data is processed or saved.
Audit Metadata