cognee-permissions
Pass
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill describes methods for retrieving information from external datasets using
cognee.recall()andcognee.search(). This functionality creates an attack surface where malicious instructions embedded in the data could potentially influence the agent's actions. - Ingestion points: Data is pulled into the agent's context through search and recall operations as documented in
SKILL.md. - Boundary markers: The skill does not provide instructions for using delimiters or warnings to ignore instructions within the ingested data.
- Capability inventory: The skill facilitates data reading, data writing, and permission management capabilities.
- Sanitization: There are no documented sanitization or filtering steps for the data retrieved from Cognee datasets.
- [SAFE]: The skill serves as legitimate documentation for the Cognee permissions module. It includes security best practices such as warning against disabling backend access control in multi-tenant environments and explaining how to handle API keys and authentication tokens securely using environment variables.
Audit Metadata