librarian
Pass
Audited by Gen Agent Trust Hub on Aug 26, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process information from external sources like GitHub, Stack Overflow, and technical blogs, which constitutes a surface for indirect prompt injection.\n
- Ingestion points: The 'Research Approach' and 'Information Sources Priority' sections explicitly direct the agent to fetch data from third-party web sources.\n
- Boundary markers: The instructions do not define specific delimiters to isolate untrusted external content from the agent's core instructions.\n
- Capability inventory: The skill's purpose implies the use of web search or HTTP retrieval tools to access documentation and repositories.\n
- Sanitization: There are no instructions to validate or sanitize the retrieved content before it is incorporated into the agent's response context.
Audit Metadata