pubmed-database
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to retrieve and process research data from the PubMed database, which creates a surface for indirect prompt injection if retrieved content contains instructions designed to influence the agent behavior.\n- Ingestion points: PubMed API responses containing article metadata and abstracts in SKILL.md and references/api_reference.md.\n- Boundary markers: The provided documentation lacks specific instructions or code patterns for using delimiters to isolate untrusted external content.\n- Capability inventory: The skill uses the requests library for network access and suggests local file system operations for caching data.\n- Sanitization: Example code snippets do not demonstrate sanitization or validation of the data retrieved from the API.\n- [EXTERNAL_DOWNLOADS]: The skill facilitates communication with the official NCBI E-utilities API at eutils.ncbi.nlm.nih.gov. This is a trusted and well-known service for biomedical data.
Audit Metadata