pubmed-database

Pass

Audited by Gen Agent Trust Hub on Jun 27, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill facilitates data retrieval from the official National Center for Biotechnology Information (NCBI) E-utilities REST API at eutils.ncbi.nlm.nih.gov. This is a well-known and trusted source for biomedical literature.
  • [COMMAND_EXECUTION]: Provides Python code templates using the requests library to execute HTTP queries against the PubMed database for automated literature analysis.
  • [PROMPT_INJECTION]: The skill possesses an indirect prompt injection attack surface because it processes external data retrieved from the PubMed database (such as article abstracts and titles).
  • Ingestion points: Data enters the agent's context through API endpoints including esearch.fcgi and efetch.fcgi as described in references/api_reference.md.
  • Boundary markers: There are no explicit delimiters or instruction-isolation markers provided in the example implementation logic.
  • Capability inventory: The skill is capable of network operations via requests.get to communicate with the NCBI infrastructure.
  • Sanitization: The provided examples do not include specific sanitization steps for the retrieved text, which is typical for skills intended to process scientific literature.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 27, 2026, 11:42 AM
Security Audit — agent-trust-hub — pubmed-database