biorxiv-database

Pass

Audited by Gen Agent Trust Hub on Jul 10, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructions include installing standard, well-known Python packages requests and pandas.
  • [EXTERNAL_DOWNLOADS]: The skill fetches research data and documents from established scientific domains including api.biorxiv.org, www.biorxiv.org, and www.ebi.ac.uk (EuropePMC).
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests and processes untrusted external data from public preprint servers.
  • Ingestion points: External metadata (titles, abstracts, author names) is retrieved via API calls in SKILL.md.
  • Boundary markers: No delimiters or explicit instructions are provided to the agent to ignore potentially malicious content embedded in the retrieved preprint metadata.
  • Capability inventory: The skill possesses the capability to write this external data to local storage using pandas.to_csv and binary file writes in SKILL.md.
  • Sanitization: There is no evidence of string validation, escaping, or filtering of the external content before it is processed or saved to the filesystem.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 10, 2026, 03:35 PM
Security Audit — agent-trust-hub — biorxiv-database