biorxiv-database

Pass

Audited by Gen Agent Trust Hub on Mar 3, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill downloads PDF documents from the official and well-known preprint repository at www.biorxiv.org.
  • [DATA_EXFILTRATION]: The skill fetches preprint metadata and paper abstracts from the official bioRxiv API at api.biorxiv.org to fulfill its primary search function.
  • [PROMPT_INJECTION]: The skill processes untrusted text from scientific abstracts, which presents a surface for indirect prompt injection.
  • Ingestion points: Preprint titles and abstracts are retrieved from api.biorxiv.org via scripts/biorxiv_search.py.
  • Boundary markers: The skill does not use specific delimiters or instructions to isolate retrieved text from agent directives.
  • Capability inventory: The script includes functionality to write files to the local system through its download_pdf method.
  • Sanitization: No filtering or sanitization is applied to the retrieved paper metadata before it is provided to the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 3, 2026, 08:47 PM