arxiv-search

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • PROMPT_INJECTION (LOW): The skill processes untrusted external data which could contain malicious instructions designed to influence the agent.\n
  • Ingestion points: In arxiv_search.py, the query_arxiv function retrieves paper titles and summaries from the arXiv API.\n
  • Boundary markers: Absent. The script formats output using simple text headers (Title/Summary) without delimiters or instructions to ignore embedded commands.\n
  • Capability inventory: The script executes network requests to the arXiv API and performs string formatting; it does not contain direct file-write or shell-execution capabilities.\n
  • Sanitization: Absent. Paper abstracts are returned directly to the agent without any sanitization or escaping.\n- EXTERNAL_DOWNLOADS (LOW): The skill depends on the arxiv Python package which must be installed from PyPI.\n
  • Evidence: Both SKILL.md and arxiv_search.py identify the arxiv library as a required dependency.\n
  • Context: While the library is a standard tool for this purpose, it is a third-party dependency not maintained by the agent's core developers.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 05:01 PM