tooluniverse-metabolomics

Pass

Audited by Gen Agent Trust Hub on Jul 30, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill depends on the tooluniverse package and retrieves data from several authoritative scientific repositories, including the Human Metabolome Database (HMDB), MetaboLights (EBI), Metabolomics Workbench (NIH), and PubChem. These are recognized services within the bioinformatics research community and are used for their intended purposes.- [PROMPT_INJECTION]: The skill handles untrusted data retrieved from external metabolomics databases which is then interpolated into generated markdown reports. This creates a surface for indirect prompt injection if the external records contain malicious instructions designed to influence downstream agent actions.
  • Ingestion points: External data is ingested through the metabolite_list, study_id, and search_query parameters in python_implementation.py and python_implementation_fixed.py.
  • Boundary markers: The report generation logic does not utilize explicit delimiters or 'ignore' instructions to separate external content from the agent's own instructions.
  • Capability inventory: The skill is capable of writing markdown files to the local filesystem and executing a wide range of external data retrieval tools via the ToolUniverse SDK.
  • Sanitization: Data retrieved from external API responses is appended directly to the report string without sanitization, escaping, or schema validation.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 30, 2026, 03:10 AM