tooluniverse-systems-biology

Pass

Audited by Gen Agent Trust Hub on Mar 8, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits a vulnerability to indirect prompt injection by interpolating user-provided biological identifiers and keywords directly into a Markdown report.
  • Ingestion points: The gene_list, protein_id, and pathway_keyword arguments in the systems_biology_pipeline function within python_implementation.py ingest untrusted data.
  • Boundary markers: Absent; user inputs are directly concatenated into the report structure without delimiters or instructions to ignore potential commands within the data.
  • Capability inventory: The skill has the capability to write files to the local file system and perform network queries to external biological databases (KEGG, Reactome, etc.) via the ToolUniverse framework.
  • Sanitization: Absent; no validation or escaping of the biological symbols or keywords is performed prior to their inclusion in the generated report file.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 8, 2026, 12:01 PM