bioservices

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • [Prompt Injection] (SAFE): No attempts to override agent behavior or bypass safety filters were found.\n- [Data Exposure & Exfiltration] (SAFE): No sensitive local data access or credential harvesting. Network activity is limited to bioinformatics APIs as described.\n- [Unverifiable Dependencies] (SAFE): The skill depends on 'bioservices', a widely used and reputable bioinformatics package. It is installed via standard package managers.\n- [Indirect Prompt Injection] (SAFE): The skill processes data from external bioinformatics APIs. While this represents an ingestion surface for external content, the data is highly structured and scientific, presenting negligible risk.\n
  • Ingestion points: scripts/pathway_analysis.py via kegg.get and kegg.parse_kgml_pathway.\n
  • Boundary markers: None explicitly defined in scripts.\n
  • Capability inventory: Network requests to bioinformatics services; file system write access for saving analysis results.\n
  • Sanitization: Handled by the 'bioservices' library's internal parsing logic.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 04:59 PM