bioservices
Pass
Audited by Gen Agent Trust Hub on Nov 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from numerous external bioinformatics web services (UniProt, KEGG, ChEMBL, etc.). This data ingestion is central to the skill's purpose but constitutes a surface for indirect prompt injection if malicious content were embedded in scientific database records.
- Ingestion points: API responses processed in
scripts/protein_analysis_workflow.py,scripts/compound_cross_reference.py, andscripts/pathway_analysis.py. - Boundary markers: Content from external APIs is typically parsed and formatted before presentation to the agent, providing natural structural boundaries.
- Capability inventory: The skill allows writing analysis results to the local file system and performing further network queries based on retrieved identifiers.
- Sanitization: The skill relies on structured parsing (XML, JSON, TSV) by the
bioserviceslibrary to extract specific biological data fields.
Audit Metadata