biopython

Pass

Audited by Gen Agent Trust Hub on Apr 16, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill facilitates data retrieval from well-known scientific databases.
  • Evidence: Utilizes Bio.Entrez and Bio.Blast.NCBIWWW to query NCBI's databases and Bio.PDB.PDBList to fetch macromolecular structures from the Protein Data Bank (PDB).
  • [COMMAND_EXECUTION]: The skill includes instructions for running local bioinformatics binaries through subprocess wrappers.
  • Evidence: Reference documentation includes examples for ClustalOmegaCommandline, MuscleCommandline, and DSSP, which execute their respective binaries on the host system.
  • [PROMPT_INJECTION]: The skill processes untrusted biological data from external files and API responses, creating a surface for indirect prompt injection.
  • Ingestion points: Reads biological sequence and structural data from files (FASTA, GenBank, PDB, etc.) and NCBI API responses via modules like Bio.SeqIO and Bio.Entrez.
  • Boundary markers: No explicit delimiters or instructions to ignore potential embedded commands within data files are provided in the documentation.
  • Capability inventory: The skill possesses network access capabilities and the ability to execute local command-line tools as identified in the download and command execution findings.
  • Sanitization: The skill relies on standard Biopython library parsers without additional sanitization or content validation of the scientific data being processed.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 16, 2026, 10:49 PM
Security Audit — agent-trust-hub — biopython