medchem

Pass

Audited by Gen Agent Trust Hub on Mar 31, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill implements molecular filtering workflows using established cheminformatics libraries. Analysis of the scripts and documentation shows no evidence of malicious intent or security vulnerabilities.
  • [EXTERNAL_DOWNLOADS]: The skill recommends installing the medchem package via uv pip. This is a standard, well-known scientific package on PyPI from a reputable source (datamol.io), posing no security risk under current evaluation criteria.
  • [COMMAND_EXECUTION]: The provided script scripts/filter_molecules.py processes molecular data files locally. It uses structured parsing via pandas and rdkit, and does not invoke subshells or execute untrusted strings as code.
  • [DATA_EXFILTRATION]: No network operations or external data transmissions were found in the skill's code. The script operates entirely on local input/output paths provided by the user.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 31, 2026, 08:21 AM
Security Audit — agent-trust-hub — medchem