medchem

Pass

Audited by Gen Agent Trust Hub on Sep 25, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses a data ingestion surface through the processing of external molecular files (CSV, SDF, TXT).
  • Ingestion points: Data enters via the load_molecules function in scripts/filter_molecules.py.
  • Boundary markers: Absent; there are no specific prompt delimiters, although the input is restricted to molecular formats.
  • Capability inventory: The skill has file read/write access via scripts/filter_molecules.py for input triage and outputting results.
  • Sanitization: Input SMILES and SDF data are sanitized and validated using the datamol and rdkit chemical informatics libraries.
  • [EXTERNAL_DOWNLOADS]: The skill requires standard scientific libraries medchem, datamol, and lilly-medchem-rules, which are installed from trusted registries like PyPI and conda-forge as documented in the installation instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 25, 2026, 02:34 PM