medchem
Pass
Audited by Gen Agent Trust Hub on Apr 16, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns, prompt injections, or unauthorized data access mechanisms were identified in the skill content or instructions.- [EXTERNAL_DOWNLOADS]: The skill installs the 'medchem' package from a standard package registry. It also refers to official documentation and repositories under the datamol.io domain, which is a recognized provider of open-source chemoinformatics tools.- [COMMAND_EXECUTION]: The skill includes a local utility script 'scripts/filter_molecules.py' for batch processing molecular data, which is a standard pattern for science-oriented skills.- [PROMPT_INJECTION]: While the skill ingests external molecular data (e.g., SMILES strings from CSV files), this ingestion surface is restricted to scientific analysis using specialized parsers with no path to natural language instruction execution.
Audit Metadata