rdkit-cheminformatics

Pass

Audited by Gen Agent Trust Hub on Jul 10, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill implements domain-specific molecular processing using the RDKit library. All scripts focus on chemical informatics tasks such as descriptor calculation and fingerprint similarity.
  • [COMMAND_EXECUTION]: Instructions include standard dependency management commands using pip and conda to install reputable scientific libraries from official registries.
  • [DATA_EXFILTRATION]: File operations are restricted to the local environment, specifically creating a results/ directory and saving CSV, SDF, and image files. No external network requests or data transmission patterns were found.
  • [PROMPT_INJECTION]: The skill ingests external molecular data (SMILES and SDF). Analysis shows that ingestion points in SKILL.md are processed via chemical-specific validation functions like Chem.MolFromSmiles, which sanitizes inputs by verifying chemical validity. The capabilities are limited to local file writes, and no boundary markers are required as the data is not re-interpolated into natural language prompts.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 10, 2026, 03:35 PM
Security Audit — agent-trust-hub — rdkit-cheminformatics