rdkit-chemdraw-cdxml

Pass

Audited by Gen Agent Trust Hub on Aug 6, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill recommends installing the epam.indigo package for rendering ChemDraw files into PNGs. EPAM Systems is a well-known global provider of life sciences and engineering software, making this a trusted source for scientific dependencies.\n- [PROMPT_INJECTION]: The skill ingests and processes chemical data from SMILES strings and CDXML files, which creates a surface for indirect prompt injection if malicious data is provided.\n
  • Ingestion points: Untrusted data enters the skill's workflow through file reading operations in check_scheme.py and build_reaction_scheme.py, which are then passed to rdkit.Chem.rdChemDraw and xml.etree.ElementTree.fromstring.\n
  • Boundary markers: The skill does not implement specific boundary markers or delimiters to isolate processed chemical data from the agent's instructional context.\n
  • Capability inventory: The skill has permissions to write files to the local file system and render images through the Indigo library.\n
  • Sanitization: The skill leverages RDKit's built-in molecule sanitization (Chem.SanitizeMol) to validate chemical structures and uses standard XML parsing libraries to ensure document structure.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 6, 2026, 02:55 PM
Security Audit — agent-trust-hub — rdkit-chemdraw-cdxml