This is the go-to cheminformatics library when you need fine-grained control over molecular operations. You get SMILES and SDF parsing, descriptor calculations like molecular weight and LogP, fingerprinting for similarity searches, substructure matching, and 2D/3D coordinate generation. It auto-sanitizes molecules on import with valence checking and aromaticity perception, which is great until you need to override it for edge cases. The skill documentation covers batch processing with suppliers, Morgan and MACCS fingerprints, Tanimoto similarity, and Butina clustering. If you just need standard workflows, use datamol instead since it wraps this with a simpler interface. Pick rdkit when you're doing custom sanitization or need specialized algorithms that higher-level tools don't expose.
npx skills add https://github.com/davila7/claude-code-templates --skill rdkit