This handles the messy reality of looking up chemical compounds where "aspirin" maps cleanly but "Vitamin D" splits into multiple forms. It cross-references PubChem and ChEMBL to resolve names to actual identifiers (CIDs, ChEMBL IDs, SMILES), fetches molecular properties and bioactivity data, and grades evidence quality based on whether databases agree. The workflow is explicit about disambiguation first, then retrieval, which matters when stereochemistry or salt forms create ambiguity. Useful when you need reliable compound data without manually checking if two databases are talking about the same molecule. It won't guess at properties and includes SMILES verification to catch malformed inputs early.
npx -y skills add mims-harvard/tooluniverse --skill tooluniverse-chemical-compound-retrieval --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
juliusbrussee/caveman
mattpocock/skills
shadcn/improve
obra/superpowers
forrestchang/andrej-karpathy-skills
vercel-labs/skills