Воспроизводимый поиск в научных базах данных через API

★ 7.0 · research

database-lookup is a Claude Code skill that performs reproducible queries against 78 documented public database APIs, returning results with explicit endpoints, filters, pagination, and full provenance. Coverage spans biomedical and genomics databases, physics and astronomy, chemistry and drugs, materials science, patents, regulatory registries, economics, and social sciences — each database has its own reference file with endpoints, query formats, and worked examples. Core operations include authoritative database selection, identifier resolution across formats such as UniProt, Ensembl, PubChem, and ChEMBL, bounded and rate-limited API calls, and count reconciliation to verify retrieval completeness. Every result includes the databases queried, parameters used, access date, and warnings about incomplete pagination or ambiguous filters, so any agent or human can repeat the lookup exactly. It is designed for situations where a scientific, regulatory, or financial fact must come from a named source rather than be inferred from the model's general knowledge.