Regulatory & Guidelines

The Regulatory & Guidelines MCP is an evidence query tool for regulatory policies and clinical norms, covering semantic evidence retrieval for clinical guidelines and FDA drug labels. It helps the AI assistant retrieve evidence snippets related to treatment guidelines, indications, dosage, safety, contraindications, warnings, and recommendations using natural-language queries. Tools included: • clinical_guideline_search: Semantically search clinical guideline evidence. • fda_label_search: Semantically search FDA drug label evidence.

• When you need evidence from treatment guidelines, recommended regimens, or clinical pathways. • When you need FDA label evidence on indications, dosage, safety, contraindications, warnings, or drug interactions. • When the question is a natural-language query, keyword combination, or requires semantic matching. • When you need evidence citations for medical, regulatory, or labeling-related answers. • Not suitable for structured drug asset or clinical trial record search.

• Provide a natural-language query, ideally including disease, drug, treatment setting, patient population, or evidence type. • Provide lang, using CN or EN. • Optionally provide top_k to control the number of returned evidence snippets. • Avoid overly generic meta-terms such as “document”, “literature”, or “report” unless they are essential to the question.

• clinical_guideline_search returns matched guideline evidence snippets, usually including text, title, source, URL, document type, subtype, and reference. • fda_label_search returns matched FDA label evidence snippets, usually including label text, source URL, document type, subtype, and reference. • Outputs are evidence snippets and reference signals, suitable for summarization, citation, and cross-checking.

• This is a semantic retrieval tool and returns evidence snippets; it may not cover every regulatory document. • lang must be CN or EN. • FDA label content is US regulatory context; other jurisdictions may not be covered. • Medical conclusions should be verified against original text, patient population, guideline version, and regulatory region. • For high-risk medical decisions, do not rely on a single snippet.