NCBI Entrez MCP Server

{ "name": "pubmed-literature-review", "description": "Workflow template for a PubMed-focused evidence scan using the Entrez MCP server.", "instructions": [ "Call `entrez_capabilities` (alias `entrez-capabilities`) with { \"tool\": \"entrez_query\", \"format\": \"detailed\" } if you need a refresher on parameter hints before starting.", "Use `entrez_query` with `operation: \"search\"` and a Boolean-rich `term`. Start with `retmax` <= 25 and `intended_use: \"search\"`.", "Summarise your primary PMIDs using `operation: \"summary\"` and `detail_level: \"brief\"` to stay token efficient.", "Pull abstracts for shortlisted PMIDs via `operation: \"fetch\"` with `rettype: \"abstract\"` when deeper reading is required.", "Persist key records through `entrez_data` `fetch_and_stage` when you anticipate SQL analysis or repeated use.", "Finish by synthesising findings and suggesting next investigative steps." ], "retry_guidance": [ "Prefer uppercase Boolean operators (AND/OR/NOT) inside the `term`.", "Restrict by publication year using `[Date]` filters before widening the scope.", "If responses get long, add `compact_mode: true` or set `max_tokens` to 300." ], "primary_tools": ["entrez_query", "entrez_data", "entrez_capabilities"] }