MCP Server with MedixHub medical and healthcare API integration, providing access to OpenTargets, FDA Drug Label, and Monarch Initiative APIs
Server has 7 tools with basic schemas and descriptions, but significant quality gaps. Tool names lack clear action verbs (queryOpenTargets, queryOpenTargetsGenetics, queryDrugByName, queryFDADrugLabel, getDrugGenericName, queryMonarch, findDiseasesForMultiplePhenotypes). Descriptions are present but generic (10-80 chars, below 194-char baseline). Parameter descriptions are minimal or absent. No documented output schemas. No error handling guidance. No tool annotations (readOnlyHint, destructiveHint). Parameter types present but lack constraints (enums, ranges, patterns). Tool 3 (queryDrugByName) has a 'getDrugGenericName' parameter typed as 'function', invalid for MCP tool parameters. No pagination support visible. Responses not shaped for agent chaining.
Find the intersection of diseases associated with multiple HPO IDs (phenotypes)
Get generic name for a drug based on brand or generic name
Main function to query OpenTargets by drug name with generic name fallback
Query FDA Drug Label API for drug information
Main function to query the Monarch Initiative API
Main function to query the OpenTargets Platform API
Main function to query the OpenTargets Genetics API
Tool names lack action verbs. 'queryOpenTargets', 'queryMonarch' use 'query' (vague); should be 'search_opentargets', 'search_monarch_diseases'. LLMs cannot infer intent from generic verbs.
Tool 3 (queryDrugByName) declares parameter 'getDrugGenericName' with type 'function'. MCP tools cannot accept function parameters, this is invalid. Must be removed or refactored as a separate tool call.
No output schemas documented. LLMs cannot plan downstream tool calls or extract required fields (e.g., drug IDs, disease IDs for chaining). Responses are opaque.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 31 | <=2025-11-25 | v2 |
Parameter descriptions are missing or trivial. 'query' param in queryOpenTargets has description 'The GraphQL query to execute' but no guidance on format, required fields, or examples. 'variables' has no description at all.
No error handling guidance. Tools return raw API responses without recovery hints. If a query fails, LLM has no guidance on retry strategy, alternative parameters, or fallback tools.