EBI PRIDE Archive MCP — proteomics (mass-spectrometry) data repository. Keyless. Wraps the public PRIDE Archive REST API (v3).
Two tools with clear verb-noun naming (search_projects, get_project) and solid descriptions (194 - 250 chars, within baseline 34 - 392). Input schemas present with proper JSON Schema structure and type definitions. However, output schemas are NOT documented, LLMs cannot predict response structure. Parameter descriptions are adequate but lack format constraints (e.g., accession pattern, limit bounds). Error handling returns generic messages without recovery guidance. No tool annotations (readOnlyHint, idempotentHint). Missing pagination details in search_projects despite returning arrays.
Get full metadata for one PRIDE Archive proteomics project by accession (e.g. "PXD000001"): title, description, sample-processing and data-processing protocols, organisms, organism parts, diseases, instruments, keywords, and DOI.
Search EBI PRIDE Archive — the largest public proteomics/mass-spectrometry dataset repository — for projects by keyword (disease, organism, technique, protein, instrument). Returns matching project accessions with title, organisms, diseases, and instruments. Keyless.
Output schemas not documented. LLMs cannot predict response structure (fields, types, nesting). search_projects returns {count, projects[]} and get_project returns {accession, title, description, ...} but this is inferred from code, not declared in tool metadata.
Parameter descriptions lack format constraints. 'accession' should specify pattern (e.g., 'PXD followed by 6 digits'). 'limit' lacks explicit bounds in description (code clamps 1 - 100 but description does not state this). 'keyword' lacks guidance on what searches (disease, organism, technique, protein, instrument).
Error responses lack recovery guidance. callTool() returns {error: string} but does not tell LLM what to do next. E.g., 'project not found' should suggest 'Try search_projects() to find valid accessions' or list available alternatives.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 59 | 2026-07-28+ | v2 |
No tool annotations. Tools are read-only (safe to retry) but lack readOnlyHint annotation. This forces LLMs to reason about idempotency instead of relying on metadata.
Pagination not fully documented. search_projects accepts page/limit and returns count, but description does not state 'Returns paginated results; use page and limit to iterate.' This forces LLMs to infer pagination behavior.