MCP Server for global museum collections: Met Museum, Art Institute of Chicago, Smithsonian — search 700k+ artworks, no API key required
Museum MCP server has 9 tools with complete input schemas and descriptions in German. All tools are READ_ONLY and follow verb_noun naming (tool_search_*, tool_get_*). However, descriptions lack LLM-optimization guidance (e.g., when to use search_artworks vs search_by_artist vs search_by_period), parameter descriptions are minimal, and output schemas are not formally documented. No error handling guidance, no recovery patterns, and no parameter validation rules stated. Tool composition has overlapping functionality (3 search variants) without clear disambiguation.
Vergleiche Kunstwerke zwischen Met Museum und Art Institute of Chicago.
Vollstaendige Details zu einem Kunstwerk abrufen.
Alle Abteilungen des Metropolitan Museum of Art abrufen.
Kuratierte Highlights aus einem Museum abrufen.
Zufaelliges Kunstwerk aus einem Museum abrufen (fuer Entdeckungen).
Kunstwerke in Museumssammlungen suchen.
Kunstwerke eines bestimmten Kuenstlers suchen.
Overlapping search tools lack disambiguation. tool_search_artworks, tool_search_by_artist, tool_search_by_period, and tool_search_by_department all perform similar queries. Descriptions do not explain when to use each variant, forcing LLMs to guess or try multiple tools.
Output schemas not documented. Code shows responses return structured dicts (artworks array with id, title, artist, date, etc.), but no formal schema definition is visible. LLMs cannot plan downstream calls without knowing response structure.
No error handling or recovery guidance. Code catches exceptions silently (try/except continue in tool_search_artworks) and returns empty results. LLMs receive no actionable error messages or retry hints.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 48 | 2026-07-28+ | v2 |
Kunstwerke in einer bestimmten Met-Abteilung suchen.
Kunstwerke aus einer bestimmten historischen Periode suchen.
Parameter descriptions are minimal. 'query' is described as 'Suchbegriff (Kuenstler, Thema, Periode, Stil, etc.)', examples in descriptions risk LLM literal reuse. 'limit' lacks min/max constraints (code enforces max:20 but description does not state this).
Tool descriptions lack context for selection. 'Kunstwerke in Museumssammlungen suchen' (Search artworks in museum collections) does not explain when to call this vs search_by_artist or search_by_period. LLMs need explicit guidance on tool selection criteria.