MCP Server for court decisions, case law, and legal research — CourtListener (3M+ cases) and EUR-Lex
Server has 6 tools with basic naming (verb_noun pattern) and descriptions in German. However, critical gaps undermine quality: (1) Parameter descriptions lack detail on constraints, formats, and valid ranges. E.g., 'court' param in search_cases lists examples but no enum or validation rules. (2) Output schemas are NOT documented, no structured response definitions visible in tool registration. (3) Error handling is absent, no recovery guidance, retryability classification, or actionable error messages. (4) No tool annotations (readOnlyHint, etc.) despite all tools being read-only. (5) Descriptions are in German, which may confuse English-speaking LLMs. Average tool score: 52/100.
Holt vollständige Details eines Falls von CourtListener.
Holt Informationen über ein bestimmtes Gericht.
Findet einen Fall anhand seiner juristischen Zitierung.
Durchsucht 3M+ US-Gerichtsentscheidungen nach Stichwörtern.
Durchsucht EU-Recht und Rechtsprechung über EUR-Lex.
Sucht Richter nach Name und zeigt deren Informationen.
Output schemas not documented. No structured response definitions visible in tool registration. LLMs cannot plan downstream calls or extract required fields (e.g., case_id, court_id for chaining).
Parameter descriptions lack constraint details. 'court' param lists examples ('scotus', 'ca1') but no enum or validation rules. 'limit' has default/max but no minimum. LLMs cannot validate inputs before calling.
No error handling or recovery guidance. If CourtListener API fails or returns 404, no actionable error message tells the LLM what to do next (retry, search differently, etc.).
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 50 | 2026-07-28+ | v2 |
Tool descriptions in German may confuse English-speaking LLMs. Descriptions should be in the same language as the MCP server instructions (English preferred for broad compatibility).
No tool annotations despite all tools being read-only. Missing readOnlyHint annotation prevents LLMs from understanding safety properties and optimizing retry logic.