install-legal-data-hunter
GitHub用于安装和配置 Legal Data Hunter MCP 服务器，连接 AI 代理以访问全球 1800 万+法律文档。支持在 Claude Code 及多种 IDE 中自动部署，提供跨法域的混合语义搜索、案例解析及全文检索功能。
Trigger Scenarios
Install
npx skills add lawve-ai/awesome-legal-skills --skill install-legal-data-hunter -g -y
SKILL.md
Frontmatter
{
"name": "install-legal-data-hunter",
"description": "Install and set up the Legal Data Hunter MCP server. Use when the user wants to connect their AI agent to legal data, install legal-data-hunter, add a legal search MCP, or search legal documents across jurisdictions.",
"allowed-tools": "Bash Read Write",
"user-invocable": true
}
Install Legal Data Hunter
Install the Legal Data Hunter MCP server and verify the connection. Legal Data Hunter provides access to 18M+ legal documents across 110+ countries — case law, legislation, and doctrine — via hybrid semantic + keyword search.
Step 1: Detect environment and install
Check if running inside Claude Code:
which claude 2\>/dev/null
If Claude Code is available, run:
claude mcp add legal-data-hunter --transport http https://legaldatahunter.com/mcp
If not Claude Code (Cursor, VS Code, Windsurf, JetBrains, Copilot, Lawvable), run:
npx legal-data-hunter --yes
This auto-detects the client and writes the MCP config to the correct location. It supports: Cursor, VS Code, Lawvable, Windsurf, Copilot CLI, and JetBrains (Junie).
Step 2: Verify installation
After installing, confirm the MCP server is connected. In Claude Code, reconnect MCP servers and check that Legal Data Hunter appears with its tools. The server exposes 7 tools:
| Tool | Purpose |
|---|---|
discover\_countries() |
List all 110+ indexed countries with document counts |
discover\_sources(country) |
List available sources for a country (e.g. "FR", "DE", "EU") |
search(query, ...) |
Hybrid semantic + keyword search across jurisdictions |
resolve\_reference(citation) |
Look up a specific citation — ECLI, CELEX, article number, case number |
get\_document(id) |
Retrieve the full text of a document by ID |
get\_filters() |
List available filter options (courts, document types, date ranges) |
report\_source\_issue(...) |
Report a problem with a data source or request a new one |
Step 3: Getting started
Once connected, here are useful first queries to try:
Explore coverage:
- Call
discover_countries()to see all indexed jurisdictions and their document counts - Call
discover_sources("FR")to see French court databases, legislation sources, and doctrine
Search across jurisdictions:
search("unfair dismissal", country="DE", doc_type="case_law")— German case law on unfair dismissalsearch("artificial intelligence regulation", country="EU")— EU legislation on AIsearch("data protection breach notification", country="UK")— UK GDPR enforcement
Resolve specific citations:
resolve_reference("ECLI:EU:C:2014:317")— look up a CJEU decision by ECLIresolve_reference("32016R0679")— look up the GDPR by CELEX numberresolve_reference("Article 49 TFEU")— find Treaty provisions
Retrieve full documents:
- Use
get_document(id)with an ID returned from search or resolve_reference to get the complete text
Tips
- Natural language works: The search engine understands legal concepts, not just keywords. "Can an employer fire someone for social media posts?" works as well as "dismissal social media misconduct".
- Cross-reference: Search one jurisdiction, then use the same query in another to compare legal approaches across borders.
- Citation chains: Use
resolve_reference()to find a law, thensearch()with that law's name to find all court decisions that cite it. - Filters narrow results: Call
get_filters()first to see what filter options are available, then pass them tosearch(). - Missing a source? Use
report_source_issue()to request a new legal source or report a problem. New sources can be indexed in under 48 hours.
Links
- Dashboard: https://legaldatahunter.com
- API Docs: https://legaldatahunter.com/docs
- GitHub: https://github.com/worldwidelaw/legal-sources
- MCP endpoint: https://legaldatahunter.com/mcp
Version History
- 7f58aaf Current 2026-07-05 11:51
