MCP server giving Claude Desktop structured access to cached labor market data — occupation outlook, skill demand, compensation benchmarks, and adjacent roles. Demonstrates stateful tool design with research session persistence.
```bash git clone https://github.com/Schoobert/mcp-labor-market cd mcp-labor-market python -m venv .venv && source .venv/bin/activate pip install -r requirements.txt python src/database.py # builds the SQLite database from raw data files — run once ``` Then wire it into Claude Desktop (see [How to try it](#how-to-try-it) below).