threat-index-mcp

An MCP (Model Context Protocol) server for the AI Career Threat Index — so your AI assistant can answer "how exposed is my job to AI?" with real, versioned data instead of vibes.

Zero dependencies. Node ≥ 18. Reads the dataset from the repo when run from a clone, otherwise fetches it from GitHub raw — the single server.js file works anywhere.

Add it to Claude Code

git clone https://github.com/Jott2121/ai-career-threat-index.git
claude mcp add threat-index -- node ./ai-career-threat-index/mcp/server.js

Add it to Claude Desktop

{
  "mcpServers": {
    "threat-index": {
      "command": "node",
      "args": ["/path/to/ai-career-threat-index/mcp/server.js"]
    }
  }
}

Tools

| Tool | What it answers | |---|---| | lookup_role | Full record for one role: score, band, four sub-scores with rationales, tasks, defense skills, salary, history | | list_roles | Roles filtered by category / risk band | | compare_roles | Side-by-side comparison of 2–6 roles | | top_movers | Biggest quarter-over-quarter score changes | | safest_roles / riskiest_roles | Extremes, overall or per category | | crosswalk_soc | Any BLS SOC 2018 code → nearest scored role |

Try: "Compare paralegal, bookkeeper and welder — which is most exposed to autonomous agents, and what should each build to defend?"

License

MIT, same as the dataset. Scores are structured editorial estimates against a published rubric, not measurements.