Federal Compass MCP server

federal-compass-mcp

Your AI compass from private sector to federal career

Federal job descriptions are 5 pages of legal language. GS grades, clearance levels, hiring paths — none of this is explained anywhere. federal-compass-mcp connects Claude to the official USAJobs API and turns it into an AI advisor that speaks plain English.

Built for tech professionals transitioning from private sector to federal/government careers.

What it does

• Searches federal job openings with optional filters — keyword, location, salary, clearance level
• Translates federal job descriptions into plain language — what they actually want, without the jargon
• Analyzes your CV against specific job postings and tells you if you qualify
• Explains federal concepts — GS grades, pay plans, security clearances, hiring paths
• Saves your CV locally so Claude can reference it automatically

Getting started

1. Get your free USAJobs API key

Register at developer.usajobs.gov — takes about 3 minutes. Verify your email and you'll receive an API token.

2. Add to your MCP client

Claude Desktop (~/Library/Application Support/Claude/claudedesktopconfig.json):

{
  "mcpServers": {
    "federal-compass": {
      "command": "npx",
      "args": ["-y", "federal-compass-mcp@latest"],
      "env": {
        "USAJOBS_API_KEY": "your_api_key",
        "USAJOBS_EMAIL": "your@email.com"
      }
    }
  }
}

VS Code / Cursor / Windsurf — same config, different config file path. See your client's MCP documentation.

3. Try it

Find me senior software developer jobs in Raleigh, NC

What is a GS-13 and how much does it pay?

Save my CV [attach file]

Look at this job posting — am I qualified?

Tools

CV storage

Your CV is stored locally on your machine — never sent to any server:

~/.federal-compass/
  cv.json
  config.json
  codelists/     <- cached USAJobs reference data

Why federal jobs are confusing (and how this helps)

Coming from private sector, federal job postings feel like a different language:

Shortened. The full README is on GitHub.