evalgate

Statistical checks an agent runs before trusting an AI eval number (is #1 real, judge bias, more).

Install

Add this to your `claude_desktop_config.json` file:

{
  "mcpServers": {
    "evalgate": {
      "command": "uvx",
      "args": [
        "eval-integrity"
      ]
    }
  }
}

Runnable packages

pypieval-integrity0.4.1stdio

What it can do

No publishable tool enumeration has been recorded.

Community

Evidence

This server has Registry metadata but no public verification run yet.