io.github.ipezygj/evalgate

not probeable (local package)

Unclaimed · Unverified

Publisher-provided description

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

Summary by lookupd

Runs statistical checks on AI evaluation results, such as whether a top ranking is significant and whether a judge model is biased.

Install

Claude Code, pypi package (stdio)

claude mcp add evalgate -- uvx eval-integrity

Codex, pypi package (stdio)

codex mcp add evalgate -- uvx eval-integrity

Cursor, pypi package (stdio) → .cursor/mcp.json

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

Other clients: docs. Or tell your agent: Set up lookupd for me: read https://lookupd.com/agent.md and follow it.

Shows this live status and links here; updates hourly.

[![lookupd status](https://lookupd.com/badge/io.github.ipezygj/evalgate.svg)](https://lookupd.com/servers/io.github.ipezygj/evalgate)