ra_evals_run: Run an agent's eval suite (dry runs)

Run an agent's eval suite as dry runs on its draft or live version and score them.

• Changes data: yes
• Scope needed: agents:write

Example call

{
  "name": "ra_evals_run",
  "arguments": {
    "id": "agt_01HZXEXAMPLE",
    "body": {
      "target": "draft"
    }
  }
}

All tools: MCP tools. Connecting: quickstarts.