@artale/pi-eval

Agent evaluation harness. Judge sessions on success, tool usage, efficiency, methodology. Inspired by opencc.

Package details

extension

Install @artale/pi-eval from npm and Pi will load the resources declared by the package manifest.

$ pi install npm:@artale/pi-eval

Pi manifest JSON

{
  "commands": [
    "eval"
  ],
  "tools": [
    "eval_judge",
    "eval_handoff"
  ]
}

Pi packages can execute code and influence agent behavior. Review the source before installing third-party packages.

Agent evaluation harness. Judge coding sessions on methodology, efficiency, and success.

npm install -g @artale/pi-eval