Refactor-Command-Router-Modules

The model receives the prompt (and optional system message). The run uses scorer rubric_json_metrics with the JSON configuration below. Pass/fail and partial credit are determined entirely by that scorer against the model output; no human grading.

Return JSON only with keys plan, refactor, tests. A CLI command router imports every command, parses arguments, validates config, and executes side effects in one file. Split responsibilities while preserving command names.

{
  "metrics": {
    "intent": {
      "checks": [
        {
          "contains": [
            "command router"
          ]
        },
        {
          "contains": [
            "split responsibilities"
          ]
        },
        {
          "contains": [
            "modules"
          ]
        }
      ]
    },
    "visible": {
      "checks": [
        {
          "contains": [
            "imports every command"
          ]
        },
        {
          "contains": [
            "parses arguments"
          ]
        },
        {
          "contains": [
            "side effects"
          ]
        }
      ]
    },
    "hidden": {
      "checks": [
        {
          "contains": [
            "command names"
          ]
        },
        {
          "contains": [
            "golden tests"
          ]
        },
        {
          "contains": [
            "same exit codes"
          ]
        }
      ]
    }
  }
}