Sec-Password-Reset-Token-Logging

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 verdict, risk, mitigation. Password reset tokens are written to application logs for debugging after each reset request.

{
  "metrics": {
    "correct": {
      "checks": [
        {
          "contains": [
            "password reset token"
          ]
        },
        {
          "contains": [
            "logs"
          ]
        },
        {
          "contains": [
            "secret"
          ]
        }
      ]
    },
    "hidden": {
      "checks": [
        {
          "contains": [
            "redact"
          ]
        },
        {
          "contains": [
            "short-lived"
          ]
        },
        {
          "contains": [
            "audit"
          ]
        }
      ]
    }
  }
}