evaluation/audit.json

{
  "passed": true,
  "host": "eai-Raider-16-HX-AI-A2XWIG",
  "completed_at": "2026-09-20T14:53:24.163149",
  "raw_rollouts": 60,
  "instances_per_variant": 20,
  "instance_ids": [
    301,
    302,
    303,
    304,
    305,
    306,
    307,
    308,
    309,
    310,
    311,
    312,
    313,
    314,
    315,
    316,
    317,
    318,
    319,
    320
  ],
  "checks": [
    "complete local campaign",
    "matching model readiness records",
    "all expected unique official raw results",
    "valid task, rollout, step counts and finite scores",
    "aggregate and per-instance JSON agree with raw results",
    "simulator source hashes unchanged"
  ]
}