{
  "schema": "gooo/released-cli-dogfood-observation/v1",
  "date": "2026-10-08",
  "release": "v0.6.7-dev",
  "compiler_source_sha": "3579b869b6025e34db87d9a283b6aabf54c39eb3",
  "release_run_id": 37719107926,
  "readiness_run_id": 37718738019,
  "local_platform": "darwin/arm64",
  "downloaded_assets_verified": 9,
  "checksum_entries_verified": 8,
  "runtime_observations": [
    {
      "observation": "dependent-first.json",
      "passed": 0,
      "total": 2,
      "runtime_model_calls": 0
    },
    {
      "observation": "dependent-next.json",
      "passed": 2,
      "total": 2,
      "runtime_model_calls": 0
    },
    {
      "observation": "dependent-replay.json",
      "passed": 2,
      "total": 2,
      "runtime_model_calls": 0
    },
    {
      "observation": "package-first.json",
      "passed": 2,
      "total": 4,
      "runtime_model_calls": 0
    },
    {
      "observation": "package-next.json",
      "passed": 4,
      "total": 4,
      "runtime_model_calls": 0
    },
    {
      "observation": "package-replay.json",
      "passed": 4,
      "total": 4,
      "runtime_model_calls": 0
    }
  ],
  "continuation_new_model_calls": {
    "dependent": 0,
    "package": 0
  },
  "initial_package_model": {
    "calls": 1,
    "predict_ns": 27208,
    "metadata": {
      "schema": "gooo/retained-path-model/v1",
      "loaded": true,
      "metadata_sha256": "f95925a277c45957722bc116213c61bee997cc07a94adae738a05d218d660059",
      "weights_sha256": "049081cc6dc62f3c4230f72e601c1fd421c514fce9cfdf7bc1eb412ca4c5514b",
      "resident_tensor_bytes": 2096,
      "setup_ms": 0.497042,
      "scope": "one constructor load; excluded from request timing; fresh source and plan each request",
      "model_schema": "gooo/tiny-shared-three-choice-path-model/v1",
      "feature_version": "triple_record_field_context_v1_shared_v1",
      "arithmetic_version": "float32_separate_v1"
    }
  },
  "numeric_refinement": {
    "feedback_counts": [
      [
        1,
        10
      ],
      [
        10,
        10
      ]
    ],
    "separate_evaluation": [
      3,
      3
    ],
    "retained_candidates": 16,
    "omitted_candidates": 92,
    "model_calls": 0
  },
  "limits": [
    "Finite named examples; general accuracy and model training independence remain unmeasured.",
    "Single model prediction observation; host CPU utilization and model-only CPU increment were not sampled.",
    "Release report FOUNDATION prose still names an older v1 denominator; current structured report and native receipts measure four targets and 26 release cases."
  ]
}