KB-4E7C rev 5

FIX7 Authority Closure — N7 Approval-Event Input Envelope (rev5, N6 availability updated)

10 min read Revision 5

FIX7 Authority Closure — N7 Approval-Event Input Envelope (rev5)

  • Date: 2026-06-10 · Patched: 2026-06-11 · rev3 patches the Codex CODEX_FIX7_FINAL_AUTHORITY_SEAL_REJECT (per-field value grammar added; provenance class system added; real N7 = encode_real_n7).
  • rev4 (2026-06-11): labels-only patch closing Codex CODEX-NNODE-ACTIVE-CONFLICT. The §5 roster annotation on membership_sha256 was changed from the deprecated (N1) alias to un-numbered membership, matching the canonical N-number reconciliation table. NO field name, roster order, N7 byte rule, DAG edge, digest, or authority semantics changed; this is an engineering-convention relabel, NOT owner/Codex ratification.
  • rev5 (2026-06-11): N6-availability prose update (lane FIX7_N7_N8_P7_AUTHORITY_INPUT_PREPARATION_MACRO_2026_06_11). After Codex ratified N6 d777e87c…b258c as a RATIFIED_ENGINEERING_VERIFIED_CANDIDATE (knowledge/dev/reports/architecture/codex-fix7-n-number-n6-targeted-recheck-2026-06-11/), the pre-ratification prose SEAL_REAL_N6_NOT_AVAILABLE in §6 and §6.2 is superseded/historical: the engineering N6 chain now exists. N6 remains an engineering candidate only — not an official seal/pin — and does not by itself authorize N7/N8/P7; the remaining blockers are the authority inputs (A1/A2/A3/A5, owner decision, Codex N8 signer/report set, P7 IDs) and Codex authoring. Text/labels only — the N6 digest, encoder, roster order, byte rule, DAG edges, and authority semantics are UNCHANGED, and this is NOT owner/Codex authoring of any seal.
  • Lane: FIX7_FINAL_AUTHORITY_SEAL_FAILOPEN_AND_PROVENANCE_PATCH_MACRO_2026_06_11
  • Authority of THIS doc: provisional-non-authority input assembly. T1 assembles candidate inputs ONLY. This document is NOT an approval event, NOT a seal, NOT N7. Codex/owner author the actual N7 at the seal.
  • Owner basis: OWNER_AUTHORIZATION_FIX7_AUTHORITY_CLOSURE_AND_SEAL_ONLY_2026_06_10 (prepare + route only; no blueprint approval; no implementation).
  • Executable contract: authority_seal_encoder.py (sha256 13344f92cafcaf0d07dcb21700bdb642f38b89351702e08080eacb0e957144b8) + authority-seal-encoder-spec.md/.json. N7's domain tag, roster, order, encoding, value grammar (§7), and provenance classes (§8) are defined there; Codex computes a real N7 by running encode_real_n7("N7", …, provenance, real_n6_available=True) — it invents nothing.

1. Explicit non-self-approval statement

T1 prepared this envelope under owner authorization limited to assembly and routing. T1 has NOT approved anything, has NOT computed or claimed envelope_manifest_sha256 as sealed, and CANNOT do so. Every value below is either (a) an engineering-verified candidate value reproduced from the Codex Recheck-9 V3 review, or (b) an explicitly marked MISSING_AUTHORITY_INPUT that only owner/Codex can supply. Any N7 value computed before the authority fields are supplied is fixture/rehearsal-only and invalid as a seal — and the rehearsal corpus is provenance-classed REHEARSAL, which encode_real_n7 rejects.

2. Packet V3 identity (engineering-verified, candidate inputs)

Field Value
packet_kb_root knowledge/dev/laws/tool-kiem-thu/packets/fix7-codex-recheck-9-2026-06-10/
packet_tree_sha256 b95df0a5d2f41f80bea0cef8621c1f8bb0f6b49a40175116418494ed4141ca6d
tracked_files 32
reconstruction Codex V3 fresh governed-MCP fetch: RECONSTRUCTION: OK, tree identical bidirectionally

3. Canonicalizer rev3 candidate identity (engineering-verified)

Field Value
document_id knowledge/dev/reports/architecture/t1-fix7-existing-system-refactor-execution-blueprint-2026-06-08/canonicalizer-fix7-canon-v1-ssot.md
revision 3
utf8_bytes 38756
sha256 49c386a9b9666c09786fc4f89bc79776b6046eaee6f4da6d8537d2c753b734d0
status CANDIDATE — independently byte-verified by Codex V3 §8; becomes authoritative pin ONLY at P7 seal
membership_digest f2bda8…fe251
provenance_class_required_for_real_seal ENGINEERING_VERIFIED_CANDIDATE (NOT REHEARSAL)

4. Codex final-seal reject (verbatim, with source paths)

  • Reject report: knowledge/dev/reports/architecture/codex-fix7-final-authority-seal-n7-n8-p7-2026-06-10/00-readme-first.md (rev1)
  • Checkpoint: knowledge/dev/reports/architecture/checkpoint-codex-fix7-final-authority-seal-n7-n8-p7-2026-06-10.md (rev1)
Verdict Value
Final status CODEX_FIX7_FINAL_AUTHORITY_SEAL_REJECT
AS-P1 PARTIALLY_CLOSED/FAIL (now closed by value grammar)
AS-P3 / AS-P4 NOT_CLOSED (now closed by value grammar + report-set)
Fail-open probes 8 ACCEPTED (now all REJECTED)
N6 provenance REHEARSAL not gated (now gated)

5. N7 fixed roster (executable — AS-P1)

N7 binds, in this exact order (domain tag FIX7_ACTIVE_AUTHORITY_ENVELOPE_MANIFEST_V1; full byte rule + value grammar in authority-seal-encoder-spec.md §3/§7):

schema_version · node_id · membership_sha256 (membership — un-numbered) · canonicalizer_sha256 (N2) · marker_fence_registry_sha256 (N3) · superseded_boundary_sha256 (N4) · guard_set_sha256 (N5) · active_corpus_sha256 (N6) · approval_event_id (A1) · approver_identity (A2) · approval_event_timestamp (A3) · owner_blueprint_decision (A5) · approval_scope

Output: envelope_manifest_sha256. N7 binds ONLY the six engineering sub-digests + the approval-event fields. N7 does NOT bind N8 or P7. Per-doc N1 digests are bound transitively through N6.

N-number label reconciliation (rev4, labels-only). membership_sha256 is the un-numbered active_corpus_membership_sha256 (frozen pin f2bda8…fe251) per the canonical reconciliation table (knowledge/dev/reports/architecture/fix7-n-node-numbering-reconciliation-under-tkt-v02-2026-06-11.json). Its historical (N1) annotation was a deprecated alias (S3:N1); under the canonical table N1 is the per-doc normalized_active_content_sha256[d] (bound transitively through N6) and membership is not numbered. N2 … N6 are already canonical. This relabel changes the engineering convention only — the encoder field name membership_sha256, roster order, N7 byte rule, DAG edges (N7 → N2,N3,N4,N5,N6,N1), and all digest math are UNCHANGED — and it is not owner/Codex ratification (binding numbering still requires ratification, G-DOC-1).

6. MISSING authority inputs + standing true blocker (DO NOT fabricate)

# Field Status Exact actor Exact next action
A1 approval_event_id MISSING_AUTHORITY_INPUT Codex/authority Codex mints the approval-event identifier (provenance AUTHORITY_INPUT) at seal time
A2 approver_identity (owner + Codex) MISSING_AUTHORITY_INPUT Owner + Codex Supplied inside the authorized seal event
A3 approval_event_timestamp MISSING_AUTHORITY_INPUT Codex/authority RFC3339 UTC stamp at seal time
A5 owner blueprint decision (OWN-1) MISSING_AUTHORITY_INPUT Owner Owner chooses an option in owner-decision-packet.md
N6 real non-rehearsal N1..N6 chain AVAILABLE — RATIFIED_ENGINEERING_VERIFIED_CANDIDATE (d777e87c…b258c, cert binding 055828db…6b96); NOT an official seal/pin; does NOT by itself authorize N7/N8/P7 Codex (consume at seal); owner/Codex (promote to OFFICIAL_PIN) Engineering half closed: real-N6 packet proved the candidate, Codex ratified it 2026-06-11. Codex sets real_n6_available=True over this candidate at seal time. The SEAL_REAL_N6_NOT_AVAILABLE text below is superseded/historical, not a current status.
A6 envelope_manifest_sha256 (N7 itself) NOT COMPUTABLE AS SEAL BY T1 Codex Codex runs encode_real_n7(…) over A1/A2/A3/A5 + a real N6 chain

6.1 Cycle correction (AS-P2)

The authoritative DAG is N7 → N2,N3,N4,N5,N6,N1; N8 → N2,N5,N6,N7; P7 → N2,N7,N8. Acyclic seal order N7 → N8 → P7 (has_cycle(EDGES) = False). Supplying detached_seal_sha256 or authority_seal_pin_sha256 to N7 is rejected SEAL_HASH_GRAPH_CYCLE.

6.2 Provenance gate (FINAL-AS-PROVENANCE, new)

encode_real_n7 requires a provenance class for every engineering and authority input. The rehearsal corpus is classed REHEARSAL and is rejected SEAL_PROVENANCE_REHEARSAL_BLOCKED. Missing → SEAL_PROVENANCE_MISSING; unknown → SEAL_PROVENANCE_UNKNOWN_CLASS; forbidden → SEAL_PROVENANCE_FORBIDDEN_CLASS.

N6 availability update (rev5, 2026-06-11). A real ENGINEERING_VERIFIED_CANDIDATE N6 chain now exists and is Codex-ratified: N6 active_corpus_sha256 = d777e87c…b258c, certificate binding 055828db…6b96, proven in knowledge/dev/reports/architecture/fix7-real-n6-provenance-under-tkt-v02-2026-06-11/ and ratified in knowledge/dev/reports/architecture/codex-fix7-n-number-n6-targeted-recheck-2026-06-11/. The pre-ratification statement SEAL_REAL_N6_NOT_AVAILABLE is therefore superseded/historical and is no longer the standing blocker. N6 remains an engineering candidate only — not an official seal/pin — and does not by itself authorize N7/N8/P7. The remaining blockers are the authority inputs (A1/A2/A3/A5, explicit owner decision, Codex N8 signer/report set, P7 authority IDs) plus Codex authoring; Codex supplies the AUTHORITY_INPUT/CODEX_AUTHORED classes and sets real_n6_available=True over this ratified candidate at seal time. Promotion of the candidate to OFFICIAL_PIN is an owner/Codex act at the seal, not performed here.

7. Routing

Route this envelope + n8-detached-seal-request.md + p7-codex-reseal-request.md + authority-seal-encoder-spec.md + authority_seal_encoder.py + owner-decision-packet.md to Codex/authority. Authority sequence (acyclic): real N6 chain → authorized approval event → Codex encode_real_n7encode_real_n8encode_real_p7.

Back to Knowledge Hub knowledge/dev/laws/tool-kiem-thu/packets/fix7-authority-closure-2026-06-10/n7-approval-event-input-envelope.md