<!-- DOC_STATUS: ACTIVE_NON_AUTHORITY -->

Checkpoint — FIX7 Authority / N-Node TKT v0.2 Alignment (2026-06-11)

Host: T1 / Claude Code. Final status: FIX7_AUTHORITY_N_NODE_TKT_V02_ALIGNMENT_NEEDS_OWNER_OR_CODEX_RATIFICATION. Codex consulted: NO. Owner approval requested: NO. Production/PG/Directus/registry/system_issues mutated: NO. T2 v0.2 framework mutated: NO. FIX7/V3 baseline mutated: NO. Real seal created: NO.

What was done (3 workstreams closed together)

1. Cross-lane alignment (A): documented T1/T2 boundary; T1 consumes 9 T2 v0.2 conventions; no parallel framework; no v0.2 mutation; no blocking T2-framework gap found.
2. Object-ID collision prevention (B): live registry read → ceiling 224; reserved non-colliding TKT-OBJ-225..240; governed via allocation addendum (no registry-JSON rewrite → no corruption risk).
3. N-node reconciliation (C): MC-1/MC-2 reconciled into a PROPOSED canonical table (labels-only, no math change) + executable n_node_checker.py (15/15, R1–R6 fail-closed). Binding numbering needs owner/Codex (G-DOC-1).
4. Defect re-verification (D) + self-Codex probes (G): encoder byte-exact reconstructed (13344f92…, 2 governed sources); every Codex-named fail-open defect re-verified CLOSED by direct bad-input execution; no patch required.
5. Packet (E) + fresh reconstruction (F): T2-convention review packet built + published; clean-dir RERUN PASS; tamper fails closed.
6. Report-vs-file (H) + firewall (I): no report-only evidence; firewall executable-by-reference to captured gates.

Captured evidence (real exit codes)

gate	result	exit
encoder --selftest	48/48 PASS	0
codex_probes.py	19/19 REJECTED, 0 fail-open	0
n_node_checker.py	15/15 PASS	0
sp_align_probes.py	10/10 fail-closed	0
commands.sh	OVERALL_RESULT PASS (5/5)	0
RERUN.sh (clean dir)	RERUN_RESULT PASS	0
tamper (file / tree)	FAIL closed	1 / 1

Encoder sha 13344f92…; packet_tree 96d00b9e….

Reality checks

Invalid input accepted? NO. PASS/digest/seal for invalid input? NO. Empty/duplicate report set accepted? NO. Empty id/signer accepted? NO. Invalid hash accepted? NO. Wrong revision/byte count accepted? NO. N1/N2 swapped outside alias? NO. N9_DIAG as load-bearing authority? NO. Unknown N10? NO. Candidate/rehearsal promoted to authority? NO. Report-only evidence? NO. Missing evidence file? NO. Object-ID collision? NO. Lane-boundary violation? NO. Production mutation? NO.

Why NEEDS_OWNER_OR_CODEX_RATIFICATION (not READY)

All safe engineering/evidence/governance work is complete and no engineering defect remains. The single thing T1 cannot finish is binding ratification of the canonical N-number table — per the governed model it "MUST be ratified by Codex/owner before binding." That, plus the standing authority blockers (SEAL_REAL_N6_NOT_AVAILABLE, A1/A2/A3/A5 inputs, OWN-1), are owner/Codex/operator only.

Deliverables

Packet root …/fix7-authority-n-node-tkt-v02-alignment-2026-06-11/ (11 files); reports: cross-lane alignment, object-id allocation, N-node reconciliation (md+json), checker-probes, defect re-verification, self-Codex probes, evidence-completeness, report-vs-file audit, authority firewall; governance allocation addendum; this checkpoint; current-state.

NEXT

Owner/Codex ratify the canonical N-number table (cheap, no production effect) → then run the real-N6 provenance macro (engineering half T1-doable) → then Codex/owner authority seal N7→N8→P7 (still gated).