FIX7 N-Node Authority Model — Next-Macro Recommendation

• Date: 2026-06-11 · Host: T1 · Codex consulted: NO · Production mutation: NO
• Basis: design addendum + source-map + N6 special + gap/contradiction reports (this macro). Sources S1..S17.

1. Decision

Recommended next macro: a small documentation-reconciliation macro FIRST, then resume the real-N6 provenance chain. Concretely, run them in this order:

1. FIX7_N_NODE_NUMBERING_RECONCILIATION_MACRO (new, cheap, no code/prod/seal) — adopt one canonical N-number→value table (closes G-DOC-1/2/3, the documentation contradictions MC-1/MC-2). T1 proposes the table (already drafted as PROPOSED CORRECTION in the addendum); Codex/owner ratify because it touches the load-bearing SSOT's numbering. Output: one ratified table referenced by the canonicalizer SSOT, doc 00, and the authority encoder spec.
2. FIX7_FINAL_AUTHORITY_SEAL_REAL_N6_PROVENANCE_CHAIN_MACRO (the paused macro) — now safe to direct, because the numbering is unambiguous. Its engineering half is T1-doable (materialize a real N1..N6 chain over the 10 real active docs in-lane, first-hand evidence, propose ENGINEERING_VERIFIED_CANDIDATE class); its authority half is Codex/owner (ratify the candidate, supply A1/A2/A3/A5, author N7→N8→P7).

Why not jump straight to N6 provenance? Because the host's own stated reason for pausing is correct: directing N6/seal work while N1/N2/N9 mean different things on different surfaces is unsafe. The reconciliation is one ratified table, changes no math/edge/digest, and removes the ambiguity that would otherwise contaminate the N6 macro's labelling.

2. The five candidate next steps, judged

Option	Verdict	Reason
Proceed with N6 provenance chain now	DEFER (do second)	Engineering is ready and even partly T1-doable, but the numbering ambiguity (MC-1) makes "which digest is N1/N2" unsafe to direct until ratified
Redesign the N-node model	REJECT	No edge/topology defect exists; the DAG is acyclic and the bindings are complete. Only labels drift. Redesign would be disproportionate and would churn a load-bearing SSOT
Ask Codex to clarify	PARTIAL-ADOPT	Codex ratification is needed for the numbering table and for N6 promotion — fold this into step 1 + step 2, not a standalone "clarify" round
Ask owner/operator for authority input	ADOPT (as part of step 2)	A1/A2/A3/A5, OWN-1 disposition, and operator materialization of the real corpus are genuinely required for the seal
Stop because contract undefined	REJECT	The contract is defined and executable (encoder, value grammar, provenance gate, cycle guard all closed). It is not undefined — it is un-fed and the labels are un-reconciled

3. Exact next actions (no work performed here)

• Step 1 actor = T1 (propose) + Codex/owner (ratify). Action: ratify the PROPOSED CORRECTION table in the addendum; cross-reference it from S1/S2/S3. No code change, no digest change, no production, no seal.
• Step 2a actor = T1 (engineering). Action: --produce over the 10 real active docs in-lane; capture exit 0, corpus_ok, membership_frozen_ok=true, active_corpus_sha256; propose provenance ENGINEERING_VERIFIED_CANDIDATE. Still not a seal.
• Step 2b actor = Codex + owner. Action: ratify the candidate (→ real_n6_available=True), supply A1/A2/A3/A5, dispose OWN-1, author encode_real_n7 → encode_real_n8 → encode_real_p7.

4. What still blocks the final seal after all of the above

• G-AUTH-1 (SEAL_REAL_N6_NOT_AVAILABLE) until a real chain is ratified — blocks seal.
• G-AUTH-2 (approval-event + Codex signer/report inputs) — blocks seal.
• G-OWNER-1 (OWN-1 do-not-approve) — blocks seal.
• IMPL-OWNER-AUTHORIZATION — blocks implementation (separate, post-seal).

None of these is curable by T1 engineering alone; all are authority/owner acts. This macro does not change that — it only makes them safe to direct.

5. Minimal safe next step (single line)

Ratify the canonical N-number→value table (PROPOSED CORRECTION in the addendum) — a labels-only, no-math, no-seal change — then resume the real-N6 provenance macro with its engineering half assigned to T1 and its authority half to Codex/owner.