Incomex AI Portal
KB-6C80

FIX7 N6 Special Analysis Report (2026-06-11)

7 min read Revision 1
tool-kiem-thufix7n-noden6provenanceactive-corpus2026-06-11

FIX7 N6 Special Analysis Report

  • Date: 2026-06-11 · Host: T1 · Codex consulted: NO · Production mutation: NO
  • Authority: provisional-non-authority, clarification only. Sources S1..S17 per fix7-n-node-source-map-report-2026-06-11.md.
  • Purpose: answer, from existing sources only, exactly what N6 is, why it is "REHEARSAL", and what would make it real — because the paused macro (…REAL_N6_PROVENANCE_CHAIN…) turns on this node.

1. What is N6 supposed to represent?

N6 = active_corpus_sha256 — a single digest over the active content of the whole 10-document active-authority corpus. Confirmed identically on all three surfaces (S1 comment, S2/S14 prose, S3 roster field 8): there is no contradiction on N6's meaning (unlike N1/N2).

  • It is NOT membership_sha256 (that is the digest over the 10 document_ids, frozen pin f2bda8…fe251).
  • It is NOT a per-doc digest (those are N1).
  • It IS an aggregate digest = digest("FIX7_ACTIVE_AUTHORITY_CORPUS_V1", sorted(corpus_records)) where each corpus_record = rec(doc_id, doc_status, section, kb_revision, per_doc_content_digest) (S1 produce()).
  • It is therefore best described as a corpus proof: it binds, per active doc, the doc's identity + status + active-section selector + KB revision + the N1 content digest. Its only DAG input is N1 (EDGES["N6"]=["N1"]).

2. Where is N6 currently produced?

By the canonicalizer SSOT (S1) --produce <docs_dir> path. It is fail-closed (R9-B2): any missing/extra/duplicate/extract-error/invalid member suppresses every candidate digest (SUPPRESSED_CORPUS_NOT_OK) and exits 4.

  • A real computed value already exists in evidence: the Packet-V3 logs/produce.log / evidence/produce-expected-output.txt show active_corpus_sha256 : d777e87c73d3b62d36789d9343f346102e98dbf301f2c93f7608470b876b258c over the 10 real active docs (membership reproduced as f2bda8…fe251).
  • So N6 is engineering-computable today and a concrete candidate digest is on record. This is important: the value is not missing.

3. Why is N6 currently "REHEARSAL"?

Not because the value cannot be computed — but because of how the authority-seal lane was wired and classed (S3 §8, S9):

  1. The authority rehearsal driver (S17) deliberately feeds placeholder corpus inputs (3/4/5/6×64) classed REHEARSAL into encode_node (the generic/rehearsal path), producing FIXTURE digests 6225f265…/b1f001b6…/3599f663… that are loudly labelled NOT-A-SEAL.
  2. The real path encode_real_n7 requires a provenance class for every input and blocks any REHEARSAL-classed input (SEAL_PROVENANCE_REHEARSAL_BLOCKED). Codex's reject (S10) was explicit: accepting the REHEARSAL-classed active_corpus_sha256 into a real N7 "would launder rehearsal evidence into authority evidence."
  3. Even with valid classes, the real path stays blocked by SEAL_REAL_N6_NOT_AVAILABLE until a real non-rehearsal N1..N6 chain is supplied in-lane with real_n6_available=True and first-hand evidence. The current authority lane never executed produce→classify→encode_real_n7 with real values — it only ran the rehearsal fixtures.

So "REHEARSAL" is a provenance-classification + lane-wiring state, not a statement that N6 is uncomputable.

4. What would make N6 non-rehearsal? (decomposed)

Step What Who T1-doable?
4.1 Run S1 --produce over the 10 real active docs in this lane, capture first-hand evidence (exit 0, corpus_ok, membership_frozen_ok, the active_corpus_sha256 value) T1 / engineering YES — engineering, deterministic, already demonstrated (d777e87c…)
4.2 Assign provenance class ENGINEERING_VERIFIED_CANDIDATE (not REHEARSAL) to the produced N1..N6 chain T1 proposes; Codex/owner ratify T1 proposes the class; ratification is authority
4.3 Feed the chain to encode_real_n7(…, real_n6_available=True) Codex (at seal) NO — gated to the real authority path the seal runs
4.4 Accept/ratify the candidate as the seal's N6 (promotion to OFFICIAL_PIN) Codex + owner NO — authority act

Key insight: the gap is part engineering-materialization (4.1, T1-doable) and part authority-ratification (4.2–4.4). The standing blocker SEAL_REAL_N6_NOT_AVAILABLE is genuine, but it is not "T1 cannot compute N6" — it is "no one has run the real chain in-lane and no authority has ratified the candidate." T1 can collapse the engineering half so the authority step becomes ratify-or-reject of a concrete, evidenced candidate.

5. Can Packet V3 engineering PASS promote N6?

No. S12 (Codex Recheck-9 V3): engineering PASS, but N7/N8/P7 remain AUTHORITY BLOCKED; the canonicalizer rev3 hash and tree "remain engineering candidates, not official authority pins." S10 repeats this for the corpus. Engineering PASS ≠ provenance promotion. Promotion to a real/official N6 is an explicit authority act, by design (anti-laundering).

6. Does Codex need to author N6? Does owner/operator?

  • N6 value: authored by T1/canonicalizer (engineering), not Codex.
  • N6 promotion (REHEARSAL/candidate → real seal input): Codex + owner/operator authority act. The provenance report (S9) and ledger (S13) both name the actor as owner/operator + Codex.

7. Is the N6 contract missing?

No. The executable contract exists and is proven live:

  • encode_real_n7(…, good_provenance, real_n6_available=True) is demonstrated to return a real N7 digest (S9 §3) — it is gated, not dead code.
  • The provenance class system, allow-lists, and the SEAL_REAL_N6_NOT_AVAILABLE gate are implemented and drift-checked (S3 §8, S9).

What is missing is inputs + authority, not contract: a real in-lane chain + its ratified provenance + the authority seal event.

8. N6 verdict

Question Answer
What is N6? active_corpus_sha256 — aggregate corpus proof over the 10 active docs; DAG input N1
Contradicted across sources? No (unlike N1/N2/N9)
Value computable now? Yes — d777e87c… on record; deterministic, fail-closed
Why "REHEARSAL"? provenance-class + lane-wiring; real path blocks REHEARSAL and stands at SEAL_REAL_N6_NOT_AVAILABLE
What makes it real? in-lane real --produce + ENGINEERING_VERIFIED_CANDIDATE class + encode_real_n7(real_n6_available=True) + authority ratification
Codex authors the value? No (T1 does). Codex+owner ratify/seal it.
Contract missing? No — gated, live, drift-checked. Inputs + authority missing.
Blocks final seal? Yes — SEAL_REAL_N6_NOT_AVAILABLE (G-AUTH-1) is a genuine standing true blocker.