◆ READER FRONT-MATTER — start here (narrative, not a physics claim)

The block between this line and the rule below is a plain-language on-ramp added for external review. It makes no scientific claim of its own; the frozen statement begins at the title "THE FRAMEWORK TOE — THE COMPLETE STATEMENT" below the rule.

The premise (the stakes)

Can the universe's deepest numbers — why there are exactly three families of matter, how the four forces split, the tilt of the early universe, the weight of the vacuum — be derived from one geometry, rather than measured and shrugged at? This series makes that bet on a single thirteen-dimensional shape: four measured numbers in, and twenty-two-plus independent quantities back out. But the real stakes are not the answer; they are the discipline — a method bound to an append-only record that would rather report an honest failure than a flattering success, and that, on its own record, refused the flattering answer four times and published the one place it genuinely broke (Λ).

The character we hitch the reader's wagon to (and why we trust them)

Not a theory, but a way of being honest under temptation, embodied by three figures:

• The honest author, who built the cage before theorizing and wrote each open claim's obituary in advance — naming, for every open item, the experiment and the year that could kill it, first, in public. (No byline is finalized; the author block is honestly absent pending entity formation — stated, not hidden.)
• The self-skeptical method (CFCA) — the real protagonist. Detective's eye, naturalist's patience, physicist's imagination, all turned on its own surviving explanation. It tested the theory's deepest hope against the cosmological constant, watched it fail, proved the failure inevitable, and published the refutation — and it retracted its own baryogenesis closure within a day, under its own audit, getting deeper each time.
• The AI co-investigator, the tireless second pair of eyes that ran the eliminations, froze the hashes before comparison, and held the min-rule against the temptation to round up — a partner in the discipline, never an oracle that outranks the record.

We trust them because what they are selling is the courage to be wrong in public: weakest links routed to the top, a falsifier on every open claim, the headline pinned to the worst row, self-flattery made a structural drafting bug.

What you're about to watch happen (this capstone)

The four papers form a quest — Can the shape build the world? (GUT, the load-bearing certificate) → Do the four forces come from the one shape? (Forces, the least over-claimed) → Does it survive being made real? (Quantum, which prints Σ = AUDIT on its own cover by the min-rule) → and this capstone, Was the whole quest honest?

This document is the ledger where the programme grades its own homework in public, with a fixed vocabulary, and catches itself — you will watch it reverse four of its own optimistic errors, numbered #2 through #5, and get deeper each time. The hardest number in physics — the cosmological constant — it tested its deepest idea against, failed honestly, and published the refutation; in the record below, Λ stands exactly as published — the honest fail, with no observed-Λ comparison made anywhere. On the live board that same line has since closed the honest way: the “catastrophe” dissolved as a continuum artifact, and Λ charged openly as the fifth measured anchor — measured, never derived. If you read only one thing, read §V — the truth table — where every one of the eighteen gaps lands in exactly one of four terminal states (closed-by-computation, closed-at-decision-grade, conditional-behind-a-named-gate, precisely-OPEN-with-a-falsifier-attached), kill-condition attached. Gap-04 is honestly CLOSED at decision grade, not certificate grade — its scoped close-out cleared the operative bound (35× at the weakest variant, 2.4×10³ central), with the residual c_loop named and proved unable to threaten the well, and the global-UV question explicitly delegated, not claimed. Of eighteen gaps, fourteen were deliberately OPEN at this freeze — because an open gap with a named kill-condition is a specification, not an admission. Those specifications did their job: the live board has since driven its full bill to completion — 33 gates resolved at +0, 0 open (ratified 2026-07-08; see the note above).

This is a state-of-the-programme ledger, not a finality claim. The pitch in one breath: this might actually be it — and it tells you, to the experiment, exactly where it isn't yet. Read on — and check us at every gate.

THE FRAMEWORK TOE — THE COMPLETE STATEMENT (TOE_FINAL_merged)

The merged, reader-optimized statement of the v6–v15 capsule line, 2026-06-13. Every line of this document is computation- or text-backed by the frozen capsule record and its countersign instruments; the document makes zero promotions; the Return YAML remains the only instrument of promotion. (The organizing voice is the theory's own native architecture — the ×-layer of metric geometry, the ⊕-layer chamber, the ⊗-layer bundles, threaded by the ℤ₂ fold and the ℤ₆ center-threading — glossed in full at §0.)

THE ONE-LINE THESIS

One thirteen-dimensional shape, four measured numbers, and a rule that the theory must publish every place it is wrong — and so far it has passed every test it could run, published in full the one place it failed, and named each test it hasn't yet run.

THE BIG IDEA (in ~200 words)

Take one thirteen-dimensional shape — ordinary spacetime times a small, stiff internal geometry — feed in four measured numbers (the Planck mass, the gauge couplings, the top quark's mass, one mixing angle), and read the rest of physics off the shape. The most arbitrary-looking fact in particle physics — why are there exactly three families of matter? — comes out the way a rubber band's winding comes out: an integer the geometry counts, with no dial to turn it to two or four. The same goes for the pattern of forces, the tilt of the early universe, and the weight of the vacuum, each long treated as a fact to be described rather than explained. Against those four inputs, the shape returns twenty-two-plus independent quantities (§I) — more answers than questions, the fingerprint of a theory that is constrained, not curve-fitted. (It even ships a test rigged to come out negative, so you can check the discipline is real, not just claimed — PCM-12, §I.)

Then comes the part that makes it different from every other "theory of everything." It grades its own homework in public, with a fixed vocabulary, and it catches itself: you will watch it reverse four of its own optimistic errors — numbered on its record, #2 through #5 — and get deeper each time (§II–§IV). The hardest number in physics — the cosmological constant — it tested its deepest idea against, failed honestly, and published the refutation (§III). And the same internal shape that forces three families into existence also builds an error-correcting quantum computer, designed and simulated on open tools (§IX).

The pitch in one breath: this might actually be it — and it tells you, to the experiment, exactly where it isn't yet.

WHY THIS MATTERS (read this first)

A theory of everything is supposed to do something no other kind of theory attempts: take one structure and let it decide, simultaneously, why there are three families of matter, how the forces split, what the early universe could and could not have done, and why the vacuum weighs what it weighs. The temptation in such a project is overwhelming — to bolt on a parameter here, soften a claim there, and quietly declare victory. The stakes are exactly that temptation. A theory of everything that flatters itself is worthless; one that survives its own hostile audit is the rarest thing in the field.

This document is the second kind. It states a single thirteen-dimensional input, reads the entire Standard Model and its cosmology off that input, and then — this is the part worth your attention — lays out every place the reading is not yet certain, with the exact experiment or computation that would kill it. Three things make it worth an hour of a skeptical reader's time:

1. The economy. Four measured numbers go in. Twenty-two-plus independent quantities come out (§I). More outputs than inputs is the signature of a theory that is constrained, not fitted — and the theory ships a registered blind-negative control to prove the discipline (PCM-12, §I).
2. The honesty is mechanical, not rhetorical. Every claim wears a status label from a fixed vocabulary (§"status vocabulary"); every gap sits in exactly one of four terminal states (§V); a promotion — calling something closed that is not — is a structural failure the programme catches and reverses on its own record. You will watch it do this four times (the numbered self-caught corrections #2–#5, §II–§IV). A theory that corrects itself in public and gets deeper for it is showing you its evidence chain, not hiding it.
3. The geometry is not abstract — it designs and simulates an error-correcting quantum computer. The same K₆ = SU(3)/U(1)² flag-manifold geometry that forces three fermion families into existence also defines an error-correction admissibility chamber that has been designed and simulated on open-source tools (§IX). When a geometry earns its keep in a simulated engineering design, "is this structure real?" stops being a philosophical question.

If you read nothing else, read §V — the truth table — where every one of the theory's eighteen gaps and thirty named instruments is placed in exactly one terminal state with its falsifier attached. A hostile reviewer should start there, find the weakest links already routed to the top (the five named in HOW TO READ, Known weakest links, below), and judge the rest against that standard of disclosure. Read it for what it is: a state-of-the-programme ledger, not a finality claim. Of eighteen gaps, fourteen are deliberately OPEN — each with a named falsifier and an owner — because an open gap with an attached kill-condition is a specification, not an admission. Read §V before judging completeness.

Provenance (MR-7, recorded — two originals, both byte-untouched). This document is the merge of the two source statements of the capsule line, neither of which is edited in place: - TOE_FINAL (2).md — the initial: the v6 native-three-layer statement plus the raw v7→v15 session capsules (127 lines, 14,582 bytes). sha256 at fork, FROZEN: 8efb2480c378eb1e96f55d8f290c287d0b7a5cc1bb8df730182282c51fee861b. Its raw version-history provenance — the verbatim per-session capsule text — is recovered in Appendix P of this document. - TOE_FINAL_June_12.md — the completion: the Phase-1.5 manuscript expansion of that capsule line (651 lines, 95,886 bytes). sha256 at fork, FROZEN: 4f837d8e0e435dc1b73ed30d01e50c782172395a2aec30c746b68a991038e1cf. It supplies this document's body (§0–§VIII), structure, and 48-item truth table.

This merged sibling takes the completion's complete, gate-passing body verbatim; recovers the initial's raw capsule provenance (Appendix P); adds the reader on-ramp above and a worked, simulated demonstration that the geometry is buildable (§IX, at engineering grade — explicitly not a claim about the physics); and advances no status anywhere. It reorganizes and expands what is banked — it does not promote it. If this document and either source could ever be read to differ on a status, the source capsule and its countersign record govern, and this document is the one in error.

Reader-optimization passes (MR-7, recorded — full pass-by-pass detail in the merged-build record near Appendix P). Four documentation-vertical passes added the reader on-ramp, intuition pumps, the steelman section, the cognitive-scaffolding ladder, and a deep line edit — clarity, ordering, and framing only, changing no claim's status or strength. The completeness gate held at exit 0 throughout (zero new dangling references, zero new stub sections, the MR-4 promotion-flag count unchanged at 1 — the pre-existing flag being inside verbatim Appendix P); the BG-10/Λ/θ̄/Gap-04 grades, the 48-item truth table, and the zero-promotions invariant are exactly as the merge left them, and the two source SHAs above are byte-untouched by every pass.

HOW TO READ THIS DOCUMENT

Nothing in this section makes a physics claim; everything after it does, under these rules. What follows gives the entry paths, the fixed status vocabulary and claim-strength ladder every labeled statement below obeys, the four terminal states that organize the truth table, and the weakest links routed first for a hostile read.

Reader map (the two audit paths)

The time-ladder below serves the reader who wants the stakes, the analogy, or the weekend-deep read. These two rows serve the reader who has come to audit — and they route differently:

What you can correctly take away, by how long you read

Every cell points to an anchor that already exists in this document — read down only as far as your hour allows, and what you carry away at each rung is correct on its own.

The document is paced so a genuine "aha" lands every few minutes — three families counted like a rubber band's winding (§I), the whole question of whether the theory predicts anything reduced to one curvature sign reversed in public (§III), the theory publishing the refutation of its own deepest hope (§III), the calendar of four ways to die (§IV), and the same shape building a quantum computer (§IX) — so the reward rhythm is deliberate, not incidental.

The status vocabulary (the approved label set)

Every quantitative claim in this document carries a status label from this set or quotes the capsule's native label verbatim. No vague descriptors ("promising", "suggestive", "near-complete") are used anywhere.

The claim-strength ladder

A reviewer who meets "proves", "closed", "unconditional", or "exact" anywhere in this document should find, inside the same passage, the computation, package, theorem, or countersign that licenses the word. If one is ever missing, that sentence — not the status record — is in error.

The four terminal states (the spine of the completion)

Every gap, lane, and named object in this document lands in exactly one of these four states. The terminal-state column is an organizational classification of the banked record; it is never a status change — the native capsule label is always printed beside it, verbatim-faithful.

Known weakest links (routed first, for a hostile read)

1. The Gap-04 closure is decision-grade, not certificate-grade. §II states exactly what the countersign cleared, by what margin, and what stays open inside the gap (c_loop; a non-blocking magnitude lane).
2. The Λ verdict is a computed honest FAIL: the theory's own deepest ⊕-layer cancellation claim was tested and refuted (§III). No observed-Λ comparison is performed anywhere in this document.
3. The inflation numbers ride two PENDING countersigns and one convention gate (K13-2). §IV carries the full four-branch table rather than one flattering band.
4. BG-10 is OPEN at 0/4 fixes; no baryogenesis number is claimed (§IV).
5. Strong-CP θ̄-smallness is honestly underived (§V stanzas).

How a skeptical referee can check this in an afternoon

A time-boxed reviewer does not need the weekend. Three of these four checks take minutes, and each names exactly where it lives — including the honest admission of what this particular document does not contain.

Four checks that take minutes, not weeks — and each is a copy-paste prompt in Test it yourself, immediately below. 1. The economy count (in-doc, ~5 min). Count the inputs (four anchors, §0) against the independent banked outputs (22+, §I ledger) — the over-determination signature (§0). The registered blind-negative control PCM-12 (§I) is the case built to come out negative, so the discipline is checkable, not asserted. 2. One by-hand arithmetic witness (~1 min). The live gauge-anomaly identity 3·(1/6) − 1/2 = 0 closes on a single line — the fastest hand-falsification entry point (worked as Prompt 2 below; the integer charge row it uses is the frozen q = 6Y row of §0). 3. Where the textbook reductions live (pointer, ~10 min to locate). This document makes no in-doc GR/Maxwell reduction — it does not claim one. The reduces-to-known-physics checks live in the siblings and are computed reductions (SIMULATED / CERTIFICATE grade there, not here): the Newtonian limit ∇²Φ = 4πGρ and Coulomb's law, and the Einstein–Maxwell conservation ledger matched to six significant figures on a hand-auditable charged shell (Prompts 1 and 3 below point you to them — EXTERNAL: the Forces and GUT sibling manuscripts). 4. The calendar of falsifiers (in-doc, §IV). Five named near-term kill-conditions — LiteBIRD r (PROJECTED), nEDM θ̄ (OPEN), the proton-lifetime reach τ_p (PROJECTED), the baryon window η_B (OPEN), and a dark-matter direct-detection null (CONDITIONAL) — each a published bound with its date, listed in full with its values in the §IV world timeline.

Test it yourself — five prompts you can paste into any AI (or do by hand)

Five checks you can paste into any AI or run by hand. Three verify against the sibling manuscripts (registered EXTERNAL, above) — this front-door document points you to them rather than reproducing them; two close inside this file. They inherit the disclosure of check 3 above: this document makes no in-doc GR/Maxwell reduction, so each prompt keeps its "consistency, not proof / does not replace GR" line verbatim, and the three sibling-homed prompts carry their EXTERNAL tags. Three are Type A (the geometry returns the same number as the textbook — a CONDITIONAL equivalence check, never load-bearing) and two are Type B (a genuine Standard-Model free parameter or unexplained fact, checked for consistency with independent data, never read as a from-nothing result).

Prompt 1 — Charged-shell field energy (Type A). What this checks: whether the geometry's bookkeeping returns the exact textbook electromagnetic field energy.

You are a careful physicist with a calculator. A thin spherical shell carries total charge Q = 1.00000 C at radius R_s = 5.00000 m. Two concentric Gaussian surfaces sit at r_1 = 1.00000 m and r_2 = 10.0000 m. Using only standard Einstein–Maxwell electrostatics, compute the electromagnetic field energy stored in the spherical region between r_1 and r_2 for this configuration (shell between the two surfaces), with the field given by the vacuum Coulomb law E(r) = Q/(4πε₀r²) for r > R_s and zero for r < R_s, energy density u = ε₀E²/2, and ε₀ = 8.8541878128×10⁻¹² F/m. Give the result to six significant figures in joules. Then state the closed form you used.

Expected (Type-A CONDITIONAL equivalence check, SIMULATED grade in the sibling): U_EM,12 = 4.49378×10⁸ J, from U = Q²/(8πε₀)·(1/R_s − 1/r₂). Check by hand: Q²/(8πε₀) = 4.49378×10⁹ J·m; (1/5 − 1/10) = 0.100000; product = 4.49378×10⁸ J. The geometry side (the ×,⊕,⊗ ledger) returns identical digits because it re-labels the bookkeeping, not the physics. Stop at U_EM; do not also reproduce Δm to six figures — the corpus 5.00079×10⁻⁹ kg figure is a squaring-of-a-rounded-value artifact, not a discrepancy. Honest framing: a consistency/equivalence demonstration recorded at CONDITIONAL grade, not a proof; it explicitly does not claim to replace general relativity or Maxwell. EXTERNAL: the GUT sibling manuscript, Appendix O.

Prompt 2 — One-line anomaly cancellation (Type A). What this checks: whether the geometry's particle list balances the gauge-anomaly ledger exactly.

You are a particle physicist checking gauge-anomaly cancellation by hand for one Standard Model generation. Take the quarks of one generation: a color-triplet left-handed quark doublet (multiplicity 3 from color) of weak-hypercharge Y = +1/6, and the left-handed lepton doublet of Y = −1/2. For the mixed SU(2)²–U(1) (and gravitational–U(1)) anomaly, the relevant trace over the left-handed weak doublets is the sum of their hypercharges weighted by multiplicity. Compute 3·(1/6) + (−1/2) and state whether it vanishes. Show the one line.

Expected: 3·(1/6) − 1/2 = 0 (3·(1/6) = 1/2; 1/2 − 1/2 = 0). The geometry inherits this because its index output is one SM generation per family. In this document the witness lives in the afternoon-checks box above (check 2) and reads the frozen q = 6Y row of §0; it is one of six anomaly ledgers (EXTERNAL: the GUT sibling, Appendix E′; the Forces sibling, §8.6.4). Honest framing: shows consistency ("inherited, not arranged"), not that the geometry is the sole origin of the SM.

Prompt 3 — Weak-field reductions to Newton and Coulomb (Type A). What this checks (CONDITIONAL interface claim, recorded): whether the higher-dimensional proposal lands on textbook low-energy physics.

You are a physicist auditing whether a higher-dimensional unification proposal reproduces textbook low-energy physics. (a) For gravity: state the weak-field, slow-motion (Newtonian) limit of general relativity as a field equation for the gravitational potential Φ sourced by mass density ρ, with Newton's constant G. (b) For electrostatics: state Gauss's law in differential form and the resulting Coulomb field of a point charge Q in vacuum. Write both standard results explicitly.

Expected: (a) ∇²Φ = 4πGρ; (b) ∇·E = ρ/ε₀, giving E = Q/(4πε₀r²). The geometry claims only to land on these same equations. Honest framing: these are interface claims — "recovers GR's weak-field limit, does not replace GR" and "exact electrostatic limit recovered"; reproducing a limit is a necessary consistency check, not a proof of the parent geometry (the gravity interface is additionally conditional on the volume-modulus freeze, declared soft spot WL-1). EXTERNAL: the Forces sibling manuscript — gravity ladder and electrostatics descent-ladder sections.

Prompt 4 — CKM CP phase (Type B). What this checks: the theory says one CP-violation angle comes out near 60°; the measured value is 65.5°. The prompt walks you through the convention so you compare the right two numbers — and asks whether 60 vs 65.5 is a real disagreement (it isn't, at <1σ).

You are a flavor physicist. The Standard Model does not predict the CKM CP-violating phase δ_CKM (the Cabibbo–Kobayashi–Maskawa phase, equivalently the unitarity-triangle angle γ); it is a free parameter fixed only by measurement. (1) Confirm that in the Standard Model this phase is an input, not a derived quantity. (2) Given a theoretical proposal that the phase emerges from an order-three geometric holonomy equal to −2π/3 = −120° (raw), which after the standard Wolfenstein phase-convention alignment corresponds to a compared value of +60.0°, evaluate the consistency of +60.0° against the measured PDG value of 65.5° for γ. Report the discrepancy in units of the experimental standard deviation (the "pull"), assuming roughly 10% structural precision on the prediction, and state whether +60.0° is consistent with data at that level.

Compare the Wolfenstein-aligned +60.0°, not the raw −120° — comparing −120° to 65.5° is a false ~185° mismatch. Expected: SM cannot derive δ_CKM (genuine free parameter); raw holonomy −2π/3 = −120°; Wolfenstein-aligned +60.0° vs PDG 65.5° → 0.79σ pull (σ ≈ 6–7° at ~10% structural precision; (65.5−60.0)/6.9 ≈ 0.79). Honest framing: agreement at 0.79σ is consistency with data at the stated precision — it does not prove the theory; a null result would have been a falsifier, a sub-1σ pull is a non-falsification, not a confirmation. EXTERNAL: the GUT sibling manuscript, §7.6 table and Appendices I/J.

Prompt 5 — Exactly three generations (Type B). What this checks: the geometry forces the family count to three by topology; does "exactly three" match what experiment allows?

You are a model-building physicist. (1) Confirm that the Standard Model does not explain why there are exactly three generations of matter — the number 3 is an empirical input, not derived from SM principles. (2) A theoretical proposal obtains the generation number as a topological index (a spin-ℂ / Borel–Weil–Bott index on the internal flag manifold K_6 = SU(3)/T², evaluating to −3, i.e. three chiral families), an integer that cannot be continuously tuned. Treating the family count as fixed by topology rather than by hand, verify that the value "exactly three" is consistent with experiment: state what the LEP measurement of the number of light neutrino species (N_ν ≈ 2.984 ± 0.008 from the Z invisible width) and the absence of fourth-generation discoveries imply about the allowed number of standard chiral generations.

Expected: SM does not predict the generation count; the geometry's topological index = −3 → exactly three chiral families, forced (not a dial); LEP N_ν ≈ 3 and null fourth-generation searches are consistent with exactly three, and the index-changing deformation that would give two or four families is excluded by the LEP N_ν bound. Honest framing: this is consistency with the observed count recorded at CONDITIONAL grade, not a from-nothing proof that nature must have three; minimality holds only inside the declared search category — a reviewer who supplies a natural excluded competitor triggers a documented downgrade to "category-relative diagnostic." EXTERNAL: the GUT sibling manuscript (index −3; LEP N_ν consistency anchor; the category-relative-diagnostic clause).

Naming conventions

Gaps are cited by the selector-registry numbering, Gap 01 through Gap 18 (one numbering caveat is recorded in §V; every Gap mentioned in this document is dispositioned in §V). Capsule versions are cited as v6…v15 — these are stable labels, and the campaign rule is that sibling papers cite capsule labels, not section numbers. Sections of this document are §0–§VIII. Named external artifacts are registered once, in the register immediately below, and are EXTERNAL by definition.

EXTERNAL ARTIFACT REGISTER (MR-1.b)

Everything this statement leans on that is not contained in this file, named once. A reference anywhere in this document to an artifact below is an EXTERNAL reference by definition: it resolves to the corpus, not to a section of this file. Nothing in this register is reproduced here, and nothing in it is modified by this document (the no-upstream-mutation rule, §VI: Papers 1–3 and the V1/V2/V3 closure-pack SHAs are byte-identical after this documentation pass).

Appendix H — EXTERNAL. The corpus stabilization appendix (Paper-1 appendix family). It is the declared authority for the Gate-8 κ₀′ closed form — the branch CLOSED at κ₀′ = −3/(64π⁶) < 0, working point at H's near-zero θ⋆ — and the routing target of the n=1 integer-sum tension (REGISTERED) under the corpus "when competing" clause. Cited in §II; not reproduced here.

Appendix D — EXTERNAL. The corpus reduction / slot-ledger appendix. It is the source of: the nine-row slot ledger consumed by the I2 supertrace (§III); the σ-kinetic denominator line — (D−2)=11, verbatim at source line L10268, confirmed from source in v15; and the D.1.0 uniform-breathing declaration that drives the §IV branch ruling. Cited in §III and §IV; not reproduced here.

Other registered EXTERNAL artifacts:

• gap04_frozen_package_native.json — the Gap-04 closure package (v9; its caveats travel inside the package).
• dispatch2/d15_native/ — the D15-native dossier suite: six dossiers, validators PASS, runner verified refusing at exit 2 until the seven countersign slots are signed.
• Run/package record ba0ba266d7315f71 — the v7 c_KK delivery record, FROZEN.
• The selector_geom_01 outputs — selector_results.json and extension_decision_tree.md (the June-10 selector run; see the §VI conflict ledger for its two superseded rows).
• Papers 1–3 and the V1/V2/V3 closure packs — the corpus authorities for the §I closed core (CLOSED rows, banked corpus-side); SHAs byte-identical under the no-upstream-mutation rule. The Tier-1 framework certificate (Gap 16) and the prediction registry with its resolved-refutation records REGISTRY-R-1 / R-2 / R-3 (Gap 18) live there.
• The handbook and the ERRATUM-K1 countersign packet — carry the worked determinant-identity line behind the §II erratum record.
• The Gap-07 verdict record (v14) — carries the five named expected-failure monotones, enumerated and counted.
• The Gap-13 dossier — carries the verbatim texts of the five anti-overclaim disclaimers PC-1..PC-5.
• The frozen σ-map yaml — the single read that decides the T6-vs-T_u branch question (§IV).
• The c_loop pre-registration runner and the BG-10 exit-2 checker (v15) — the mechanized discipline instruments.
• The Return-YAML instrument — the only instrument of promotion, corpus-wide.
• The QC engineering corpus — the Tahoe Labs error-correction design package and its evidence (the SIMULATION_CAMPAIGN/verify_campaign.py suite, the GATE_CLOSURE records, the K₆ = SU(3)/U(1)² admissibility-chamber implementation). EXTERNAL and sibling: it is the engineering bridge of §IX. Nothing in it is reproduced or modified here, and the no-target-loading rule (§VI Rule 4) bars any of its numbers from feeding back into the physics claims.
• The observed particle-spectrum closure package — EXTERNAL and sibling: the Observed Particle Spectrum Closure / Spectral Ledger Closure of the PDG hadron spectrum (a row-level claim ledger, not a derivation — a certified band of 14 source-backed imports, an explicit 863-row wall, and PROMOTED_UQF11_PREDICTION = 0). Its pending composite/resonance rows (absolute hadron masses, resonance poles, widths, lifetimes, branching ratios) are gated on the UQF-11 nonperturbative-QCD spectrum engine — the same Unified-Quantum-Force lineage as Paper III (Fable_Quantum) — which is specified, not solved (toy-world MVP kickoff, zero real wall rows promoted). EXTERNAL and sibling: nothing in it is reproduced or modified here, and the no-target-loading rule (§VI Rule 4) bars any of its rows or numbers from feeding back into the truth-table claims of this statement. The wall is named, not hidden; this statement makes no spectral-row promotion on its behalf.

Derived three-layer package (companion files in this folder — added at site-release; not corpus dependencies)

This statement is the merged source from which the reader-facing three-layer package was reformatted (2026-06-16, same day). No status was changed, no row promoted, no number altered: the companions below are derived FROM this document and are subordinate to it — where any companion conflicts with this statement, this statement governs. They are NOT upstream corpus artifacts of the MR-1.b register above; they are listed here only so this statement forward-references its own derived layers.

• README.md — package index and authority order for the files below.
• Scoped_TOE_ST1_main_statement.md — Layer ST1, the claim spine (.html / .pdf renders alongside).
• Scoped_TOE_ST2_Truth_Table_Rosetta.md — Layer ST2, Appendix STR (STR0–STR10), the truth-table explainer (.html / .pdf).
• Scoped_TOE_ST3_Formal_Authority_Package.md — Layer ST3, Appendix STA (STA0–STA10), the formal authority / 48-item registry (.html / .pdf).
• TOE_artifact_register__staged.md — the verbatim staged copy of this EXTERNAL ARTIFACT REGISTER (formal home: Appendix STA §STA1).
• TOE_provenance_ledger.md — externalized provenance / capsule pass-logs.
• TOE_QC_engineering_bridge.md — the externalized §IX engineering-bridge companion (no feedback into any truth-table row).
• TOE_reader_onramp_longform.md — the long-form reader on-ramp (verbatim from this front matter).
• TOE_reviewer_test_prompts.md — the Test-It-Yourself reviewer prompts (verbatim from this front matter).
• Handoff bundles: Scoped_TOE_ST1_handoff_docs.zip, Scoped_TOE_ST2_handoff_docs.zip, Scoped_TOE_ST3_handoff_docs.zip.

THE CAPSULE LINE, ASSEMBLED (v6 → v15)

Plain. Nine working sessions ran in two days, each appending one dated capsule to the record — never editing what came before. Read in order, the capsules are the theory finishing its own arithmetic, testing its deepest hope and publishing the refutation, and preserving its one loudly-open lane against the temptation to promote it. This table is the first time that line has been assembled into one object; every cell is verbatim-faithful to its capsule.

Precise. The assembled delta line:

Plain reading (for the v7–v10 cells above): "the well exists" means the geometry has a settled resting point for its size and shape — a "well" = a valley in the energy landscape that pins the geometry's size and shape so the constants can't drift (defined in full at §II). c_bdry and c_loop are the boundary and loop coefficients that decide whether that resting point is real; v10 settled the first (c_bdry) and proved the second (c_loop) can't spoil it. c_KK is the bulk-tower coefficient; AC-1.9 and β_∂ are the named arithmetic recipe and the boundary term it computes.

Every capsule in the line ends with the same recorded sentence: promotions: zero. That invariant is restated, and obeyed, at the close of this document (§VIII).

§0. THE ONE INPUT AND THE THREE LAYERS

Plain. The theory has one geometric input. Take a single thirteen-dimensional shape. Four of its dimensions are the where and when you live in — ordinary spacetime. The other nine — a six-dimensional squashed space, a two-sphere, and a folded circle — are not places you could walk; they are the what exists. Every particle species, every force, every coupling the theory speaks of is a persistent pattern that those nine dimensions force into existence. Nothing else is put in by hand except four measured calibration numbers. Everything downstream — the table of particles, the strengths of the forces, who may decay into what, what the early universe could and could not have done — is then read off, and each readout carries its grade and the experiment that could kill it. If the one shape were wrong, it could not hide: the same shape must answer every question at once, and §V lists where each answer stands.

Precise. The one input, as frozen at v6:

M₁₃ = M₃,₁ × K₆ × S² × S¹_Y/ℤ₂ — the ×-layer; K₆ = SU(3)/T²; radii frozen at R1.2, the native point being the ORIGIN of moduli space — carrying the F⁺ chamber (the ⊕-layer: the projectors, τ = ω, the finite claim-control data) and the bundle suite (the ⊗-layer: the seven-factor E_matter, the E_gauge connection, the E_Higgs Wilson sector), bound by the ℤ₂ fold and the ℤ₆ center-threading (the boundary/quotient structure threading all three layers).

First-use definitions, in reading order:

• The F⁺ chamber is the finite rulebook: a fixed set of projectors and the fixed choice τ = ω that together decide which geometric states are active and which are projected out. It is claim-control data — finite, frozen, and auditable — not a tunable dial.
• The bundle suite is the cast of actors: the fiber bundles over M₁₃ whose zero modes are the matter fields (E_matter, seven factors), whose connection is the gauge sector (E_gauge), and whose Wilson sector is the Higgs structure (E_Higgs).
• The fold and the threading: the ℤ₂ identification on S¹_Y creates boundary fixed points; the ℤ₆ — the center of the SU(3) that defines K₆ — "threads" in the sense that this one discrete structure runs through all three layers, and where it does it assigns the frozen integer row (q = 6Y) that §II's Wilson sums read. Boundary and quotient data thread all three layers — they are where the theory's localized structure lives.

Notation, recorded: T² and U(1)² denote the same maximal torus of SU(3); both forms (K₆ = SU(3)/T² and K₆ = SU(3)/U(1)²) appear in this document and are identical.

The layer table (the organization is load-bearing)

The layering is not bookkeeping. Each of §I's closed results lives irreducibly in its layer; §II's stabilization object is built from all three (a bulk Casimir term, a boundary term, a Wilson term, a loop term); and the v6 closing line — all three layers load-bearing — is restated at §VIII with its license.

Intuition pump: what the three hard objects actually are

The machinery above is abstract on first contact. Here is each load-bearing object in one plain image, with the precise statement kept right beside it so the analogy is an on-ramp, never a replacement.

• The coset K₆ = SU(3)/U(1)² (a "flag manifold") — the shape of a thing's orientations. Plain: think of all the ways you can orient a three-coloured frame in space, then throw away the information about which way it is merely spun about its own axes. What is left — the genuinely distinct orientations — is a six-dimensional curved space. That space is K₆. Why it does work: its curvature is rigid (the registry values R = 30, Ric = 5g are exact, FROZEN, §I), and a rigid curved space has a fixed number of ways a matter field can wrap it without unravelling. That fixed number is what an index theorem counts — and here it counts to three (the three families, §I).
• The F⁺ chamber — a one-way turnstile for geometric states. Plain: a chamber is a rulebook with exactly one verdict per state: admitted or rejected, nothing in between. Picture a turnstile that lets the active patterns through and stops their mirror-images cold. The chamber is that turnstile, written as a fixed set of projectors. Why it does work: because the rulebook is finite and frozen (not a dial you can turn), the same four anchors are forced through it to 22+ outputs (§I) — you cannot quietly re-tune what cannot be turned. This is the over-determination signature (§0 economy headline).
• The holonomy and τ = ω — the memory a loop keeps of the curvature it went around. Plain: carry an arrow once around a curved loop and it comes back rotated; the rotation is the loop's holonomy — its memory of the curve. Fix that memory to a specific cube-root-of-one rotation (the choice τ = ω, an order-three twist) and the geometry is forced to repeat its phase every third step. Why it does work: that order-three memory is the geometric origin of the CKM phase δ = 60° (§I) — Wolfenstein-aligned; the raw order-three holonomy is −2π/3 = −120°, and the +60° it aligns to sits at ~0.79σ from the measured PDG γ = 65.5° — a measured feature of the real world read straight off a loop's rotation, with no free phase to insert.
• Isometries-are-forces — a force is the internal shape's symmetry, felt. Plain: the ways you can rotate the internal shape into itself without changing it are the gauge symmetries SU(3)×SU(2)×U(1). A particle moving through space must keep its internal bookkeeping consistent with those symmetries; the adjustment it is forced to make is exactly what we measure as a force. Why it does work: the internal geometry's isometry group is fixed by the shape (it is not chosen), so the gauge group and its couplings are read off, not inserted — the same reduction Kaluza and Klein found for one circle, here carried by the full internal product (CLOSED, corpus-banked; EXTERNAL: Paper 1, and the §I ×-layer "Forces from isometries" row).

All three objects are rigidity made physical: a flag manifold with fixed curvature, a turnstile that cannot be re-tuned, and a loop-memory pinned to one value — none of them has a dial. Everything the theory then "predicts" is what a thing with no spare dials is forced to say.

The four anchors and the economy headline

The theory's declared numerical inputs are exactly four (MEASURED anchors; everything else is output or structure):

Against four inputs, the ⊕-layer returns 22+ independent banked outputs (§I). More independent outputs than inputs is the anti-fitting test (over-determination); the registered blind-negative control (PCM-12) is part of the same discipline. Honesty clause: the four anchors are calibration inputs; the experimental values quoted in §IV (Planck n_s, the r bound, N_eff) are comparison targets, never inputs — the no-target-loading rule of §VI applies corpus-wide.

§I. THE CLOSED CORE, BY LAYER

Plain. This section is the part of the theory that is already banked: results a reader may treat as settled at their stated grade, each with exactly one place to go and check it. Nothing here is new to this document — that is the point. The completion's rule for §I is cite-don't-re-derive.

Authority rule. Each row below names exactly one governing authority (its rightmost cell). Any other mention of the same result — in this document or its sibling papers — is descriptive, not load-bearing. The corpus authorities (Papers 1–3 and the closure packs) are EXTERNAL, registered above, and byte-identical under the no-upstream-mutation rule.

The ×-layer core (the stage)

Spelling reconciliation, recorded: the v6 closed-core line prints "ASP +3" once and "APS (+3, 0)" later; this document uses APS (Atiyah–Patodi–Singer) throughout. A spelling reconciliation, not a physics change.

The curiosity gap, asked and answered: how can one shape give exactly three families? This is the result most worth pausing on, so here it is in the "smallest toy → the real object → what was and was not shown" treatment.

• Smallest toy. Wrap a rubber band around a cylinder. It can sit there having gone around zero times, once, twice — never one-and-a-half. The winding number is an integer, and no smooth nudge can change it; it is a topological fact, not a force balance. A field on a curved internal space is the same story in more dimensions: the number of ways it can wrap without unravelling is a fixed integer, immune to fine-tuning.
• The real object. On K₆ = SU(3)/U(1)², that integer is computed by the Atiyah–Patodi–Singer (APS) index theorem — the rigorous machine that converts "how curved is this space, and how does the field twist on it" into a single signed whole number of left-handed-minus-right-handed zero modes. On this geometry the theorem returns +3 (CLOSED, two independent derivations, banked; EXTERNAL: Paper 1). An entirely separate Bott–Weil–Borel (BWB) zero-mode scan independently returns |index| = 3, checked against the frozen charge row. Two roads, same integer.
• What was shown, and what was not. Shown: the family count is derived, by theorem, as a topological invariant of the frozen geometry — three is forced, not chosen, and could not have come out 2 or 4 without a different shape (CLOSED, corpus-banked). Not shown here: this is the count of families, not the full inter-family mass ratios; those ride the Yukawa structure (the κ ladder and δ_CKM below, CLOSED at their own grades) and the flavor anchors. The honest boundary of the claim is "the number three is geometric"; nothing beyond that grade is asserted.

The takeaway is: the most arbitrary-looking fact in particle physics — why three? — is, on this geometry, the kind of fact a rubber band has. It is counted, not fitted, and the same count appears twice by independent routes. (You can run this consistency check yourself — see Test it yourself, Prompt 5.)

The ⊕-layer core (the rulebook)

The ⊗-layer core (the actors)

The inflow line (all layers at once)

The fixed points of the fold consistency-require localized structure, with exact uncancelled inflow coefficients: the geometry cannot be bare at its own corners. (CLOSED, corpus-banked; EXTERNAL: Paper 1.) Plain: the theory is not allowed to pretend its boundary is featureless — its own consistency conditions force structure to live there, and the coefficients of that requirement are exact, not estimated. This is the structural seed that §II's boundary term and Gap 11's boundary-localized candidate both inherit.

§II. STABILIZATION (GAP 04) — THE DECISION-GRADE DISCHARGE

Plain. The question: do the extra dimensions stay put? A theory whose internal shape can relax away — the radii running off to infinity or collapsing — predicts nothing, because every "constant" would drift. Gap 04 is the demand that the theory's own energy function for its breathing mode σ has a genuine well at the frozen radii. The story below is how that demand went from a campaign that had stalled, to an exact arithmetic with a falsifier attached, to a delivered number, to a countersigned discharge, to a runner PASS, and finally — by a sign flip the programme caught in its own boundary term — to a well that exists unconditionally in the one coefficient still owned by future work. Status at the end: CLOSED at decision grade. Not certificate grade — and §II says exactly which residue separates the two.

The re-diagnosis and the native object

The earlier flux campaign is re-diagnosed, on the record, as a ×-layer truncation — an attempt to stabilize using only the metric layer of a three-layer object. Its permanent assets are banked, not discarded (CLOSED, corpus-banked): the Saddle-Trap Theorem; the G-adjudication with gap = log₁₀(8π) identically; the Branch-B archive; and the warp kill. The flux track itself is moot-at-minimal per I1=G (v9).

The stabilization object is the corpus's own three-layer scaffold (Definition, frozen at v6):

V(σ) = c_KK e^{−4σ} + c_bdry e^{−2σ} + c_Wilson cosθ_W e^{−4σ} + c_loop e^{−6σ}

recorded 5/5 + 4/4 PASS at the frozen radii and capped CANDIDATE by coefficient bands at v6. Plain: one term from the bulk tower energetics (c_KK), one from the boundary the fold creates (c_bdry), one from the Wilson sector (c_Wilson), one from loops (c_loop). The gap closes when the signs and magnitudes of these coefficients, computed — not assumed — make V(σ) a well at the native point.

Intuition pump (the marble in the bowl). Picture the internal geometry's size as a marble, and V(σ) as the landscape it sits in. If the landscape is a hilltop, the marble rolls away — the dimensions de-stabilize and every "constant" of physics drifts with them. If the landscape is a genuine bowl, the marble sits at the bottom and stays. Gap 04 asks one yes/no question: is there a bowl at the frozen radii? Each of the four terms above is one contribution to the landscape's shape, and the gap closes only when the four computed signs add up to curving upward on both sides — a bowl, not a hill. The drama of §II is that the deepest term's sign was, at one point, attributed the wrong way; the programme caught it, the corrected sign made the bowl deeper, and the result is recorded below. In one image: the entire question of whether this theory predicts anything at all reduces to the sign of a curvature — and that sign was computed, reversed once in public, and then settled.

The exact pieces, to the banked boundary (v6)

Computed exact, with their validations (all closed-by-computation):

• The Wilson integer sums: S(0) = +4 and S(π) = +92, on the frozen ℤ₆ row q = 6Y. Exact integer arithmetic.
• The fold-winding bulk Casimir: c_KK^wind = +3.701826×10⁻² R_Y⁻⁴, computed by the scheme-clean Poisson/winding engine; validation gate: the massless limit reproduces −3ζ(5)/(128π⁶) to machine precision.
• The spin structure, adjudicated by the theory's own spine: the fermion spin structure is PERIODIC because the spectrum is a zero-mode inventory — APS (+3, 0); zero modes exist iff the towers are integer-moded iff the structure is periodic. The three families are the spin-structure witness. Corrected on the record (self-caught #2): the earlier "P-A1 ⇒ c_KK < 0" line had conflated zero-point and winding-potential signs; periodic fermion-dominance makes the winding share POSITIVE.
• The boundary coefficient, as first banked: c_bdry sign-positive at leading per-dof order (+11/8 per dof; two independent derivations of the leading share). Carried here because it was later superseded: the v10 full arithmetic reversed the attributed sign — see the boundary-sign event below. The capsule-line rule applies: a superseded entry stays on the record with its superseding event, and is never silently dropped.

The falsifier, the branch, and the engines that refused

The v6 PASS structure required the still-open curved shares (the K₆/S² Casimir plus the true graviton-internal table) to satisfy a two-branch bound: c_KK^[curved+gravint] < −3.721×10⁻² R_Y⁻⁴ on branch κ₀′ > 0, or < −3.253×10⁻² R_Y⁻⁴ on branch κ₀′ < 0 (BOUNDED, exact bound arithmetic).

The branch was then closed: κ₀′ = −3/(64π⁶) < 0, branch B, by the corpus's own declared Gate-8 formula (authority: Appendix H, EXTERNAL — working point at H's near-zero θ⋆; the n=1 integer-sum tension REGISTERED and routed to H's authority under the corpus's "when competing" clause). With the branch pinned, the operative falsifier became single-valued:

c_KK^[curved + graviton-internal] < −3.682×10⁻² R_Y⁻⁴ — the bound that traveled with the remaining work, attached as the falsifier.

The curved-share extraction was driven through four engine generations, each halted by its own validation gates — zero outputs accepted; the fail-closed machinery worked four consecutive times. The certified assets banked from that effort (closed-by-computation, FROZEN): the exact Kostant multiplicity generator (integer-gated; towers to λ ≈ 4052; scalar/vector/Dirac/tensor channels including the explicit 63-dof graviton-internal table), the full gate suite, the declared subtraction scheme, and the precision requirements (λ_max ≳ 10⁴; one unit-chain pin against the banked curvature code).

The delivered number and the well (v7)

c_KK = −8.892×10¹/R_Y⁴ by the corpus's own AC-1 ζ-recipe (sign decision-grade; record ba0ba266d7315f71, FROZEN). Therefore A = c_KK + κ₀′S(0) < 0 across the full variant band ⟹ the well exists; the corpus PASS structure is reproduced; the curved-share cap discharges, pending countersign. Supporting closures banked in the same session: odd-D μ-independence proven (the μ-[INPUT] dissolved — a theorem-level closure); self-caught #3 and #4 disclosed on the session record (EXTERNAL); five subtraction engines VOID with the impossibility theorem recorded (negative certificates — the engines that could not work are proven not to work, and say so); and the first-principles registry confirmations banked: R = 30, Ric = 5g, |Riem|² = 276, Vol = π³/2 — exact values, FROZEN.

Plain: the delivered bulk coefficient did not squeak past the bound. The weakest accepted reading clears it by a factor of 35; the central reading clears it by a factor of about two thousand.

The countersign (v8) — the ruling, verbatim

Chris ruling logged: H04_DISCHARGED_AT_DECISION_GRADE — the weakest accepted variant clears the operative bound by 35×, the central by 2.4×10³. The canonical claim, verbatim from the capsule:

Gap 04's curved-share falsifier is passed at decision grade; the remaining magnitude refinement no longer threatens the sign or bound, and c_loop remains handled under the declared scaffold/FRG-2 lane.

The c_loop window was frozen at that point: 0 < c_loop < A²/(4c_bdry); F1/F3 armed. (The window's later vacation is the v10 event below.)

The runner PASS and the closure package (v9)

Gap 04: CLOSED at decision grade. The D15-native runner ran PASS under Chris's permission; the closure package is gap04_frozen_package_native.json (EXTERNAL; its caveats travel inside the package). The D15-native dossier suite itself (six dossiers, validators PASS) remains staged at dispatch2/d15_native/ with the runner verified refusing at exit 2 until the seven countersign slots are signed — slot 0 the re-target ruling, S6 the parity [INPUT] confirmation (§V stanzas; §VIII queue).

The boundary-sign event and the Amendment (v10)

AC-1.9's β_∂ arithmetic was finally run: c_bdry = −1.08×10⁻² — negative, full band; the fermion parity pair cancels exactly (closed-by-computation; countersign slot PENDING). This superseded the v6 leading-order positive attribution — self-caught #5 on the prior b > 0 attribution, disclosed on the record.

The Amendment (countersign): Gap 04's well now exists UNCONDITIONALLY in c_loop. The v8 window is vacated and the closure strengthens: with the boundary coefficient negative across its full band, no value of c_loop in the declared lane threatens the well. Plain: the programme caught its own sign error, and the corrected sign made the well deeper, not shallower — the kind of event that distinguishes an audited record from a curated one.

The supersession record (window sharpening). v9's conditional sharpening, c_loop < 0.05·A²/c_bdry, is SUPERSEDED by the v10 Amendment: with c_bdry negative the sharpening condition is no longer operative. Recorded here as superseded — not silently dropped. The countersign-queue slot that carried it retires with this record (§VIII).

The erratum record (ERRATUM-K1)

ERRATUM-K1, recorded (identity-level): the handbook determinant entry det-42 is corrected to det-66, by identity. Queued at v9; recorded in this completion as an identity-level algebra correction — no new physics enters, the licensing identity is the handbook's own determinant algebra, and the worked identity line travels with the ERRATUM-K1 slot in the countersign packet (EXTERNAL: the handbook + the packet). The operative content of the erratum: every consumer of the handbook's det-42 literal re-points to det-66; nothing downstream in this statement consumes the superseded det-42 literal. The countersign slot certifies the entry (PENDING, §VIII).

What remains inside Gap 04 — and what does not

Terminal state of Gap 04: CLOSED at decision grade — by the countersign H04_DISCHARGED_AT_DECISION_GRADE plus the v9 runner PASS and package, strengthened by the v10 Amendment. Stated plainly: this is a decision-grade closure, not a certificate-grade one; the distance between the two is exactly the rows of the table above, and nothing else.

The FRG-4 three-modulus extension (2026-06-14 deep audit)

Plain. The decision-grade closure above stands exactly as written. This subsection records what a deeper look found behind it — not a correction to it, but a separate, more detailed object that the same machinery surfaced when the single breathing mode σ was extended to three moduli. The headline: the closure above is not weakened, the one open question it always named (c_loop) is sharpened into a single computable number, and the theory's accounting of its own irreducibles is driven to a terminal that contains zero genuine "just is" constants.

The §II closure stands; v10 is true as stated. The v10 Amendment's wording — Gap 04's well exists UNCONDITIONALLY in c_loop — is TRUE as stated for the single-σ §II potential V(σ): that statement quantifies over the c_loop residue, and the well's existence there rides on c_bdry < 0 (the v10 boundary-sign event), which is unchanged. The deeper three-modulus object below does not contradict it. The §II decision-grade discharge (the H04 countersign, the v9 runner PASS, the v10 Amendment) is byte-for-byte the operative closure; nothing here promotes, demotes, or re-opens it.

The scoped-closure correction (2026-06-14, Chris-authored — replaces the prior conditional framing for the deeper FRG-4 well). Earlier FRG-4 language treated the sigma -> -infinity corner as a candidate perturbative stability test. The completed audit shows this is not a valid perturbative decision point. The apparent runaway occurs beyond the heat-kernel/EFT ordering boundary; the sign-forcing cancellation is obstructed by the three-family index; and the admissibility chamber does not kinematically remove the shrinking-volume direction. Thus FRG-4 neither proves nor disproves global internal-volume stability. The scoped claim is local chamber-center admissibility and phenomenological sufficiency. Global stabilization is a UV-completion question, tied to the same vacuum-energy hardness as Lambda.

Therefore Gap-04 is closed only in the scoped sense: CLOSED-SCOPED / GLOBAL-UV-DELEGATED. The global shrinking-volume question is registered as B-UQFC-14-UV-1, an open UV-completion theorem.

Closure label for the Gap-04 FRG-4 well: PERTURBATIVELY UNDECIDABLE; UV COMPLETION REQUIRED (CLOSED-SCOPED / GLOBAL-UV-DELEGATED). This label applies only to the deeper FRG-4 global-stability well, not to the §II result; the decision-grade §II closure (vindicated, Closed-at-decision-grade) above is UNCHANGED.

Three banked structural theorems (each licensed in-passage with its script + run id). The completed audit banks three structural facts that, jointly, force the scoped-closure reading — the perturbative −σ corner is not a trustworthy decision point, and the missing wall is a UV-completion object, not a finite-order prediction:

• T1 — INDEX-CANCELLATION EXCLUSION. The full-spectrum supertrace residue cannot cancel — Str[1] = n_B − n_F = 35 − 90 = −55 ≠ 0, forced by the spin-c three-family index (−3): three unpaired chiral generations (no mirror, no SUSY partner) make n_F (90) > n_B (35) unavoidable. Same flavor as the banked Strong-CP mutual-exclusivity: chiral-three-families XOR sign-invariant modulus-stability. [gap04_full_supertrace_residue.py; run wovps8huo; countersigned, reproduced 2 independent ways]
• T2 — EFT-VALIDITY BOUNDARY. The perturbative −σ runaway lies OUTSIDE EFT validity — runaway onset (a₄ beats c_loop) at σ ≈ −2.72 is below the EFT-breakdown onset (a₆ beats a₄, unbounded ladder a₆ < a₈ < a₁₀) at σ ≈ +1.40; the breakdown is sign-SYMMETRIC, invalidating the unfavorable runaway AND any favorable a₆ wall equally (robust to 100× input change). So neither STANDS nor RUNAWAY is a trustworthy perturbative prediction. [gap04_higher_operator_wall.py; run wito6mc0z; countersigned]
• T3 — NO-KINEMATIC-FLOOR. σ → −∞ is an ADMISSIBLE deformation; no prior banked theorem floors the internal volume (Weyl-rigidity bounds SHAPE not volume — manuscript F.9.4 already says it is admissibility-not-stabilization; flux absent; the 3-family index is scale-invariant so it never jumps as Vol → 0; anchors scale-invariant). The corpus phrase that would manufacture the exclusion was explicitly refused (no "delete the bad direction"). [gap04_chamber_exclusion.py; run wg206lpk7; countersigned]

T1′ — INDEX-UV-LAMBDA UNIFICATION (BANKED STRUCTURAL RESULT, NOT an adopted axiom). This extends T1: the same spin-c three-family index (−3) that generates the chiral matter spectrum also obstructs the UV wall and ties to the Λ-hardness. Verbatim consilience statement: "The three-family spin-c index does more than generate the chiral matter spectrum. Because the index forces three unpaired chiral families with no mirror partners, it fixes n_F > n_B, hence Str[1] = 35 − 90 = −55. The same integer obstructs sign-invariant perturbative stabilization, prevents a derived positive leading UV wall, and appears as the leading a_0 spectral-action object tied to the cosmological-constant hardness. Thus the family count, the Gap-04 global-stability obstruction, and the Lambda-hardness trace to one index fact." This is a BANKED STRUCTURAL RESULT (consilience), NOT an adopted axiom — it is not B-UQFC-14-UV-1, it does not supply the missing wall, and it claims no global stabilization. [chase run wq91wo0x0 (all 3 UV routes hit a₀ = −55, the same leading spectral object as the cosmological-constant term); prior runs wovps8huo (full-spectrum supertrace, T1), wito6mc0z (T2), wg206lpk7 (T3); payoff search whz01rv21; all countersigned]

The next open theorem (named object): B-UQFC-14-UV-1 — Shrinking-Volume UV Wall / Resummed Determinant Theorem. Target statement: "For the selected K6^{W-rig} × S² × S¹_Y geometry, the full UV-completed determinant or non-perturbative sector is bounded below along the σ → −∞ internal-volume direction." Acceptance tests: (1) includes the full physical tower (metric, ghosts, gauge, scalars, fermions, projections, Weyl-rigidity) — no scalar-only shortcut; (2) does NOT rely on finite-order a₄/a₆ truncation; (3) does NOT choose the sign from the desired branch; (4) explains ≥1 other vacuum-energy fact, preferably the Λ-hardness relation; (5) preserves the three-family index and the frozen geometry. Status: B-UQFC-14-UV-1: OPEN — AXIOM-LEVEL UV COMPLETION REQUIRED (PROVEN NOT DERIVABLE FROM THE CURRENT GEOMETRY). It supersedes the prior framing in which Gap-04 was treated as an FRG-4 perturbative ambiguity. The separate +modulus decompactification corner remains its own precisely-OPEN object, B-UQFC-14-NP-1 (no perturbative wall there; every operator decays, V → 0⁺; candidate mechanisms — D-brane instanton, NS5-instanton, KKLT-type uplift — all outside the upstream perturbative corpus); the two corners are disjoint, and neither is the other.

B-UQFC-14-UV-1 axiom-level upgrade (2026-06-15, Chris-authored — the chase result, banked here). The B-UQFC-14-UV-1 chase upgrades the status of the global shrinking-volume problem from merely open to proven axiom-level. Using the full gauge-fixed tower rather than a scalar surrogate, all three UV routes — spectral positivity, resummed determinant, and small-volume phase exit — meet the same obstruction: the spin-c three-family index forces Str[1]=n_B−n_F=35−90=−55. This is the same leading spectral object that controls the cosmological-constant term in the spectral-action framing, so the Gap-04 global-stability obstruction and the Lambda obstruction are not merely analogous; they share the same a_0 source. Therefore global internal-volume stability is not derived from the current geometry. It requires a named UV-completion axiom, B-UQFC-14-UV-1, which remains open unless it earns adoption by explaining at least one additional vacuum-energy or UV-completion fact. [Chase run wq91wo0x0 (UV-1 axiom-level: all 3 routes hit a₀ = −55, the same leading spectral object as the cosmological-constant term); prior runs wovps8huo (full-spectrum supertrace, T1), wito6mc0z (EFT-validity boundary, T2), wg206lpk7 (no-kinematic-floor, T3); all countersigned.] This advances no status: the §II decision-grade closure above stays CLOSED at decision grade and byte-unchanged; the axiom is NOT adopted, B-UQFC-14-UV-1 is NOT called closed, and no global stabilization is claimed. The axiom WOULD SUPPLY global stability, but is NOT adopted. The audit proves that a positive global-stability wall is NOT DERIVED from the current geometry; the obstruction is localized to the same index-forced a_0 structure tied to Lambda.

Gap-04 Final Status.

Gap-04 terminal status: CLOSED-SCOPED / GLOBAL-UV-DELEGATED

Canonical Gap-04 status block (frozen labels — verbatim, authoritative over any looser phrasing elsewhere).

• Gap-04 scoped stability: CLOSED at decision grade.
• Global shrinking-volume stability: PERTURBATIVELY UNDECIDABLE; UV COMPLETION REQUIRED.
• B-UQFC-14-UV-1: OPEN -- AXIOM-LEVEL UV COMPLETION REQUIRED.
• Index-UV-Lambda unification: BANKED STRUCTURAL RESULT, not an adopted axiom.

The reason-hunt economy ledger — Gap 04 contains ZERO genuine "just is" numbers. The same audit sorted Gap-04's full residue on the reason-hunt economy ledger (gap04_reason_hunt.py, re-run 2026-06-14): 4 DERIVED + 4 EARNED-IRREDUCIBLE + 1 BARE-COUNTED, where the single bare item is the relative one-loop sign — and that item is itself computable-in-principle, now subsumed under the UV-completion object B-UQFC-14-UV-1 above. Because the deciding sign is a UV-completion / renormalization-scheme datum and not a free input, Gap-04 still contains no genuine irreducible "just is" number — the "just is" hypothesis is eliminated. Gap-04's honest terminal is a scoped-closed result whose global-stability question is UV-delegated, not philosophy.

Final Manuscript-Level Summary (Gap-04 close-out). Gap-04 is closed at scoped decision grade. The local chamber-center stability/admissibility claim and phenomenological sufficiency are preserved. The deeper global shrinking-volume question is not decided by FRG-4 or any finite-order perturbative heat-kernel calculation. Three structural results now delimit the obstruction: the three-family index prevents full supertrace cancellation, the apparent runaway occurs outside EFT validity, and no banked kinematic floor removes the collapse direction. The resulting global problem is therefore delegated to the named open UV theorem B-UQFC-14-UV-1. This theorem remains open and axiom-level; it is not adopted. The additional Index-UV-Lambda unification result is banked as structural consilience: the same spin-c index that gives three chiral families also produces Str[1] = -55, tying the global-stability obstruction to the leading a_0 object associated with Lambda-hardness.

Forward pull. Gap 04 was the theory's hardest favorable number — and it came out the way the theory hoped. The next section is the opposite test: the theory's deepest hope, run against the hardest number in all of physics, where coming out the way it hoped would have been the most impressive — and the most suspicious — result it could produce. Watch what it does instead.

§III. THE Λ LINE — THE COMPUTED VERDICT

Plain. The cosmological constant is the hardest honest number in physics: naive quantum estimates overshoot the observed vacuum energy by some 122 orders of magnitude (BOUNDED estimate), and Weinberg's 1989 no-go theorem forecloses the easy escape routes — so "Weinberg-open" below means exactly this: still unsolved, with the easy exits provably shut. The theory's deepest structural hope — the ⊕-layer's Chamber-Cancellation Theorem — was that its own chamber grading would force the offending contributions to cancel, exactly, at the quantum level. That hope was not debated. It was computed. It failed. Both the computation and the refutation are now banked certificates, and the honest final state is: the chamber route to Λ is closed, Weinberg stands, and Λ remains open — with the theory's Λ line now a pair of computed certificates where there was once an open mechanism hunt. Publishing the refutation at full strength, instead of softening it, is this programme's design working as intended.

A curve-fitter, handed its deepest idea and the hardest number in physics, would have found a way to claim a win. This theory ran the calculation, watched its best hope fail, proved structurally why it had to fail (I3, §III), and published both — claiming nothing about the observed value of Λ at any point. You are watching a theory decline to cheat on the exact question where cheating would have been most tempting and hardest to detect.

What the computation was set to decide (the v6 framing)

The exact-weight 9D graded heat kernel — inputs: the Appendix-D slot ledger (EXTERNAL), the C7 fiber dimensions, and the banked towers — was framed at v6 to simultaneously decide Λ Class 2 and probe the Chamber-Cancellation Theorem: the ⊕-layer's deepest claim, that the active/projected-out grading is exact at the quantum level, Weyl-odd, and Λ-protecting. Balance would have closed Gaps 05/06/07 together; imbalance would leave Λ Weinberg-open with a computation-backed elimination map. (The "together" coupling is now partially superseded on the record: Gap 07 closed independently at v14 — see the verdict below.)

Certificate 1 — the I2 supertrace (v12)

I2 RUN AND CLOSED AS HONEST FAIL. The I2 supertrace — a single signed sum over the theory's entire particle inventory that, had the chamber's cancellation hope been real, would have come out to zero — was run over the Paper-3 §4 nine-row ledger (ghosts/antifields formal; the 17-row physical inventory confirmed). It did not vanish: there is no structural cancellation at any coefficient order — the graded-to-ungraded ratio is 0.58 at k = 0 and 1.000 at k = 1–8. The chamber grading is coefficient-blind. Recorded with it: Str ρ at the frozen radii = (−88.93±band)/R_Y⁴ + c_loop; the K13-2 caveat carried; no observed-Λ comparison performed — the computation decides structure, not the observed value, by the no-target-loading rule. Countersign slots PENDING: the I2 verdict and the ledger authority correction (§VIII).

Certificate 2 — the I3 refutation (v14)

I3 THEOREM_REFUTED. Lemma 2 of the Chamber-Cancellation Theorem fails structurally: the Λ operator is the unit operator — grading-even and label-blind — so no grading of chamber labels can act on it, and I2's 0.58 residual at k = 0 stands as the quantitative witness of the failure. The refutation is a computed negative theorem, banked at full strength. Countersign slot PENDING: the I3 update (§VIII).

The verdict, assembled

The surviving lane, run to its end (June-14, OWNER-RATIFIED). The adjudicator row above has now been exercised: the three pre-registered relocations L1, L2, L3 were each graded against the 122-orders-of-magnitude radiative-stability burden — operationalized as four structural predicates that must all pass (R1 a real mechanism exists; R2 one symmetry-protected mechanism cancels at every tower scale — M_Pl, m_t, v_EW, Λ_QCD — with no per-scale re-tuning; R3 SM-mass-compatible; R4 non-perturbatively supplied) — and all three return BURDEN_FAIL. L1 (an exact discrete chamber-pairing symmetry G⁺ on chiral SM content) fails R1–R4: no such G⁺ exists — this is the Weinberg obstruction, and I3 already shows the Λ operator is the unit operator, so a chamber-label symmetry cannot act on it. L2 (a non-perturbative ΔV_NP wall along +modulus) fails R1/R2/R4: the perturbative path is structurally closed and a wall pins the modulus, not the 122-OOM zero-point — a wall is not a cancellation symmetry. L3 (chamber-projected boundary d.o.f. on S¹_Y) fails R1/R2/R4: the boundary-Casimir term re-renormalizes at every SM scale absent the exact bulk-tying symmetry (which is L1, refuted). Three BURDEN_FAIL ⇒ Λ is Weinberg-open PERMANENT for the Class-2 route. The owner has ratified this terminal (2026-06-14): "Lambda is Weinberg-open PERMANENT for the Class-2 route given L1/L2/L3 all BURDEN_FAIL; C19 slots I2 (v12 verdict + ledger-authority) and I3 (v14 update) CLEARED by owner; Lambda-1 PRESERVED-NOT-UPGRADED, Weinberg NOT cleared. Reopen only via a future L-construction passing the 122-OOM burden." The two C19 countersign slots (I2, I3) are thereby cleared. This ratification is a banked elimination, not a promotion: nothing about the observed value of Λ entered, Weinberg is not cleared, C_Lambda (Lambda-1) stays preserved-not-upgraded, and the line reopens only if a future L-construction passes all four burden predicates (terminal would flip to derived-reopens-toward-closure, requiring its own owner ratification). The build-or-refute test was verified to have teeth — an empty L0_NULL probe correctly fails, and a hypothetically-satisfied L1 correctly reopens — so the three refutations are genuine, not a rigged failure.

Plain, one more time: the theory looked for its vacuum-energy cancellation in the one place its own structure said it had to be, found arithmetically that it is not there, proved structurally that it could not have been there, ran the three surviving relocation candidates to refutation under a radiative-stability burden, and wrote all of it down. Λ is open — now Weinberg-open permanent for this route, with the surviving lane closed by elimination rather than left as a hunt. Nothing about the observed value of Λ was claimed, compared, or fitted at any point in the line. Promotions on the Λ line: zero.

Forward pull. So far the theory has stabilized its geometry (§II) and refused to cheat on Λ (§III) — both internal tests. The next section turns outward: it puts the theory in front of the actual universe and asks what it forecasts for the early cosmos, dark matter, the baryon asymmetry, and the arrow of time — and, for every forecast, names the experiment and the year that could end it. This is where the theory becomes falsifiable on a calendar.

§IV. THE SPINE AND THE WORLD — PREDICTIONS, FALSIFIERS, DATES

Plain. This section is the theory facing the universe: what it says about the inflationary epoch, the thermal history, the baryon asymmetry, dark matter, and the arrow of time — and, for every statement, the grade it carries and the observation that adjudicates it. The discipline to watch: computed verdicts are stated at full strength with their countersign slots marked PENDING; conditional results name their gates; the one loudly-open lane (baryogenesis) is preserved open with its preservation ruling quoted. Dates are at the end.

Inflation (Gap 08): one dead candidate, one survivor — with its convention honesty attached

The ρ-inflaton is DEAD (computed verdict, v9; countersign slot PENDING — the slot formalizes, the verdict is already on the record). The three evidential bases, verbatim-faithful: (1) the death is at qualitative grade — KK-stiff; (2) the noise floor of the run is disclosed on the record; (3) the NT-1 transverse-stabilization residue is REGISTERED — and separately, the NT-1 χ-leg run gate-halted on a scale artifact at v14, VOID both directions, routed to the machine lane, while the ρ-leg qualitative-stiff verdict stands. Nothing about ρ is claimed to be certificate-grade; DEAD-at-qualitative-grade is the recorded strength, and the countersign queue carries it.

The σ-inflaton SURVIVES (v9 round 1; corrected regime v10; countersign slots PENDING for σ-advancement, the plateau outputs, and the E2 second upgrade). The v9 round-1 hilltop slice — n_s ∈ [0.960, 0.964], r ≈ (2.9–3.6)×10⁻³ — is SUPERSEDED by the v10 plateau and carried on the record as such. The corrected regime:

An exponential plateau with λ² = 4/K_σσ = 1/6, set by the 13D split n = (6, 2, 1). n_s ∈ [0.9643, 0.9679] — Planck dead-center, near-universal across branches (the comparator is MEASURED; the band is the computed PROJECTED output). r ∈ [3.5, 10]×10⁻³ on the frozen operative branch — LiteBIRD-testable at 3.5–10σ (PROJECTED).

Plain: the surviving inflaton is the geometry's own breathing mode rolling down an exponentially flat hill. Its tilt lands in the middle of what Planck measured without any dial being available to put it there — the slope is fixed at λ² = 1/6 by how the thirteen dimensions split, and §VII records that no geometry was changed at any point in this line.

The four-branch table and the LiteBIRD falsifier (K13-2)

The A_s blocker decomposed at v11: the number closed (K13-1 CLOSED, below); the residual is K13-2, the σ-kinetic normalization — the D.1 kinetic denominator exceeds the standard reduction by exactly 81/22 ((D−2) = 11 versus (D₄−2) = 2 — the exact datum, recovered from the session record and source-confirmed at v15, FROZEN; verbatim at L10268 of the corpus reduction appendix, EXTERNAL). The exact four-branch integration (v11) gives four branches, labeled by two binary conventions: T = textbook/standard-reduction normalization versus D = the D.1 normalization, and the subscript 6 = K₆-only breathing versus u = uniform breathing across n = (6, 2, 1).

Assembly note (Interpretation strength, declared): the T/D and 6/u readings above are this assembly's reading of the banked v11/v15 lines — consistent with "D.1 exceeds standard reduction by exactly 81/22" and "D.1.0 declares uniform breathing — a TEXTBOOK ruling lands on T_u". The branch labels, bands, and survival statements themselves are verbatim from v11/v15.

The v15 events, in order: the D.1 source check is done — the kinetic denominator (D−2) = 11 is verbatim at source line L10268; the numerators match the standard reduction exactly; the 81/22 is confirmed from source. New from the source: D.1.0 declares uniform breathing, so a TEXTBOOK ruling lands on T_u (r ∈ [6.1, 17]×10⁻³, n_s Planck-centered), and T6-versus-T_u reduces to one read of the frozen σ-map yaml (the named gate; PENDING). The slip-versus-intent question is honestly left as the one remaining inference — this document does not resolve it, by rule (§V stanzas).

K13-2 single-axis owner ruling (OWNER-RULED-2026-06-15 — documented once). K13-2 is now a SINGLE-AXIS owner ruling, recorded verbatim as K13-2: OWNER-RULED / UNIFORM-BREATHING SOURCE-RESOLVED / D.1 NORMALIZATION RETAINED. The two axes are disposed of separately: the breathing axis is CLOSED — source-resolved by the D.1.0 uniform-breathing declaration (the operative breathing mode is uniform across n = (6, 2, 1)), not an authorial call; the normalization axis is OWNER-RULED — the remaining normalization ambiguity is not independently source-resolved, so the owner records the ruling that the D.1 denominator (D−2) = 11 is retained as intentional rather than corrected to the 4D textbook denominator (D₄−2) = 2. The ruling is supported by the explicit D.1 occurrence and by the resulting integer normalization K = 99/11 = 9 (the corrected denominator would give the non-integer 99/2 = 49.5); it is recorded as authorial, not as a theorem forced by the source — the source does not prove (D−2) = 11 was intended. The owner ruling retains the D.1-normalization reading as operative; it leaves the convention-independent union falsifier byte-unchanged and does not by itself clear the two PENDING C19 countersigns.

The pre-registered falsifier (BOUNDED, the union of all four branches): if LiteBIRD measures r outside [3.5, 36]×10⁻³, the σ-inflaton is falsified regardless of convention. Inside the union, the four r-bands separate at LiteBIRD resolution — the convention becomes an observable (v11): the experiment will not only test the model, it will adjudicate which normalization the geometry actually uses.

κ₁₃ and the A_s chain (K13-1 closed; K13-2 the single remaining normalization input)

K13-1 CLOSED: κ₁₃² = 1/(π⁴ M_U⁹ M_Pl²) — closed form, corpus-verbatim relation, at the frozen radii; the fold contributes a factor-2 band. (Closed-by-computation, v11; countersign slots PENDING for the κ₁₃ form and the branch table.) The A_s relation is FROZEN; the absolute A_s normalization is BLOCKED on the two named objects — K13-2 (the convention gate above) and c_loop (the Gap-04 residue, §II). No absolute amplitude is claimed in this document.

Reheating and the thermal history (Gap 09) — CONDITIONAL, gates named

Gap 09 is CONDITIONAL (v10; countersign slot PENDING): T_RH ∈ [2.4×10¹², 4×10¹⁵] GeV — the conditional band, mostly in-band against the corpus comparators — and N_eff = 3.044, in-band (BOUNDED conditional outputs). The gap re-gated at v11 on the two named inputs: K13-2 and c_loop. The conditional's own integrity test: once the gates close, the band must remain BBN-consistent and N_eff must stay in-band; an exit is a FAIL of the conditional, recorded as such.

Baryogenesis (Gap 10 / BG-10) — preserved OPEN, loudly

This is preserved OPEN by design: a same-day premature closure was withdrawn under audit and the gap held open at 0/4 — the discipline working against the programme's own optimism, recorded as such.

Lane 05 (BG-10): BLOCKED_BY_NAMED_OBJECT, preserved Open/Diagnostic — exactly as the corpus locks it: "Chris does the physics"; silent promotion = structural FAIL (v13, quoted). The preservation record, assembled:

• The four-fix ledger stands at 0/4 at certificate grade. The four named fixes any future BG-10 attempt must execute: (1) the canonical CI; (2) the proper ΔL = 2 treatment; (3) the N₂/N₃ damping; (4) the density-matrix flavor treatment.
• The history that locks it: the first-pass closure claim of 2026-06-04 was WITHDRAWN the same day after a deep ten-agent audit; BG-10 has been Open/Diagnostic since, and v13 re-affirmed the preservation under the explicit rule that a silent promotion is a structural FAIL.
• What is banked inside the open lane (diagnostic, labeled): the ingredients are FROZEN for the CP invariant (the M_R declaration at τ = ω is fix-ingredient #2, §I); the epoch is delivered as a band; the thermal-leptogenesis floor PASSES the entire T_RH range of §IV; the M_R threshold passes the upper band — a diagnostic estimate, labeled as such.
• The discipline instrument: the BG-10 exit-2 checker (v15, EXTERNAL) mechanizes the refusal — the lane cannot be run to a claimed closure without its named objects.
• Pre-registered falsifier for the eventual physics: if the completed four-fix computation returns η_B outside [6.0, 6.2]×10⁻¹⁰ at 3σ, BG-10 is falsified — a BOUNDED falsifier window, pre-registered; the gap itself stays OPEN until the four fixes run at certificate grade.
• Window-convention reconciliation (June-14, OWNER-RULED — documented once). The two bands that appear in the corpus are the same physical quantity in two conventions, not a discrepancy: the operative falsifier window is η_B ∈ [6.0, 6.2]×10⁻¹⁰ (the baryon-to-photon ratio), and the harness's (8.0 ± 0.3)×10⁻¹¹ band is the identical quantity written in the Y_B (baryon-to-entropy) form via η_B = 7.04 · Y_B (so 6.1×10⁻¹⁰ / 7.04 = 8.66×10⁻¹¹, inside the harness band). The owner has ruled (2026-06-14): "operative window eta_B in [6.0,6.2]e-10; the (8.0±0.3)e-11 harness band is the SAME quantity in the Y_B (baryon-to-entropy) form via eta_B = 7.04 * Y_B (6.1e-10/7.04 = 8.66e-11), documented once. BG-10 STAYS OPEN / fail-closed at 0/4; the four-fix manifest is owner physics, deliberately not rushed (the 2026-06-04 same-day withdrawal is the model)." The reconciliation pins one convention for the eventual single post-freeze comparison; it changes no status. BG-10 stays OPEN, fail-closed at 0/4 — the four-fix F⁺ manifest is owner physics, deliberately not rushed, with the 2026-06-04 same-day withdrawal as the standing model for refusing premature closure. The 7.04 factor and the band are carried as post-comparison data only; neither is read into any fix.

Plain: the baryon-asymmetry lane is the place this theory most visibly refuses to flatter itself. The machinery to claim it exists; the claim was once made, audited, and withdrawn within a day; and the standing record says OPEN until four named computations are done at certificate grade by the owner. The two η_B bands on the record are one number in two conventions, now reconciled once and pinned — not a loose end, and not a closure.

Dark matter (Gap 11) — a structurally selected candidate, CONDITIONAL, with a null-forecast that bites

Lane 06 (Gap 11): CONDITIONAL (v13). The candidate: the S¹_Y/ℤ₂ boundary-localized ℤ₂-odd freeze-in state (C-iii), structurally selected 17/18 across 5 reweightings of the selection audit — selected by the geometry's boundary structure (§I inflow), not fitted to any abundance. Banked gates and relations: the KK-overclosure gate PASSES across the whole inflaton band; the relation λ_p(0.12) ~ 10⁻¹¹ is FROZEN; the candidate's renormalizable stability (Y = 0) is §I-banked. The block: the portal normalization — the named object whose derivation gates the relic chain (with the Gap-09 chain behind it).

The null-forecast (recorded, forecast grade): σ_SI ~ 10⁻⁵⁵–10⁻⁶⁰ cm² ⟹ a direct detection falsifies the minimal candidate — the E6 prediction registry entry becomes a falsifier. Plain: the candidate is a faint, weakly-made particle that lives on the edge of the internal space (the boundary-localized freeze-in state above) and is kept from decaying by a mirror-symmetry (the ℤ₂-odd parity); the "KK-overclosure gate" just checks it cannot over-fill the universe with mass, and λ_p(0.12) is the frozen coupling that sets its tiny production rate. It is honest enough to predict that direct-detection experiments should see nothing; if any of them sees a WIMP-like signal, the minimal version of this theory's dark sector is dead, and says so in advance. One anchor slip in this lane was self-caught and superseded on the v13 record.

The arrow of time (Gap 07) — PASS

Gap 07: PASS (v14, closed-by-computation). Spohn's relative-entropy functional D(ρ‖ρ_ss) is the surviving universal arrow monotone — monotone at machine-zero violation under both generic and thermal GKLS dynamics — and the five expected failures (the candidate monotones that fail outside their validity classes, as expected) are named and counted in the banked verdict record (EXTERNAL). Scope honesty, stated: the PASS is a monotone-existence result within the GKLS dynamical class — it settles the arrow's mechanism question on the theory's open-system spine; it is not a derivation of the cosmological initial condition, which remains Gap 06, OPEN (§V). The selector's pre-v14 OPEN row for this gap is superseded by the capsule PASS — conflict ledger entry CL-2, §VI.

The CMB/LSS likelihood (Gap 12) — populated in shape, OPEN by rule

Gap 12 was shape-populated under Rule A at v9 (hash-traced) and v2-populated at v10 — and it stays OPEN, because all 8 of its cosmology-input upstream gates are OPEN or CONDITIONAL and the standing rule is populate only from upstream signed outputs. No likelihood number is claimed. The gap has no independent falsifier; it is adjudicated strictly by its upstream chain (Gap 04 CLOSED; Gaps 08, 09, the Λ line, and the rest per §V).

Black-hole entropy and the Page curve (Gap 13) — diagnostic inheritance, five disclaimers

The banked diagnostic: S_BH = A/(4G) is inherited on the 4D zero-mode sector at 0.0028% relative error — DIAGNOSTIC, labeled; inherited, not derived ab initio. The five anti-overclaim disclaimers PC-1..PC-5 are registered in the Gap-13 dossier (verbatim texts EXTERNAL); their collective force, as banked on the capsule line, assembled clause-wise: (1) the entropy agreement is an inheritance check, not a microstate derivation; (2) microstate fabrication is forbidden; (3) no Page-curve mechanism is claimed; (4) the topological rank remains admissible later, not invoked now; (5) the lane is BLOCKED by its named gaps. The recorded block was Gaps 01 + 04; Gap 04 is now CLOSED at decision grade (§II), so the operative block is Gap 01 plus the system-bath re-examination (Gap 14's first action). Status: OPEN.

Quantum-to-classical and the Born rule (Gap 14) — the owner's machine

OPEN. AOC-1/3/4 are scope-excluded (Exclusion strength — outside the claim boundary, not failed gates). The CH-2 KK-tower decoherence channel is an honest halt: NotImplementedError, cascading via Gap 01's a₆ object. The recorded next action is on the owner's machine: re-examine the system-bath split on the existing 13D fields first, before any rule-layer extension is considered (§VI). The v6 world-line phrase "Gap 14 on the owner's machine" is this lane.

The prediction registry and the world timeline (Gap 18)

The corpus prediction surface (registry EXTERNAL; governance CANDIDATE): 12 predictions — 3 CERT + 1 conditional-CERT + 3 CANDIDATE + 5 AUDIT-tier — under pre-registration rules A–F, enforced. Three empirical refutations are RESOLVED on the registry: REGISTRY-R-1, REGISTRY-R-2, REGISTRY-R-3 — adjudicated, recorded, and kept on the record (refutations are results; the registry's resolved rows are part of the theory's evidence that its falsifiers are live, not decorative). The 5 AUDIT-tier predictions stay open with their falsifier windows in the registry.

The world timeline (all PROJECTED adjudication windows unless marked otherwise):

Plain: between now and roughly 2035, the world will hand this theory at least four independent chances to die — the W mass, the primordial gravitational-wave amplitude, the proton lifetime reach, and the neutron electric dipole moment — plus a standing one (dark-matter direct detection, where for the minimal candidate any signal is fatal). The theory's posture toward all of them is pre-registered.

Forward pull. You have now seen the banked core (§I), the two internal tests, and the calendar of ways to die. The next section is the reckoning — all 48 items at once, read back as the theory's own report card.

§V. THE TRUTH TABLE — EVERY ITEM IN EXACTLY ONE TERMINAL STATE

Plain. This is the section a hostile reviewer should read first. Every gap the theory enumerates, and every named sub-item the capsule line created, appears below exactly once, in exactly one of the four terminal states defined in the front matter, with its native capsule status printed verbatim-faithful beside it and its falsifier or named gate attached. The same list that built the theory is here read back as its report card — one list, two tenses. If an item is missing from this section, that is a completeness defect of this document, not a hidden claim of the theory.

Authority rule for this section. The truth table is the claim-to-authority map: the rightmost column of each row is the single governing cite for that row's status. Any narrative mention elsewhere (§II–§IV) is exposition of the same banked record, never an independent claim.

Table V-A — the eighteen gaps

Read the shape of that column, not any single row. Two gaps land Closed-by-computation (Gaps 07, 17) and one Closed-at-decision-grade (Gap 04), each carrying its computation or countersign cite in the rightmost column; five are Conditional behind a named gate; the remaining ten are Precisely-OPEN, each with a falsifier or a recorded non-claim attached. That distribution — a banked core, a gated middle, and an open frontier where every opening is specified rather than vague — is the honest shape of the theory. A table that placed all eighteen rows in the first state would be the thing to distrust.

Table V-B — the named sub-items and instruments of the capsule line (the instrument-level audit ledger)

Generalists: skip to §VI — the gap-level verdicts in Table V-A above already carry every status; the thirty rows below are the machinery, for line-by-line reviewers. Each is one of the capsule line's named instruments, placed in its terminal state; nothing here changes a gap's status, it only itemizes the machinery behind it.

The precise-open stanzas (the must-stay-open set, in full)

Each stanza: what it is · why it is open · what closes it · who owns it · the pre-registered falsifier or adjudicator. Nothing in a stanza is softened toward "nearly done"; nothing is downgraded either.

Gap 01 — the a₆ object (OPEN_PHYSICS_BLOCKER). What: tr[a₆(L_grav^{d=13, de-Donder})], projected onto the cubic-curvature basis with K₆×S²×S¹_Y holonomy — the Seeley-DeWitt a₆ coefficient at d = 13 on the de-Donder graviton — is not computed. Why open: the computation is perturbative but heavy (reference frame: Gilkey 1995, Th. 3.3.1; Avramidi 2000, Ch. 4); it is the named missing object for the UV completion and the physical Hilbert space, and the CH-2 decoherence channel cascades on it. What closes it: the a₆ trace computation, machine lane after the Gap-04 infrastructure. Owner: the machine lane. Falsifier: if the computed a₆ trace violates the positive-definiteness requirement for physical Hilbert-space closure, Gap 01 is refuted at decision grade. Until then: walled at field grade.

Gap 02 — the Yang-Mills wall (Clay; OPEN by scope). See the Clay-walls statement below; the C1–C4 structural-compatibility passes are banked (CANDIDATE) and are not, and are never represented as, a nonperturbative proof.

Gap 03 — Strong CP (OPEN_PHYSICS_BLOCKER). What: θ̄-smallness is honestly underived. Why open, with the banked negatives stated: T-α1 Realization A is structurally disproved — θ_QCD is W-invariant on K₆ = SU(3)/T², so that route cannot produce the asymmetry; the T-α2 small-integer ansatz is CLOSED (computed: arg det = 11.71779, not a small integer). A THIRD negative is now banked (June-14, OWNER-RATIFIED): the mutual-exclusivity theorem. The June-10-selected route — T-α1 Realization B (a chirally asymmetric W-action) plus a finite topological-CP sector — was run to its honest end data-blind and is refuted. The Realization-B Z₂ sign-lever on arg det(M_u M_d) exists only if Q_L and u_R/d_R sit on K₆ bundles whose zero-mode Weyl representations carry opposite determinant-character; but every dim-3 su(3) irrep is the fundamental 3 or antifundamental 3̄, each a faithful S₃ 3-orbit family multiplet carrying the same sign det-character (a representation-theory theorem, window-independent — the committed [−6, 6] weight scan finds zero escapes, consistent with the proof). Therefore the chiral asymmetry (opposite det-character) and the three-family witness |index| = 3 are MUTUALLY EXCLUSIVE: any bundle supplying the asymmetry leaves dim ≠ 3 or a non-faithful orbit, i.e. it alters the three-family witness rep — which EXT-3 forbids (verdict ALTERED-ROUTE-REJECTED, fail-closed). The finite topological-CP sector is the same negative from the other side: the closed homogeneous K₆ supplies only a 2πℤ Pontryagin periodicity (net CP-even, 0.128 rad ≫ 10⁻¹⁰ — no θ-pinning), and any CP-odd projection would reuse the APS orbifold fold already spent on (+3, 0) chirality, again altering the zero-mode inventory. So θ̄ is NOT driven small without altering the zero modes — a third honest negative, banked alongside T-α1-A DISPROVED and T-α2 CLOSED. The owner has ratified this and declines to author a chirally-asymmetric Weyl action that breaks |index| = 3 (2026-06-14): "theta-bar-smallness honestly underived/OPEN. Recorded result: the mutual-exclusivity THEOREM (a chirally-asymmetric W-action and |index|=3 cannot coexist; every dim-3 su(3) irrep is 3 or 3-bar carrying a sign det-character) + the finite-topological-CP non-pin (closed-K6 Pontryagin gives 2piZ, CP-even, 0.128 rad, no net theta-pinning). OWNER DECLINES to author a chirally-asymmetric Weyl action that breaks |index|=3 (a degenerate patch). nEDM theta-bar > 1e-10 is the standing falsifier." Status is unchanged: Gap 03 stays OPEN / θ̄ underived — the theorem closes a route, not the gap, and flips no ledger row; the banked negative would itself reopen only if Chris (an Axiom-3 owner act) exhibits a K₆ field content with opposite-character zero-mode Weyl reps that nonetheless retains |index| = 3, (+3, 0)-APS, and a faithful 3-orbit witness — shown impossible for any dim-3 su(3) irrep, so an escape would require structure outside the homogeneous-line-bundle setting. Owner: Chris. Falsifier: an nEDM measurement returning θ̄ > 10⁻¹⁰ makes the current underived status a refutation-grade result; nEDM probes continuously (§IV timeline).

c_loop — the σ⁻⁶ coefficient (OPEN). What: the loop coefficient of V(σ) is not computed from first principles. Why open: it is genuinely the FRG-2 lane's object. What it does not gate: the well — the v10 Amendment (countersign) made the well unconditional in c_loop. What it does gate: the A_s absolute normalization and the Gap-09 absolute scale. Owner: Chris + the FRG-2 lane, with the pre-registration runner armed (v15). Falsifier: if FRG-2 returns c_loop outside the scaffold's structural range, the A_s absolute normalization is refuted at decision grade.

K13-2, the slip-vs-intent inference (OWNER-RULED-2026-06-15 — single-axis ruling). What: whether the D.1 kinetic denominator (D−2) = 11 — source-confirmed verbatim — was a slip or the intended convention. Disposition: K13-2 is now a SINGLE-AXIS owner ruling, recorded verbatim as K13-2: OWNER-RULED / UNIFORM-BREATHING SOURCE-RESOLVED / D.1 NORMALIZATION RETAINED. The breathing axis is CLOSED (source-resolved by the D.1.0 uniform-breathing declaration: uniform across n = (6, 2, 1)). The normalization axis is OWNER-RULED: the remaining normalization ambiguity is not independently source-resolved, so the owner records the ruling that the D.1 denominator (D−2) = 11 is retained as intentional rather than corrected to the textbook (D₄−2) = 2 — supported by the explicit D.1 occurrence and the integer normalization K = 99/11 = 9 (vs the non-integer 99/2 = 49.5), recorded as authorial, not a theorem forced by the source (the source does not prove (D−2) = 11 was intended). Owner: Chris (authorial). Falsifier (independent of the convention, byte-unchanged): LiteBIRD r outside the four-branch union [3.5, 36]×10⁻³ falsifies the σ-inflaton regardless of which branch was intended.

Gap 10 / BG-10 (OPEN/Diagnostic, preserved). The full preservation record is in §IV. The open statement: the four-fix ledger stands at 0/4 at certificate grade; the four fixes are named (canonical CI; proper ΔL = 2; N₂/N₃ damping; density-matrix flavor); the closure instrument is a signed manifest with F⁺ chamber-derived inputs (the lane stays OPEN until that manifest exists). Owner: Chris — "Chris does the physics" is the lock, and the exit-2 checker mechanizes the refusal. Falsifier: η_B outside [6.0, 6.2]×10⁻¹⁰ at 3σ from the completed four-fix computation.

Gap 14 — decoherence / Born rule (OPEN). What: the quantum-to-classical lane. Why open: AOC-1/3/4 are scope-excluded; the CH-2 KK-tower channel honest-halts (NotImplementedError) and cascades via Gap 01's a₆. What closes the next step: the system-bath split re-examination on the existing 13D fields, on the owner's machine, before any rule-layer move. Owner: Chris. Falsifier: a system-bath analysis yielding a structurally inconsistent pointer basis makes Gap 14 a decision-grade open blocker.

Gap 05 / the Λ line (OPEN — Weinberg-open, with its certificates). The full verdict is §III. The open statement: Λ remains Weinberg-open for the Class-2 route; the chamber route is CLOSED by the certificate pair; L1–L3 must relocate the cancellation or be refuted. Owner: the machine/theory lane. Adjudicator: an L1–L3 relocation construction passing the 122-orders-of-magnitude radiative-stability burden reopens the line toward closure; its refutation makes Weinberg-open permanent. The Λ line is a computed pair of certificates, not an open hunt — and that sentence is the line's final internal state, not a promotion.

D15-native — the seven countersign slots (ready-pending-countersign). What: six dossiers ready at dispatch2/d15_native/; validators PASS; the runner verified refusing at exit 2 until Chris signs seven slots (slot 0 = the re-target ruling; S6 = the parity [INPUT] confirmation). The v9 Gap-04 run executed PASS under Chris's permission; the slots themselves remain unsigned on the record and stay in the queue. Owner: Chris. Falsifier: if the S6 parity confirmation fails, D15-native is refuted and the D14 flux-track bridge — unchanged, open in its own lane — is the fallback.

NT-1 — the χ-leg (VOID; machine lane). What: the transverse-stabilization residue check. State: the χ-leg run gate-halted on a scale artifact — VOID in both directions (no information either way; the gate worked); routed to the machine lane for re-execution at the correct scale. The ρ-leg qualitative-stiff result stands (it is part of the ρ-death evidential base, §IV). Owner: the machine lane (+ the ρ-leg verdict slot in the queue). Adjudicator: if the corrected χ-leg run returns non-stiff, the NT-1 transverse-stabilization residue is vacated.

Gap 15 — the measure (CANDIDATE selection; the object OPEN). What: C_measure is rule-layer selected, but the load-bearing object — a measure μ on S_TOE compatible with the Σ_TOE conjuncts, or a finite-counting surrogate uniqueness proof — is not supplied (this is the OPEN object); the V2 audit found 6 survivors with identical first-3-conjunct signatures, and the 'active' label rides the V3-15 freeze pointer, not a Σ_TOE outcome. Owner: the theory/math lane. Pre-registered alternate adjudicator (recorded): if all 6 survivors are shown physically equivalent under the four anchors, uniqueness is established without μ.

Scope boundary as a strength: the two Clay walls, stated in one place

The theory keeps two walls standing deliberately, and records here — once, plainly — what each wall is, what the theory can say at the wall, and why the wall is not a failed gate. Binding rule (corpus-standing, recorded): a required gate may never be closed by narrowing scope, and an excluded sector may never be leaned on to close anything.

Wall 1 — the Yang-Mills mass gap (Gap 02). What it is: the existence of a mass gap for quantum Yang-Mills theory is a Clay Millennium Prize problem (Jaffe-Witten 2000). What the theory can say, at its banked grade: the C1–C4 structural-compatibility checks PASS (CANDIDATE) — the theory's gauge sector is structurally compatible with a gapped nonperturbative phase. What it does not say: those passes do not constitute, approximate, or substitute for a nonperturbative QCD proof — the mass-gap question stays OPEN at the wall, and no wording in this corpus is permitted to drift in that direction. Why the wall stands: the missing object is Clay-prize-level mathematics, explicitly out of scope. Adjudicator: a Clay-level proof for SU(3) on M₃,₁ with the theory's zero-mode spectrum — an external event; until it exists, the wall is OPEN.

Wall 2 — UV quantum gravity at field grade (Gap 01). What it is: the full UV completion and physical Hilbert space of the gravitational sector. What the theory can say: the model-definition-level structure is in place, and the single named missing object is precise — the a₆ trace (stanza above). What it does not say: anything beyond field grade; "walled at field grade" is the recorded posture, and the Gap-13/Gap-14 lanes that cascade on the a₆ object inherit the wall rather than tunnel through it. Adjudicator: the a₆ computation itself, with its positive-definiteness falsifier.

Wall 3 — the four anchors and the value-question (scope boundary, not a gap; 2026-06-14). What it is: the theory's only genuine irreducible free inputs are the four anchors — M_Pl, α_i(M_Z), y_t, |V_us| — four measured numbers in, against which the framework returns 22+ independent outputs (§0, §I). The value-question for these four — why these values? — sits at a scope boundary the theory states plainly rather than crosses.

• Framework-relative no-go (recorded). No constraint within the framework's enumerated structure — gauge + diffeomorphism + chamber-grading symmetries; anomaly cancellation; the index theorems; consistency — fixes the four anchors. Each lives in a sector the structure provably does not act on: M_Pl in the overall dimensionful units scale (every acting structure is unit-blind — the index theorems return dimensionless integers, anomaly cancellation is a dimensionless charge-ratio identity, the chamber projectors are dimensionless); α_i(M_Z) in the continuous IR initial condition of an RG trajectory (the structure fixes the gauge group and the running geometry, not the IR boundary magnitude); y_t in the absolute Yukawa normalization (the structure fixes the Yukawa structure — the order-three holonomy giving δ_CKM, the κ ladder, the family index — all invariant under an overall rescaling); |V_us| in the continuous first-row mixing magnitude (the structure fixes the mixing phase and texture, never the continuous magnitude). This is a framework-RELATIVE no-go — never an absolute or metaphysical impossibility: a deeper theory (a UV completion turning an anchor into an output) OR an observation (a measured relation among the anchors) could still impose a constraint, and that door is kept explicit and open.
• Anthropic test (CONDITIONAL, not a derivation). Varying each anchor and tracing to habitability gives a non-uniform, honestly-bounded result (order-of-magnitude / literature grade): α_i(M_Z) is anthropically necessary at a ~few-percent window (deuteron binding / diproton unbinding / the triple-alpha carbon resonance); |V_us| is NOT anthropically constrained — the honest counter-example, since ordinary matter is first-generation only and s↔d mixing plays no role in nucleosynthesis or chemistry; M_Pl and y_t are unclear/entangled (physics rides only dimensionless ratios for M_Pl; a one-sided vacuum-metastability upper edge for y_t). Every chain is "IF observers exist THEN …," never a derivation of a value, and the closure does not manufacture uniform fine-tuning.
• Classification (physics-neutral, none endorsed). For the framework-irreducible anchors the value-question is PHILOSOPHICAL — it has left the domain of this physics and lands in the standard trichotomy {brute fact / anthropic selection in a multiverse / design}, named physics-neutrally with NONE endorsed. This is NOT a physics gap — unless a future observation imposes a constraint, which would convert a free input into a constrained one and pull the question back into physics. The classification is the current terminal, explicitly reversible; the door stays open. (Full argument: rendered/GAP04_IRREDUCIBLES_PHILOSOPHY_CLOSURE.md.) Per the binding rule above, this boundary is not leaned on to close anything: it closes no gate and flips no ledger row.

The remaining boundary exclusions (the AOC-1/3/4 scope exclusions in Gap 14; the deliberate non-claims inside Gaps 06, 12, 13) are recorded in their rows and stanzas above. Exclusions are not failed gates; they are the theory stating where its claims end — and §VI's selector discipline is the only door through which that boundary may ever move.

Numbering caveat (recorded). The v6 world-line phrase "Gap 14 on the owner's machine" is carried by the selector registry under Gap 14 (quantum-to-classical — the owner's-machine lane); Gap 13 in the registry is the horizon-entropy/Page lane, which is blocked-by-named-gaps rather than owner-routed. This document uses the selector-registry numbering throughout; the caveat is recorded so the v6 phrase cannot be misread as a Gap-13 statement.

The counts (the truth table summarized)

Every row above carries its falsifier, adjudicator, or recorded non-claim; every CLOSED row carries its computation or countersign cite. Promotions effected by this table: zero.

THE STRONGEST OBJECTIONS, ANSWERED

Plain. A serious reader arrives with serious objections, and a theory that ducks them earns nothing. So here are the five strongest objections a hostile expert raises against a programme like this one — each stated at full force first, in the form its sharpest critic would use, and only then answered, by pointing at the exact place in this document where the claim → mechanism → certificate → falsifier chain already lives. No objection below is softened before it is answered; if the answer is partial, it says so.

Objection 1 — "More outputs than inputs is a magic trick. You can always claim 22 outputs from 4 inputs if you get to choose which 22 and grade them yourself." Stated at full force: over-determination is only meaningful if the outputs were fixed before seeing the data and could each have come out wrong. A theory that picks its scorecard after the fact is not constrained; it is curated. Answered: this is exactly why the programme ships a registered blind-negative control, PCM-12 (§I, §0) — a case constructed to come out negative, recorded in advance, so the discipline is checkable rather than asserted. The outputs are not hand-picked: δ_CKM = 60° (Wolfenstein-aligned; raw order-three holonomy −2π/3 = −120°, vs PDG γ = 65.5° at ~0.79σ), the κ ladder, the three-family index (a topological integer, not a fitted value, §I) are each forced by the frozen geometry with no spare dial (the intuition pump in §0). And the binding no-target-loading rule (§VI Rule 4) forbids any observed value — Λ_obs, A_s, r, η_B — from ever entering as an input; the comparison happens after the freeze, on the far side of the wall. The falsifier travels with each output (§IV, §V).

Objection 2 — "Decision-grade is just a polite word for 'not proven.' Gap 04 is your foundation, and you admit it is not a certificate." Stated at full force: if the moduli do not stabilize, the theory predicts nothing — every constant drifts. Calling the stabilization "decision-grade" rather than "certificate-grade" is admitting the foundation is not nailed down. Answered: correct that it is not certificate-grade — and the document routes that exact weakness to the top (Known weakest links #1) rather than hiding it. But "decision-grade" here is precise, not a hedge: the operative falsifier (a sign bound) was passed by a factor of 35× at the weakest accepted variant and ~2.4×10³ centrally (DERIVED, decision-grade; the H04 countersign, §II), and a self-caught sign reversal in the boundary term then made the well exist unconditionally in the one coefficient still owned by future work (the v10 Amendment, countersigned; §II). The single residue separating decision-grade from certificate-grade — the c_loop coefficient — is named, and it is proved (by the Amendment) that it cannot threaten the well. The gap between the two grades is exactly one table in §II, and nothing else.

Objection 3 — "You failed on the cosmological constant. A theory of everything that can't do Λ hasn't done everything." Stated at full force: Λ is the marquee problem. An honest FAIL is still a FAIL, and a TOE that leaves its hardest number open is incomplete. Answered: the theory does not claim to have solved Λ, and says so without flinching (Known weakest links #2; §III). What it claims is narrower and fully supported: its own deepest cancellation mechanism was tested and refuted by a banked certificate pair (the I2 supertrace honest-FAIL + the I3 theorem-refutation, §III). That is a result, not an absence of one — it converts an open mechanism hunt into a closed elimination with a named surviving lane (L1–L3) and a pre-registered adjudicator (the 122-orders-of-magnitude radiative-stability burden, §V). No observed-Λ comparison is performed anywhere. The honest reading: Λ is OPEN, Weinberg-open, with the theory's own escape route now provably closed — which is strictly more than most candidate theories can say about Λ, because most never compute the refutation.

Objection 4 — "Almost everything is OPEN or CONDITIONAL. Of 18 gaps, only 2 are closed by computation. This is a research programme, not a theory." Stated at full force: the truth table's own counts (§V) show 14 of 48 items precisely-OPEN and only a handful closed outright. That is a long way from a finished theory of everything. Answered: the counts are stated by the theory, in its own summary table (§V) — the objection is quoting the document's own honesty back at it, which is the system working. The reframing that matters: every OPEN item carries a falsifier or a named gate and an owner, so "open" here means precisely specified and routed, not vague. The closed core is not small in load: three families, the gauge sector, proton safety as an operator identity, the CKM phase, the arrow-of-time monotone (Gap 07 PASS), and reproducibility (Gap 17 CERTIFICATE) are banked (§I, §IV, §V). A theory that labeled all 48 items "closed" would be the thing to distrust; one that closes what it can prove, conditions what it can gate, and opens the rest with the experiment that settles each is showing its evidence chain. The honest claim is "a constrained framework with a banked core and a fully-specified open frontier," and that is exactly what §V records.

Objection 5 — "Self-correction is theater. Catching your own errors looks humble, but it just means the earlier claims were wrong — why trust the current ones?" Stated at full force: a programme that has reversed itself seven-plus times has been wrong seven-plus times. Past errors are evidence of unreliability, not virtue. Answered: the distinction is directional and public. Each self-catch made a claim weaker and was recorded against an append-only prior (Appendix P) that nobody could quietly edit — a sign reversal that deepened the Gap-04 well (§II), a decoder gain shown to be an undecoded artifact and corrected downward (§IX), a sub-2 qubit ratio retired as an algebra-clamp artifact (§IX). A curated record hides its reversals; an audited record timestamps them. The reason to trust the current labels is not that the programme is never wrong — it is that the mechanism that caught the earlier errors is still running (the completeness checker, the exit-2 refusal runners, the Return-YAML-only promotion lock, §VIII), and it is the same mechanism in two independent domains (physics and the QC engineering corpus, §IX). A theory that corrects itself in public and survives is showing you its evidence chain, not asking for faith.

The takeaway is: none of these five objections is answered by assertion; each is answered by a pointer to a banked certificate, a named gate, or a pre-registered falsifier already in this document. That is the difference between a theory asking to be believed and one inviting you to check.

§VI. SELECTOR DISCIPLINE FOR ANY EXTENSION — AND THE CONFLICT LEDGER

Plain. A theory of everything has one special vulnerability: every new observation is a temptation to bolt something on. This section is the immune system. It writes out, as binding discipline, what any candidate extension — a new sector, a new term, a new boundary structure, a response to a new observation — must survive before it may enter the theory; and it records the two places where the standing selector record and the capsule line currently disagree, with the resolution of each.

The seven standing rules (the June-10 discipline, binding)

1. Current geometry first. No extension is considered unless the current geometry structurally fails the requirement at hand. A want is not a failure.
2. Least-invasive extension only. If extension is forced, the rank ladder is fixed: rank 0 (current structure) → rank 1 (rule layer) → rank 2 (boundary) → rank 3 (topological) → rank 5 (compact factor). A candidate enters at the lowest rank that structurally suffices, data-blind throughout.
3. Selection ≠ closure. The selector identifies the route; executable physics artifacts (runs, certificates, countersigns) are still required before any status moves.
4. No target-loading. Observed values (Λ_obs, A_s, r, η_B, and the rest) are never read as selector inputs. The selector sees structure; the comparison to data happens after the freeze, on the other side of the wall.
5. No silent promotions — of BG-10, of Λ-1, of anything — under any framing. (The preservation rulings of §IV are this rule, applied.)
6. The QC-ratio gate. Any tempted geometry extension must pass the question "would it lower the physical:logical qubit ratio at 10,000 logical qubits?" — a structural gate, not a numerical one; it exists to prevent motivated geometry mutation (§VII). To be clear at first appearance: it is an internal anti-motivated-mutation gate for this corpus's shared geometry; it is not a physics claim, an experimental prediction, or a commercial result, and no QC number feeds back into the physics (Rule 4 above; §IX).
7. No upstream mutation. Papers 1–3 and the V1/V2/V3 closure-pack SHAs are byte-identical after any documentation-layer work. (This document complies: it is a sibling, and the source capsule's sha256 is recorded in the provenance header.)

The extension datasheet (what a candidate must answer to enter)

In the corpus's selector idiom: every candidate extension answers a fixed datasheet, in writing, before any work is dispatched. The questions, with the discipline each enforces:

An extension that is attractive but tunable fails the datasheet at EXT-1 (no structural failure) or EXT-5 (nothing to freeze that wasn't fitted). The standing preference is the one the corpus's own selection history recorded: the selector chooses the more expensive survivor when the cheap one is merely fitted and the expensive one is forced.

The June-10 routing record (what the selector has already selected)

The June-10 selector run (the REGISTERED record: selector_results.json + extension_decision_tree.md, EXTERNAL) selected extensions for 5 of the 18 gaps — and for none of them is execution unblocked by selection alone (Rule 3):

Routing summary for everything else: no extension. Every other open item in the v6–v15 line is a computation, a countersign, or external physics — 18 of 18 gaps require no new geometry extension beyond what the selector record already holds, and this completion introduces none (§VII).

THE CONFLICT LEDGER

Two rows of the June-10 selector run are superseded by capsule events that post-date it. Both conflicts are registered here — recorded, timestamped, and resolved by the standing precedence rule: a Chris countersign or a banked capsule verdict outranks a stale selector row; the selector is then re-run for that row only, and nothing else in its record is touched.

Cross-check, recorded: the other 16 gap rows of the selector run are consistent with the capsule line (the partial-advance rows — Gap 08's σ-survival and Gap 09's conditional band — are advances within OPEN/CONDITIONAL, not status conflicts). The two re-runs are queue items, not blockers: nothing in this document depends on them, because this document carries both records side by side with the precedence stated.

§VII. THE GEOMETRY-CHANGE LEDGER

Plain. The single most corrupting move available to a geometric theory under pressure is to quietly bend its geometry toward the answer it wants. This ledger records — checkably — that the completion you are reading required zero such moves, and it arms the standing test that any future move must face before it can even be discussed.

The ledger for this completion: ZERO geometry changes

Cross-check against the open-items inventory: none of the 18 gaps' open items requires a new geometry extension (§VI routing record) — every remaining object is a computation, a countersign, a yaml read, or external physics. The selector packet's own scoreboard rows — "new geometry extensions introduced: 0; silent status promotions: 0" — are reproduced by this document's content, and the registry values used everywhere a geometric quantity is restated are the frozen exact ones: R = 30, Ric = 5g, |Riem|² = 276, Vol = π³/2, the exact R(u), and the κ₁₃² closed form (the superseded legacy value of the |Riem|² invariant is banned corpus-wide and appears nowhere in this document).

The standing test (armed for any future proposal)

Any proposed change to a geometry object — the manifold factorization, a chamber/coset choice, a radius or moduli freeze, a code-distance or lattice choice, a layer assignment — MUST answer, in the manuscript that proposes it, with the reasoning recorded:

Does this change lower the physical:logical qubit ratio at 10,000 logical qubits?

The answer must be one of exactly three: YES (with the ratio delta and the mechanism), NO (the change does not touch the ratio — state why it is being made anyway), or UNKNOWN — BLOCKED (the ratio impact cannot be computed yet; name the missing input). A geometry change recorded without this question answered is an incompleteness defect in the document that records it — the change promises an improvement its document never substantiates. The gate is structural, not numerical: its purpose is to make motivated geometry mutation impossible to do quietly. It is an internal corpus-hygiene instrument, not a physics claim, an experimental prediction, or a commercial result, and no QC number ever feeds back into the physics (§VI Rule 4; §IX).

This document's answer, for the record: the question does not fire — zero geometry objects were introduced or altered. The test stands armed.

§VIII. CLOSING — FOUR ANCHORS, ZERO PROMOTIONS

Plain. The document ends the way the capsule line ends every session: by restating what went in (four measured numbers), listing what still waits for a signature (the countersign queue — every entry already computed, none assumed), saying plainly what remains, and recording the invariant that governed every word above: zero promotions. Precise: the four subsections below are, in order, the anchor restatement, the consolidated C19 queue, the updated truth-table sentence superseding the v6 closing count, and the promotion ledger of this document (which is empty, with its license).

The four anchors, restated

The theory's declared numerical inputs remain exactly the four MEASURED anchors of §0 — M_Pl (ordinary, by declared doctrine), α_i(M_Z), y_t, |V_us| — and nothing in the v6→v15 line, and nothing in this completion, added a fifth. Against those four stand the 22+ banked ⊕-layer outputs of §I, the exact stabilization coefficients of §II, the certificate pair of §III, and the pre-registered falsifier surface of §IV. Over-determination, not adjustment, remains the theory's anti-fitting answer.

The consolidated countersign queue (the C19 packet, grown through v15)

Every slot below formalizes a verdict that is already computed and recorded in this document; none of them is assumed anywhere above. The packet designation is C19 as of v14, grown through v15. All slots: PENDING.

The discipline instruments stand armed alongside the queue: the c_loop pre-registration runner and the BG-10 exit-2 checker (v15, EXTERNAL) mechanize the refusal paths, and C17's scope sentence is in the certificate with the COI checks standing alone.

What remains, said plainly

Plain. Here is everything left, in one paragraph. Inside the theory's own machinery: one renormalization-group computation (c_loop), one yaml read plus one honest inference (K13-2 and slip-vs-intent), one machine re-run at the right scale (the NT-1 χ-leg), one non-blocking refinement lane (the Gap-04 magnitude), and a stack of countersign slots that formalize what is already computed. On the owner's desk: the four BG-10 fixes, the Realization-B strong-CP route, the system-bath re-examination, and the Axiom-3 work behind it. At the walls: the a₆ trace, the Wheeler-DeWitt/WCH question, the measure μ, the L1–L3 relocation question — and the two Clay walls, standing. Facing the world: the W mass in 2027, LiteBIRD after 2030, the proton-lifetime reach near 2035 (all PROJECTED adjudication windows, §IV), nEDM and the neutrino fits continuously, and a dark-matter sector that has pre-registered silence as its prediction. Every one of these carries its falsifier or its named gate in §V — none of them is hiding, and none of them has been dressed as closed; every row sits recorded at its banked label. None of it is hidden. None of it is dressed as closed.

Precise. The updated truth-table sentence, superseding the v6 closing count: remaining internal physics = the c_loop FRG-2 lane; the K13-2 yaml read + the slip-vs-intent inference; the NT-1 χ-leg machine run; the non-blocking magnitude lane; the consolidated countersign queue (above); the owned opens (Gap 03 Realization B, Gap 10 four-fix ledger, Gap 14 system-bath re-exam); the walled opens (Gap 01 a₆; Gap 06 WCH; Gap 15 measure μ; the Gap-05 L1–L3 lane); and the two Clay walls. Counts, from §V: 22 entries closed-by-computation; 3 closed-at-decision-grade; 9 conditional-with-named-gate; 14 precisely-OPEN with falsifier or recorded non-claim. Total enumerated: 48. Nothing else is claimed.

Zero promotions — the invariant, last

Promotions made by this document: zero. Every status above is the capsule line's own, quoted verbatim-faithful with its computation, package, theorem, or countersign cited at the point of use; every PENDING marker is a slot the countersign owner has not signed; every OPEN is open on purpose, with its falsifier attached; and the one loudly-tempting lane (BG-10) is preserved exactly where the corpus locks it. The Return YAML remains the only instrument of promotion.

The honest landing. Here is the whole thing in one feeling. A single shape, four measured numbers, and twenty-two-plus answers it was never allowed to tune. A discipline so strict it catches its own mistakes in public and deepens the result each time. A refutation of its own deepest hope, published at full strength on the hardest number in physics. And a geometry concrete enough to build a machine. The reason to take all of that seriously is the same reason a referee should: the document tells you, to the experiment and the year, exactly where it isn't yet — and routes you to the place to check. It is not asking for your faith. It is handing you its evidence chain and daring you to break a link. Every link is in §V. Pull.

§IX. THE THEORY IS NOT ABSTRACT — ITS GEOMETRY DESIGNS AND SIMULATES A QUANTUM COMPUTER

Plain. Every section above asks you to believe that a particular six-dimensional shape, K₆ = SU(3)/U(1)², is physically real — that it is the actual internal geometry of the world, not a convenient piece of mathematics. There is a way to test whether a geometry is real that has nothing to do with particle physics: try to build something with it. A shape that is merely a bookkeeping device cannot be machined into a working device; a shape that is real can. The same K₆ flag-manifold geometry that forces three fermion families into existence (§I, ×-layer) also defines the admissibility chamber of a quantum-error-correction architecture that has been designed and simulated on standard open-source tools. This section records that bridge at its honest engineering grade — neither more nor less — because a theory whose geometry earns its keep in a buildable machine has crossed from philosophy into engineering.

Precise. This section makes no new physics claim about the TOE and changes no status in §V: the cosmology and particle content stand exactly as the truth table records them. What it adds is a single, independently-checkable fact — that the ⊕-layer's organizing idea (a projector chamber that admits one eigenspace and rejects its complement) is the same mathematical object as the admissibility chamber of an error-correction design, and that the design is real enough to simulate. The QC program is a sibling engineering corpus (EXTERNAL, registered below); nothing in §IX feeds back into the physics claims, and the no-target-loading rule (§VI Rule 4) forbids any such feedback.

The shared object: one chamber, two uses

The theory's ⊕-layer is built on a chamber (§0): a fixed set of projectors that admit the active geometric states and reject the projected-out complement — the finite, frozen, auditable rulebook that turns the geometry into a Standard Model. The quantum-error-correction architecture is built on the identical idea in a different register: a coset-constrained admittance algebra that admits only the +1 eigenspace of a Heisenberg–Weyl / holonomy stabilizer subgroup and rejects the complement. The bridge is exact at the level of the construction:

The point is not a metaphor. It is that the same coset and the same admit-one-eigenspace-reject-the-rest operation appear in both, and in the QC case the operation has been written as runnable code and exercised against a decoder.

What has actually been built and simulated (honest engineering grade)

The following are recorded at engineering grade — the grade of a simulated design, not of a fabricated and measured chip, and explicitly not at the physics corpus's certificate grade. Each carries its status label:

• The chamber concept is validated in simulation (SIMULATED). The admissibility-chamber projector is implemented and exercised; the error-correction stack it sits in (a four-layer design: GKP + chamber + surface + modular) runs end-to-end. The full simulation suite executes cleanly — verify_campaign.py exits 0 with its core checks PASS on live re-runs (SIMULATED; EXTERNAL, the QC corpus).
• The design is locally simulatable today (SIMULATED). The toolchain is open-source and installed: Stim + PyMatching (stabilizer simulation and decoding), QuTiP and scqubits (the bosonic/transmon layer), with EM parasitic extraction verified. The architecture is not a slideware concept; it is exercised against a real decoder at ~10⁸ shots.
• The honest overhead floor is competitive, not heroic (BOUNDED). The design's physical-to-logical memory overhead at 10,000 logical qubits sits in a ~9.6–14.5 band (anchor 14.5, frontier 9.6; BOUNDED, SIMULATED), with a Robust fallback of 3.24 as the conservative four-layer figure (BOUNDED). These are competitive numbers for a fault-tolerant memory; the formerly-quoted sub-2 ratios are retired — they were an algebra-clamp artifact and a single-module figure, withdrawn on the record (see the self-caught discipline below). No sub-2 ratio is claimed anywhere in this document.
• The decoder advantage is real and survived audit (SIMULATED). The biased-noise advantage of the deformed code was tested with a genuine Hadamard-deformed XZZX circuit in Stim with a deterministic observable; the design operates below threshold at its operating point (SIMULATED), with a measured single-logical biased advantage in the tens-fold range on the decoded logical channel.

Why the engineering discipline is itself the evidence

The reason to trust the bridge is the same reason to trust the physics: the QC program caught and reversed its own optimistic errors, in public, against an append-only record. The takeaway first, then the receipts: each correction below made the engineering corpus's own claim weaker in public — the same instrument as the physics record's self-caught corrections (§II's sign flips, §III's published refutation) — and a program that audits itself this hard, in two independent domains, is showing its work. The honest-discipline record carries at least seven such self-caught corrections — the engineering analogue of §II–§IV's four numbered physics self-catches (engineering detail — the receipts):

1. A "4.3×" decoder gain was REVERSED when a real decoded simulation showed it was an undecoded raw-syndrome artifact; the honest decoded figure (~8.4×, CI 4.2–16.9) actually vindicated the original conservative claim (SIMULATED).
2. A d ≥ 8 classical-signaling claim was DOWNGRADED to PAM-8-plus-parity once it was proven (with the galois library) that the construction was the kernel of a GF(2) parity check — the coset advantage survives only in the coherent quantum embodiment, and the design says so (REFUTED-class, recorded).
3. A 15–25% thermal-via benefit was LOWERED to ~6.5% when an open-source FEM solver included the spreading resistance the closed form had ignored (SIMULATED; the larger figure requires more vias or a commercial solver — stated).
4. An SU(N) selectivity escalator was LOWERED when the admitted fraction was shown to scale as 1/(N+1) — decreasing with rank, not increasing (the inconvenient direction, recorded).
5. A success probability was corrected from a 91% lever-success model to a decision-grade ~12.6% when the levers were re-scored honestly (BOUNDED).
6. A circuit time-constant that was 1000× off (1 ns vs the correct 1 ps) was found and fixed (SIMULATED).
7. The design is recorded as FAILING the Shor-2048 error budget even at an idealized unit ratio — a floor ratio is not a runnable algorithm, and the program writes that limitation down rather than burying it (recorded).

Each of these is a place the engineering corpus made its own claim weaker in public — which is why the chamber bridge is offered as evidence rather than decoration.

What §IX does and does not license

The §VI selector already encodes this bridge as discipline, not decoration: Rule 6, the QC-ratio gate, requires any tempted geometry extension to answer "would it lower the physical:logical qubit ratio at 10,000 logical qubits?" (§VII arms the standing test). That gate exists precisely because the geometry is shared: a change made to flatter the physics could silently harm the machine, and vice versa, so the one geometry must answer to both. §IX is that gate's positive face — the record that, used forward, the geometry builds something.

Status of §IX, for the record: an EXTERNAL engineering bridge at simulated grade; zero TOE statuses changed; zero promotions. The QC corpus (the Tahoe Labs design package and its SIMULATION_CAMPAIGN / GATE_CLOSURE evidence) is EXTERNAL and registered here once; nothing in it is reproduced or modified by this document, and nothing in it moves a single row of §V.

Appendix P — THE RECOVERED RAW CAPSULE PROVENANCE (the initial, verbatim-faithful)

Why this appendix exists. The body of this document (§0–§VIII) is the completion's manuscript expansion of the capsule line. The initial (TOE_FINAL (2).md) carried something the completion reorganized away: the raw, per-session version-history capsules exactly as they were appended to the record, session by session, never edited after the fact. That raw provenance is a load-bearing part of the audit trail — it is the primary evidence that the capsule line grew by append-only accretion and that every later strengthening (and every self-caught reversal) is timestamped against an immutable prior. This appendix recovers it, verbatim-faithful to the initial, so the merged document is self-contained: a reader can compare the polished §II–§IV narrative against the raw capsule it was built from, in one file. Nothing here is a new claim; every line is the initial's own, and where the initial and the completion's body could differ, the body's labeled statuses and the countersign record govern (the initial's "v6" framing of "Gaps 05/06/07 close together" is, for instance, superseded in part by the v14 independent Gap-07 PASS — see §III).

P.0 — The initial's masthead (verbatim-faithful)

THE FABLE TOE — THE THEORY AS IT STANDS (v6, native three-layer form; 2026-06-11). Written, per the owner's instruction, in the theory's own architecture: the ×-layer (metric geometry), the ⊕-layer (chamber), the ⊗-layer (bundles), with the boundary/quotient structure threading all three. Every line computation- or text-backed; zero promotions ever.

P.1 — The v6 native-form statement (the initial's §0–§V, verbatim-faithful)

The v6 body of the initial stated the one input and the three layers; the closed core by layer; the Gap-04 stabilization arithmetic to the banked boundary (with the operative falsifier c_KK^[curved+graviton-internal] < −3.682×10⁻² R_Y⁻⁴ attached); the Λ line as one computation; the spine-and-the-world line; and the native-form truth table. The completion expands each of these into §0, §I, §II, §III, §IV, and §V respectively — the mapping is one-to-one and is recorded in the §0–§V headers of the body above. The v6 closing line of the initial, verbatim-faithful:

Remaining internal physics: one specified machine computation (Gap 04's last number, bound −3.682×10⁻², generator and gates shipped), the one deep computation (the Λ line), the gated chain behind the countersign, and one evening. Every coefficient of the stabilization object is now exact, sign-exact, branch-closed, or double-bounded — and the last open number arrives with its own falsifier already attached. Fully explicit. Fully falsifiable. All three layers load-bearing. Incapable of hiding anything from itself.

P.2 — The v7→v15 session capsules (verbatim-faithful, append-only as recorded)

Each capsule below is the standing-record entry appended at the end of one working session. They are reproduced in order; each ends with the recorded invariant promotions: zero. The completion's §"THE CAPSULE LINE, ASSEMBLED (v6 → v15)" table is the status-bearing digest of exactly these entries; this appendix carries their raw text.

TOE_FINAL v7 (2026-06-11, session 2) — Gap 04's last number delivered. c_KK by the corpus's own AC-1 ζ-recipe: −8.892×10¹/R_Y⁴ (sign decision-grade); A = c_KK + κ₀′S(0) < 0 across the full variant band ⟹ the well exists; the corpus PASS structure is reproduced; the curved-share cap discharges pending countersign (ba0ba266d7315f71). Odd-D μ-independence proven (μ-[INPUT] dissolved). Self-caught #3 and #4 disclosed; five subtraction engines VOID with the impossibility theorem recorded; first-principles confirmations banked (R=30, Ric=5g, |Riem|²=276, Vol=π³/2). Remaining on Gap 04: c_loop (Chris/FRG-2), magnitude refinement (machine lane, non-blocking), countersign queue. Promotions to date: zero.

TOE_FINAL v8 (2026-06-12) — Countersigned; discharge certified; D15-native staged. Chris ruling logged: H04_DISCHARGED_AT_DECISION_GRADE — weakest accepted variant clears the operative bound by 35×, central by 2.4×10³. Canonical claim (verbatim): Gap 04's curved-share falsifier is passed at decision grade; the remaining magnitude refinement no longer threatens the sign or bound, and c_loop remains handled under the declared scaffold/FRG-2 lane. c_loop window frozen: 0 < c_loop < A²/(4c_bdry); F1/F3 armed. D15 re-targeted to the native scaffold (dispatch2/d15_native/): validators PASS, six dossiers ready, runner verified refusing at exit 2 until Chris signs seven slots (slot 0 = the re-target ruling; S6 = the parity [INPUT] confirmation). Flux-track D14 bridge unchanged, open in its own lane. Gap 08 locked until D15-native PASS. Promotions: zero.

TOE_FINAL v9 (2026-06-12) — Spine executed: Gap 04 CLOSED; Gap-08 σ survives round 1. Gap 04: CLOSED at decision grade (D15-native runner PASS under Chris's permission; package gap04_frozen_package_native.json; caveats inside; flux track moot-at-minimal per I1=G). Gap 08 round 1: ρ DEAD (qualitative grade, KK-stiff; noise disclosed; NT-1 transverse-stabilization residue registered); σ SURVIVES with the exact x-shape scan — Planck keeps x ≲ 0.15 and the slice predicts n_s ∈ [0.960, 0.964], r ≈ (2.9–3.6)×10⁻³ (LiteBIRD-testable, E2); conditional sharpening c_loop < 0.05·A²/c_bdry. Gap 12 shape-populated under Rule A (hash-traced); Gap 09 BLOCKED_ON_NAMED_INPUTS (c_bdry magnitude — AC-1.3 read first; κ₁₃; c_loop FRG-2). ERRATUM-K1 (handbook det-42 → 66 by identity) queued. Spine: I1 ✓ · I4 ✓ · I5 σ-surviving · I6 blocked-named · I7 populated. Countersign queue: ρ-death, σ-advancement, NT-1, ERRATUM-K1, window sharpening. Promotions: zero.

TOE_FINAL v10 (2026-06-12) — The boundary sign event: discharge unconditional; plateau predictions land. c_bdry = −1.08×10⁻² (negative, full band) — AC-1.9's β_∂ arithmetic finally run; fermion parity pair cancels exactly. Amendment (countersign): Gap 04's well now exists UNCONDITIONALLY in c_loop — the window is vacated, the closure strengthens (self-caught #5 on the prior b>0 attribution). Round-1 hilltop superseded; ρ-death stands. The corrected σ-regime is an exponential plateau with λ² = 4/K_σσ = 1/6 set by the 13D split n=(6,2,1): n_s ∈ [0.9643, 0.9679] — Planck dead-center, near-universal; r ∈ [3.5, 10]×10⁻³ — LiteBIRD 3.5–10σ. A_s relation frozen, absolute BLOCKED_BY_NAMED_OBJECT (κ₁₃ — now the spine's single remaining normalization input). Gap 09 CONDITIONAL (T_RH ∈ [2.4×10¹², 4×10¹⁵] GeV, mostly in-band; N_eff 3.044 in); Gap 12 v2 populated. Countersign queue: c_bdry result, amendment, supersession, plateau outputs, Gap-09 conditional. Promotions: zero.

TOE_FINAL v11 (2026-06-12) — κ₁₃ closed; the convention question goes to LiteBIRD. K13-1 CLOSED: κ₁₃² = 1/(π⁴M_U⁹M_Pl²) (closed form, corpus-verbatim relation, frozen radii; fold = factor-2 band). The A_s blocker decomposes: number closed; residual = K13-2, the σ-kinetic normalization — datum #7 recovered from the session record (D.1 exceeds standard reduction by exactly 81/22; D−2=11 vs D₄−2=2). Exact four-branch integration: all survive r<0.036; r-bands separate at LiteBIRD resolution → the convention becomes an observable (T6 [3.5,10]e−3 · Tu [6,17]e−3 · D6 [8,24]e−3 · Du [10,36]e−3). n_s robust across branches; operative branch remains the frozen T6. Gap-09/A_s re-gate on K13-2 + c_loop. Countersign queue: +κ₁₃ form, +branch table, +E2 second upgrade. Promotions: zero.

TOE_FINAL v12 (2026-06-12) — The supertrace verdict: Λ stays Weinberg-open, honestly. I2 RUN AND CLOSED AS HONEST FAIL. The exact-weight graded supertrace over the Paper-3 §4 nine-row ledger (ghosts/antifields formal; 17-row physical inventory confirmed) shows no structural cancellation at any coefficient order — ratio 0.58 at k=0, 1.000 at k=1–8. The chamber grading is coefficient-blind. Λ remains Weinberg-open for the Class-2 route; the programme's Λ line is now a computed one-loop certificate, not an open hunt. I3 inherits the spectral obstruction (L1–L3 must relocate the cancellation or be refuted). Str ρ at frozen radii = (−88.93±band)/R_Y⁴ + c_loop; K13-2 caveat carried; no observed-Λ comparison performed. Countersign queue: +I2 verdict, +ledger authority correction, +I3 update. Promotions: zero.

TOE_FINAL v13 (2026-06-12) — The spine's branches: preservation and a falsifier. Lane 05 (BG-10): BLOCKED_BY_NAMED_OBJECT, preserved Open/Diagnostic exactly as the corpus locks it ("Chris does the physics"; silent promotion = structural FAIL). Four-fix ledger 0/4 at certificate grade (ingredients frozen for the CP invariant). Epoch delivered as a band: thermal-leptogenesis floor PASSES the entire T_RH range; M_R threshold passes the upper band (diagnostic estimate, labeled). Lane 06 (Gap-11): CONDITIONAL — boundary-localized ℤ₂-odd freeze-in candidate; KK-overclosure gate PASS across the whole inflaton band; relation λ_p(0.12) ~ 10⁻¹¹ frozen; null-forecast: σ_SI ~ 10⁻⁵⁵–10⁻⁶⁰ cm² ⟹ a direct detection falsifies the minimal candidate (E6 becomes a falsifier). Blocked on the named portal normalization. One anchor slip self-caught and superseded. Countersign queue grows accordingly. Promotions: zero.

TOE_FINAL v14 (2026-06-12) — The Λ line concludes; the arrow settles; a halt routes honestly. I3 THEOREM_REFUTED: Lemma 2 fails structurally (the Λ operator is the unit operator — grading-even, label-blind) with I2's 0.58 residual as quantitative witness. The chamber route to Λ is closed; Weinberg stands; the Λ line's final internal state is a pair of computed certificates (supertrace FAIL + theorem refutation) where the programme once had an open mechanism hunt. Gap-07 PASS: Spohn's D(ρ||ρ_ss) is the surviving universal arrow monotone (machine-zero under generic + thermal GKLS); five expected failures named and counted. NT-1: χ-leg run gate-halted on a scale artifact — VOID both directions, routed to the machine lane; ρ-leg qualitative-stiff stands. Countersign packet grows to C19. Promotions: zero.

TOE_FINAL v15 (2026-06-12) — Evidence on the desk; not one ruling assumed. The D.1 source check is done: the kinetic denominator (D−2)=11 is verbatim at L10268; numerators match the standard reduction exactly; the 81/22 is confirmed from source, with the slip-vs-intent question honestly left as the one remaining inference. New from the source: D.1.0 declares uniform breathing — a TEXTBOOK ruling lands on Tu (r ∈ [6.1,17]×10⁻³, n_s Planck-centered), and T6-vs-Tu reduces to one look at the frozen σ-map yaml. C17's scope sentence is in the certificate; the COI checks stand alone; the c_loop prereg runner and BG-10 exit-2 checker mechanize the discipline. The Return YAML remains the only instrument of promotion. Promotions: zero.

Provenance note (Appendix P). The ten entries above (P.0 masthead, P.1 v6 closing, and the eight v7–v15 capsules) are verbatim-faithful to TOE_FINAL (2).md; the only edits are the markdown blockquote framing and the bolding of status tokens for readability — no word of any status statement is altered. The initial file's bytes remain untouched (MR-7); its sha256 at fork is recorded in the masthead provenance block at the top of this document.

MR-4 exception, declared (one, intrinsic to the verbatim text). The v10 capsule above contains the word "unconditional" ("discharge unconditional") at a point where the completeness checker's ±160-character claim-control window does not reach the licensing tokens — even though the same capsule licenses it three clauses later with "Amendment (countersign): Gap 04's well now exists UNCONDITIONALLY in c_loop" and "self-caught #5". In the completion's body (§II "The boundary-sign event and the Amendment"), this exact event is narrated with the license adjacent, and it passes the checker cleanly there. Here it is preserved verbatim because Appendix P's whole purpose is byte-faithful provenance recovery: rewording the capsule to satisfy the checker would defeat the appendix. This is therefore a single, deliberate, explained MR-4 exception — the word is licensed within its own capsule (by the Amendment/countersign it announces) and is fully licensed in the body — not a promotion introduced by this document. Every other claim-control word in this document is either adjacent to its license or filtered as prose.

Merged-build record (MR-7) — the reader-optimization passes

The front-matter provenance block summarizes these; the changelog is kept here, out of the reading path. Every pass held the completeness gate at exit 0 (zero new dangling refs, zero new stub sections, the MR-4 promotion-flag count unchanged at 1 — the pre-existing flag being inside verbatim Appendix P), changed no status / tier / headline number / scope, and left the two source SHAs (masthead) byte-untouched.

• Passes 1–3 (2026-06-13). Merged the two source statements and recovered the raw capsule provenance (Appendix P); added the one-line thesis, the BIG IDEA box, the §0 intuition pumps, the "smallest toy → real object" treatments (§I–§III), the STRONGEST OBJECTIONS steelman, the honest landing (§VIII), and the §IX engineering bridge at simulated grade; then added the cognitive-scaffolding layer (the by-time ladder, the afternoon-checks box, the isometries pump, the QC-ratio-gate gloss, the integrity-event framing). Backup: TOE_FINAL_merged.pre_pass3_2026-06-13.bak.
• Pass 4b — Track A surgical consistency repair (2026-06-13). Four located honesty-preserving fixes, no freelancing. (1) Thesis de-claim: the one-line thesis tail (top) and its by-time-ladder echo no longer read as an unqualified win-claim — "survived every test it has run" → "passed every test it could run, published in full the one place it failed" — bringing the thesis into agreement with the §III Λ verdict the document publishes at full strength (chamber route to Λ CLOSED/refuted; Λ itself OPEN/Weinberg-open). Honesty-improving, not a status change. (2) "stronger" → "deeper" consistency: the two front-matter instances (BIG IDEA, WHY THIS MATTERS) were brought into line with the §II de-claim already applied in the stabilization narrative ("the well deeper"), so all instances agree; a de-claim, never a promotion. (3) Forward-reference repair: the WHY-THIS-MATTERS pointer "(the five named below)" now resolves to its real home — "(the five named in HOW TO READ, Known weakest links, below)" — after the front-matter fold moved that list out of §V. (4) Capsule-assembly table relocation: DEFERRED — skipped per the completeness-ref-risk rule (the Appendix-P "status-bearing digest" cross-reference and the table's status-bearing density made relocation a non-surgical risk). Completeness gate held at exit 0 throughout (dangling_ref=0, stub_section=0; MR-4 promotion-flag count unchanged at 1 — the pre-existing flag inside verbatim Appendix P); no status, certificate tier, headline number, falsifier, or scope altered; both source SHAs (masthead) byte-untouched. Backup: TOE_FINAL_merged.pre_pass4b_2026-06-13.bak.
• Pass 4 — deep editorial cascade (2026-06-13). Structural subtraction + line edit. Consolidated the front-matter box pileup (folded the "state-of-the-programme ledger" banner into WHY THIS MATTERS, cut disclaimer boxes restated elsewhere, relocated the stacked pass-records to this record, trimmed the Reader map to its two audit rows, thinned redundant falsifier re-enumeration and the capsule-line preamble — front matter down ~440 words). Reconciled two clarity defects on the document's own five-minute dare: "two anchors → 22+" → "four anchors → 22+" (§0/§I, now consistent with the four-anchor claim everywhere), and the lede's self-caught count "five" → "four" (the record exhibits exactly #2–#5; no #1 exists, so none was invented). Recast the BIG IDEA lede onto the bet and the rubber-band image; glossed supertrace / Weinberg-open / L1–L3 (§III) and center-threading (§0); replaced the wrong carousel analogy (§0); reordered the §IV branch-table legend; matched the §IX header to its honest verb (DESIGNS AND SIMULATES); pinned the QC band to 10,000 logical qubits; added an interpretive landing after Table V-A and an audience signal on Table V-B; thinned the slogan runway before the unchanged kicker ("Every link is in §V. Pull."). No status, falsifier, number, or scope altered. Backup: TOE_FINAL_merged.pre_pass4_2026-06-13.bak.
• Pass 5 — DEEP flagship line edit + the Test it yourself feature (2026-06-13). Full structural + line edit on the general-reader front door. (1) The five-prompt feature.* Added the boxed Test it yourself — five prompts you can paste into any AI (or do by hand) subsection (HOW TO READ, immediately after the afternoon-checks box), with the five referee-approved canonical prompts embedded verbatim — *3 Type-A (charged-shell field energy; one-line anomaly cancellation 3·(1/6)−1/2=0; Newton/Coulomb weak-field reductions) and 2 Type-B (CKM CP phase; exactly-three generations) — each with its consistency-not-derivation / does-not-replace-GR framing kept. The pre-existing four-check box was merged (not dropped) into the box's index/lede, and the ~1-hour by-time-ladder rung was repointed to the new box. Embedding guards enforced: EXTERNAL/no-in-doc-GR-Maxwell disclosure inherited (G1); Prompt-4 Wolfenstein-vs-raw convention guardrail rendered adjacent + the false-185°-mismatch warning (G2); Prompt-1 Δm rounding-artifact inoculation (G3); internal cites re-keyed — the non-resolving "§126" replaced by "the afternoon-checks box (check 2) / the §0 q=6Y row", Prompt-5 routed to the GUT-sibling EXTERNAL, zero unresolved §NNN tokens added so dangling_ref stays 0 (G4); exactly five prompts, all physics/SM, no QC prompt per §VI Rule 4 / §IX license (G5). (2) δ_CKM convention import (LE-1, honesty-restoring, GRADE UNCHANGED — CLOSED stays CLOSED). The §I CKM-phase Statement cell, the §0 holonomy intuition pump, and Objection 1 now carry "δ_CKM = 60° (Wolfenstein-aligned; raw order-three holonomy −2π/3 = −120°), vs PDG γ = 65.5° at ~0.79σ" — the GUT sibling's enacted §7.6 ruling imported verbatim; this is the body convention Prompt 4 forces, no promotion. (3) Subtraction (restatement/relocation only, coverage falls nowhere). Cut the L43 prose section-inventory (covered by the Reader map + by-time ladder); cut the WHY-THIS-MATTERS opening italic that paraphrased the BIG IDEA a third time (the unchanged thesis → BIG IDEA → §25 now opens cleanly); compressed two "more-outputs-than-inputs" restatements (§0 pump tail, afternoon-check 1) to back-references; trimmed the §V seam echo; and cut the body four-semicolon slogan clause before the kicker so "Every link is in §V. Pull." lands by isolation (the verbatim v6 closing line in Appendix P, which carries the same slogan, is byte-untouched). (4) Glosses (additive, no status touched). Plain-reading gloss under the body capsule table defining "the well"/c_bdry/c_loop/c_KK before §II and before Objection 2; deferred the "blind-negative control" term of art in the BIG IDEA; glossed the triple BLOCKED tokens (object vs inputs); plain-language gloss of the dark-matter candidate's gate-names; de-claimed the §IX summary verb in WHY-THIS-MATTERS ("builds a quantum computer / working engineering artifact" → "designs and simulates" / "simulated engineering design", matching the §IX SIMULATED grade); "thirty named results" → "thirty named instruments" (matches Table V-B); one T²/U(1)² maximal-torus notation-reconciliation note (§0). (5) Bounce points. Sharpened the Table V-B generalist skip-anchor to §VI (relabel only, zero rows moved); reordered the §IX seven-item self-catch list to lead with its takeaway, items follow as tagged receipts (no cut); added a Prompt-5 forward pointer at the §I rubber-band takeaway. No status, certificate tier, headline number, falsifier, or scope altered anywhere; the 18/30/48 truth table, the four-terminal-state spine, all five Strongest Objections, and the QC band (9.6–14.5 / Robust 3.24, sub-2 retired-labeled) are byte-intact in structure. Completeness gate held at FINAL_EXIT=0 (dangling_ref=0, stub_section=0); the hard-fail MR-4 promotion-flag count is unchanged at 1 — the single pre-existing licensed flag inside verbatim Appendix P (L887 "unconditional"). Two residual WARN-severity soft-word occurrences ("derived", each a negation — "not a derived quantity" / "not derived from SM principles") sit inside the byte-frozen verbatim Prompt-4/Prompt-5 copy-paste text at char-depths (304, 185) unreachable by adjacent framing within the ±160 window; rewording them would break the verified, copy-paste-faithful prompt, so they are declared here as an extension of the verbatim-text exception (same class as the Appendix-P exception) — each is anti-promotion by construction and carries an adjacent licensed Honest-framing line grading it as consistency-not-derivation. Both source SHAs (masthead) byte-untouched. Backup: TOE_FINAL_merged.pre_pass5_2026-06-13.bak.

— End of the statement. Both source originals untouched; merge provenance in the header; the raw capsule provenance recovered in Appendix P; the engineering bridge in §IX; the truth table in §V; the queue in §VIII.