# AI CLOSURE COMPENDIUM — ARTICLE 1: GUT

**Subject:** The Fable scoped Grand Unified Theory (Article 1 of the four-paper set).
**Source manuscript:** `C:/Users/cberg/Desktop/VS Code Projects Random/physics_Journal_and_patents/Fable_Version/rendered/GUT/GUT.md`
**Authoritative per-gate map:** `C:/tmp/toe_closure/GATE_CLOSURE_MAP.md`
**Frozen branch:** `dcc66f1b2685` / `a5b1e6f9d951` — **READ-ONLY, unmutated.**
**Date:** 2026-06-24 · **No status was ever upgraded.**

---

## 0. Reading rules for this compendium (binding)

**Closure vocabulary — used EXACTLY, nothing else:**

| Token | Meaning |
|---|---|
| **DERIVED-CLOSED** | Follows by theorem / exact arithmetic from stated mathematics, independent of any selected input. |
| **DERIVED-GIVEN-E** | Derived *given the bundle E (the SM chiral content) and/or the SELECTED 13D geometry* as an input. The result is a genuine computation, but it consumes the selection. |
| **AXIOM-CLOSED:declared** | Closed only by adopting a named, openly-confessed posit; not derived. |
| **PARTIAL** | A real result that does not close the full requirement — a worked sub-result, a consistency check, or a certificate complete only under declared assumptions. |
| **OPEN** | Not derived and not reduced to any declared axiom. Genuine residual. |

**THE LOAD-BEARING DISTINCTION (governs every gate below):**

> **Selection ≠ derivation.** The GUT gates are derived **GIVEN the SELECTED geometry + E**. The 13D shape (𝔅, K₆=SU(3)/T², the Z₆ quotient, the orbifold, the three-layer ×/⊕/⊗ structure) and the bundle E are an **INPUT**, not an output of this article. *Why this geometry?* is a separate question, addressed by the minimal-shape suite, and is NOT claimed here. The manuscript itself encodes this as the data-use firewall (§4.4, §4.9, §1.3 point 1): a gate's *constraint-face* reads architecture only (which groups, how many families, no mirrors); its *certificate-face* reads frozen numbers. Passing the architectural filter never pre-pays the numerical test.

**Honest partials carried verbatim into every relevant gate:**
- Anomaly closes **only ~10/16 classes at certificate**; the 4D quantum (BV-BRST descent) anomaly is **AUDIT**, not proven.
- Flavor is **certificate-complete UNDER declared assumptions** with a **disclosed ~4.4σ m_u miss**.
- Unification is **consistency-only**: M_U ~ 10¹⁶ GeV is **frozen-imported** (RG-transport output of declared anchors), not independently predicted.
- Proton decay **τ_p is OPEN** — the proton-safety *operator identity* Π_q M Π_ℓ = 0 is **NOT a lifetime prediction**.

**Internal-vs-prompt gate numbering.** The manuscript's own Section-6 gate cards number gauge recovery as Gate 2, charge as Gate 3, chirality+generations as Gate 4, anomaly as Gate 5, stabilization Gate 6, thresholds Gate 7, Higgs Gate 8, flavor Gate 9, proton Gate 10. This compendium follows the **prompt's 9-gate spine** and cites the corresponding manuscript card for each.

---

## GATE 1 — Gauge group SU(3)×SU(2)×U(1) → SU(3)×U(1)_em

**(i) REQUIREMENT.** A GUT must recover exactly the three observed gauge forces — color SU(3)_c, weak SU(2)_L, hypercharge U(1)_Y — with no extra unobserved factor and no missing factor, and exhibit the electroweak breaking to SU(3)_c × U(1)_em.

**(ii) CLOSURE STATUS: DERIVED-GIVEN-E** (selection-grade).

**(iii) CLOSURE ARGUMENT + kernel.** The surviving 4D gauge algebra is read off as the **isometry algebra of the internal factors of the selected geometry**:
- K₆ = SU(3)/T² (the flag manifold) carries SU(3) exactly → routes color;
- S² carries SU(2)_L → routes the weak force;
- S¹_Y/Z₂ carries U(1)_Y → routes hypercharge.
The non-vacuous control is in the elimination table (§3.5): a bare circle gives one U(1) but every fermion keeps a mirror (killed by chirality); T⁶ gives six U(1)s, not the non-abelian SU(3) (killed by gauge recovery); the cheaper CP² = SU(3)/U(2) carries SU(3) but its family count is a *tunable bundle choice* (killed by the anti-fitting rule). Only K₆ × S² × S¹_Y/Z₂ survives. The teaching principle (Appendix GP, *Isometry*): **the symmetries of the internal shape become the forces of the 4D world.** Electroweak breaking to U(1)_em is delivered by the Wilson-line/Hosotani mechanism of Gate 5 (EWSB), with Q = T₃ + Y on every multiplet (see the Gate-6 hypercharge kernel).

**(iv) EXACT CAVEAT.** This is a derivation **GIVEN the SELECTED factor set** K₆ × S² × S¹_Y/Z₂. The factor set is itself a SHAPE posit — the elimination table is a *failed-search over an enumerated menu*, not a proof that no other shape works. So the gauge group is forced *given* the geometry; the geometry is not forced. **selection ≠ derivation.**

**(v) EVIDENCE.** GUT.md §6.2 (Gate-2 card), §3.5 (elimination table), Appendix D (Standard Model Recovery), R1.2/R1.4 isometry data.

---

## GATE 2 — Chiral fermion content (no surviving mirrors)

**(i) REQUIREMENT.** The world is chiral: left- and right-handed fermions transform differently, and no vectorlike "mirror" partners survive at observed energies. A GUT must produce a chiral spectrum and project mirrors out.

**(ii) CLOSURE STATUS: PARTIAL / DERIVED-GIVEN-E.** The classical mirror removal is derived given the geometry; the **4D quantum (BV-BRST) descent of the projection is AUDIT**, not proven.

**(iii) CLOSURE ARGUMENT + kernel.** The orbifold S¹_Y/Z₂ has two endpoints that impose handedness-filtering boundary conditions. The **Atiyah–Patodi–Singer one-sided boundary index** on S¹_Y/Z₂ returns (n_L, n_R) = (+3, 0): three left-handed survivors, zero right-handed mirrors, for the relevant bundles (frozen parity table `ac4d2df3e708`). This is the classical statement and it is clean. **What is NOT closed:** the quantum descent — the statement that the BV-BRST-quantized measure inherits the same projection without a residual gauge/anomaly obstruction — is carried at AUDIT tier (this is the UQF-7 "quantized BV-BRST descent measure" open object in the TOE ledger; classical layer certificate-closed, quantum descent measure genuinely-open).

**(iv) EXACT CAVEAT.** Derived **GIVEN the SELECTED geometry + E**: the index is computed *with* the bundle E as input, so it certifies that the chosen shape yields a chiral spectrum — it does not force E from a deeper principle, and the quantum-descent half stays AUDIT. Asserting the projection survives quantization is an honest open audit, not a proven theorem.

**(v) EVIDENCE.** GUT.md §6.4 (Gate-4 card), §3.5 (orbifold row), Appendix E (chirality half), orbifold freeze R1.3 `ac4d2df3e708`.

---

## GATE 3 — Three generations (χ = −3)

**(i) REQUIREMENT.** Exactly three families of matter — not two, not four. A GUT must deliver the family count as a fixed quantity, not a fitted dial.

**(ii) CLOSURE STATUS: DERIVED-GIVEN-E** (the count is an integer topological invariant, not a dial).

**(iii) CLOSURE ARGUMENT + hand-checkable kernel.** The family count is the **spin-ℂ Borel–Weil–Bott index of the relevant bundle on K₆ = SU(3)/T²**, and that index equals **−3**. The magnitude 3 is the number of generations; the sign encodes chirality. The load-bearing property: an **index is a topological integer** — deform the shape smoothly and the count cannot move (Atiyah–Singer). There is **no dial**: unlike CP² (whose family count is a tunable bundle choice and is therefore eliminated under the anti-fitting rule), K₆'s index is rigid. This is the cleanest "forcedness" claim in the article: χ(K₆, E) = −3 is an integer, not a parameter.

**(iv) EXACT CAVEAT.** The index is computed **WITH E as input** (DERIVED-GIVEN-E). Therefore it **cannot circularly force E** — it certifies that *this* geometry+bundle yields three families; it does not prove three is the unique possibility across all geometries, nor is the family number itself fixed by anomaly cancellation (a separate gate). The integer is rigid *given* the selected shape.

**(v) EVIDENCE.** GUT.md §6.4 (Gate-4 card, "Borel–Weil–Bott family index on K₆ returning −3"), §3.5 (flag-manifold row: "returns −3: three families, as a topological count with no dial to change it"), Appendix E.

---

## GATE 4 — Anomaly cancellation

**(i) REQUIREMENT.** Every gauge and gauge–gravity anomaly must cancel exactly on the surviving fermion content; a nonzero trace makes the theory quantum-inconsistent (probabilities that do not sum to 1). Six independent ledgers must vanish: SU(3)³, SU(2)³ (Witten parity), U(1)_Y³, the two mixed traces, and gauge–gravity.

**(ii) CLOSURE STATUS: PARTIAL / AUDIT.** A hand-witness and the perturbative six-ledger check pass; but **only ~10/16 anomaly classes close at certificate**, and the **4D quantum descent is AUDIT**.

**(iii) CLOSURE ARGUMENT + hand-checkable kernels.**

*Front-matter one-line witness (the SU(2)²–U(1) / mixed ledger):*
> **3·(1/6) − 1/2 = 0.**
Three colors of the quark doublet at hypercharge Y = +1/6, summing to +1/2, exactly cancel the lepton doublet at Y = −1/2. This is the front-matter "check it by hand in five minutes" rung (GUT.md line 42).

*The Y / Y³ ledger (one SM generation, left-handed basis, §3.6.2 — verifiable with pencil and paper):*

| Particle (LH basis) | Components | Y | Σ Y | Σ Y³ |
|---|:---:|:---:|:---:|:---:|
| Quark doublet Q_L | 3×2 = 6 | +1/6 | +1 | +1/36 |
| Up singlet u_R^c | 3 | −2/3 | −2 | −32/36 |
| Down singlet d_R^c | 3 | +1/3 | +1 | +4/36 |
| Lepton doublet L_L | 2 | −1/2 | −1 | −9/36 |
| Electron singlet e_R^c | 1 | +1 | +1 | +36/36 |
| **Total** | | | **0** | **(1−32+4−9+36)/36 = 0** |

Both balance by exact rational arithmetic. The **non-vacuous control**: Σ Y² = 10/3 ≠ 0 — not *any* combination cancels, only the six consistency-critical ones (and the cubic vanishes only in the consistent same-handedness bookkeeping; mixed conventions give a −4/9 trap). The doublet count per family is 3+1 = 4, even → Witten condition holds.

*What is NOT closed.* The hand-witness and the perturbative six-ledger trace are a **FILTER on E**, not a determiner of E (anomaly cancellation has infinitely many solutions in general — Allanach 2111.04148 — so it cannot single out the SM). And the full *quantum* anomaly accounting closes **only ~10/16 classes at the certificate level**; the BV-BRST 4D descent (and the coset-twist descended anomaly, UQF-4) remain **AUDIT/OPEN**. The generalized / center-anomaly framework invoked for these descended (non-perturbative) classes is **GKSW (Gaiotto–Kapustin–Seiberg–Willett, arXiv:1412.5148)**; GKKS (1703.00501) is deliberately **not** attached here — there is no θ-angle / time-reversal data in the GUT descent rows, so that framework stays absent. The certificate G05 verifies the rational sums on the frozen spectrum with a negative control, but the spectrum it consumes is Gate-4's (conditional claim).

*Provenance of the "16-class / descent" denominator (read this before grepping GUT.md).* The **six-ledger certificate is GUT.md's own** (§6.5 / Appendix E′, "Claimed certificate pass" — GUT.md lines 1167/2605): SU(3)³, SU(2)³ Witten, U(1)_Y³, the two mixed traces, and grav²–U(1)_Y. The **10/16-class denominator and the 4D BV-BRST descent AUDIT accounting are NOT GUT.md objects** — they are imported from the deeper **TOE-level quantum-anomaly ledger** (UQF-4 coset-twist descended anomaly = OPEN; UQF-7 quantized BV-BRST descent measure = quantum-descent-OPEN; `GATE_CLOSURE_MAP` lines 138-139). They are surfaced here so the gate is **not rested on the GUT.md perturbative pass**. A reviewer grepping GUT.md will correctly find **six** ledgers, not sixteen; the sixteen is the TOE quantum object, deliberately the more conservative anchor.

**(iv) EXACT CAVEAT.** Anomaly cancellation here is **inherited**, not an independent miracle: the geometry's index output (Gate 3) is one SM generation per family, and the SM's own cancellation comes along. It is a **FILTER on E given the SELECTED geometry**, not a derivation of E, and ~6/16 quantum classes plus the descent measure are not at certificate. **selection ≠ derivation; AUDIT ≠ PROVEN.**

**(v) EVIDENCE.** GUT.md §3.2–3.7 (worked deep example), §3.6.2 (the table), front-matter line 42 (3·(1/6)−1/2=0), §6.5 (Gate-5 card), Appendix E′ (full six-ledger trace, conventions A(R̄)=−A(R), T(fund)=1/2), certificate `certificates/G05_anomaly_cancellation/`.

---

## GATE 5 — Electroweak symmetry breaking / Higgs

**(i) REQUIREMENT.** A scalar must break SU(2)_L × U(1)_Y → U(1)_em, set the electroweak scale v ≈ 246 GeV, deliver the Higgs mass m_h ≈ 125 GeV, keep the Higgs light against the high scale by a *structural* mechanism (not a tuned counterterm), and preserve custodial symmetry (ρ ≈ 1).

**(ii) CLOSURE STATUS: PARTIAL.** EWSB mechanism + two near-hits are real; the hierarchy and ρ ≈ 1 are **not derived**.

**(iii) CLOSURE ARGUMENT + kernel.** The Higgs is a **Wilson-line / Hosotani mode** on a named cycle of K_gauge, one doublet n_H = 1, protected by an **integer winding count** (an integer cannot be continuously tuned to zero → structural protection). Outputs, regenerated post-RG from the R1.8 inputs + geometry:
- v_pred = **246.02 ± 3.5 GeV** (0.06σ post-freeze match to PDG);
- m_h = **123.82 ± 1.8 GeV** (0.48σ — a near-hit; **read with the caveat that the quartic is entangled with the OPEN 13D β-vector / chamber data** — the 0.48σ figure is NOT free-standing), via
  m_h² = (d²V_Hos/dθ_H²)|_{θ_H*} / (2πR_γ)², with the Hosotani potential's n⁻⁵ tail converging absolutely (the structural reason m_h is finite, set at the EW scale by the √η_BK/(2π) ratio rather than at M_Pl).

**What is NOT closed:**
- The **hierarchy** v/M_Pl: the Hosotani protection (Operator A) demonstrably removes the δm_H² ~ M_Pl² counterterm and is cutoff-independent, but protects **only up to ~10¹⁴ GeV — twelve orders short of M_Pl**. The smallness of v itself is read out from θ_H*, not derived (v_EW is the disputed *second* dimensionful anchor; "exactly one anchor" was retracted).
- **ρ ≈ 1 / custodial symmetry is not derived** in this article.
- The quartic is entangled with the OPEN 13D β-vector / chamber data.

**(iv) EXACT CAVEAT.** v and m_h are genuine **post-freeze near-hits GIVEN the SELECTED geometry**; they are NOT a solution to the hierarchy problem, and ρ ≈ 1 is assumed not derived. The matches are real; the *smallness* is undelivered.

**(v) EVIDENCE.** GUT.md §6.8 (Gate-8 card: v=246.02, m_h=123.82, η_BK=0.009721281516312024), §6.6 (Gate-6 stabilization), Appendix H (Higgs protection), Appendix I.2.1.

---

## GATE 6 — Yukawas / CKM / PMNS / CP

**(i) REQUIREMENT.** Reproduce the structured pattern of quark and charged-lepton masses, the CKM mixing matrix and its CP phase (Jarlskog), the PMNS angles and δ_CP — from frozen operators, NOT per-observable fits.

**(ii) CLOSURE STATUS: PARTIAL** — manuscript label *"Certificate-complete under declared assumptions."* Over-determined, but with a disclosed ~4.4σ m_u miss and IMPORT-dependent assumptions.

**(iii) CLOSURE ARGUMENT + hand-checkable kernel (the over-determination count).** The F⁺ flavor chamber supplies frozen operators O_u, O_d, O_e, O_ν and frozen Yukawa maps. **Exactly two numerical anchors are declared and counted in the ledger:**
- y_t(M_Z) = 0.9665 fixes the up-sector normalization N_u (hash `548d7099ef18`);
- |V_us| = 0.22436 fixes the chamber angle θ_F (hash `a1bc510bc7cd`).

**There is no per-family anchor — but there is one absolute-scale normalization per sector.** Against the two declared anchors {y_t, |V_us|}, the chamber returns **19+ independent frozen outputs**: m_t/m_c, m_c/m_u, m_b/m_s, m_s/m_d, |y_t/y_b|, the six off-diagonal |V_ij|, δ_CKM/J_CKM, m_e/m_μ/m_τ, and the neutrino Δm² and angles. CKM is the **misalignment V_CKM = U_u† U_d** of two frozen diagonalizations (not an inserted unitary); the CP phase is read from frozen holonomy. Most pulls are <1σ (charged leptons 0.01–0.07σ); ~2280 PDG citations back the comparison. **Honest input ledger (Gate-9 finding 2026-06-24):** what is genuinely derived, with *no* per-family knob, is every within-sector mass *ratio* and all the mixings, phases, and J — that is the real over-determination. What each sector still requires is *one absolute mass-scale*: N_d, N_e, N_ν are calibration normalizations pinned to m_b, m_τ, and the neutrino splitting respectively (they are **not** derived from N_u — no such closed form exists), and the seesaw scale M_R is asserted but not yet computed. So the absolute masses m_b, m_τ, and the neutrino splitting are anchor-consistency checks, not independent outputs. Counting these effective inputs, the honest economy is ~22 outputs from ~5–6 inputs ≈ **4× (3.7–4.4×)** — real over-determination on the ratios and mixings, more modest than a bare "2-in/19+-out" reading.

**The disclosed weakest link (carried verbatim):** the **m_u row at ~4.4σ** — a rigid-ladder prediction with **no m_u anchor**. This is registered, not hidden.

**(iv) EXACT CAVEAT.** This is **VALIDATION / over-determination GIVEN E and the SELECTED geometry**, not a from-deeper-principle derivation; the manuscript's own status is **"Certificate-complete UNDER declared assumptions."** The 4.4σ m_u miss is real and disclosed. A reduction attempt for |V_us| (deriving it from K₆ coset overlap) is **OPEN/NOT_DERIVABLE** — |V_us| stays one of the declared anchors (χ=−3 is deformation-invariant and carries no overlap data; the blind 1/√3 guess lands z=+552). If a flavor output were used as an input, or the chamber reopened after comparison, Gate 9 downgrades.

**(v) EVIDENCE.** GUT.md §6.9 (Gate-9 card), §7.3–7.5 (anchor fixing, frozen output ledger, anti-fitting), §7.4 (the 2-vs-19+ table; the "m_u row at ~4.4σ" disclosure), Appendices I/J/K, certificate `certificates/G09_flavor/`.

---

## GATE 7 — Neutrino masses

**(i) REQUIREMENT.** Account for the small neutrino masses and the PMNS mixing — including the mass-squared splittings Δm²_21, Δm²_32 and the three mixing angles — via a stated mechanism.

**(ii) CLOSURE STATUS: PARTIAL** (subsumed under the flavor certificate; some neutrino quantities labeled *Pending — not used in claim*).

**(iii) CLOSURE ARGUMENT + kernel.** A neutrino chamber operator **O_ν** plus a **Type-I seesaw** yields the neutrino *mixing* quantities and mass-squared *ratio* as frozen outputs: sin²θ_12, sin²θ_13, sin²θ_23, δ_CP^ℓ, and Δm²_21/Δm²_32 — counted among the 19+ outputs of Gate 6, with **no per-family neutrino knob**. The *absolute* splitting scale, however, requires one sector normalization (N_ν) fixed to the measured Δm², and the seesaw scale **M_R is asserted-derived but not yet computed (open)** — so the absolute neutrino splitting is an anchor-consistency check, not an independent prediction, and N_ν is effectively an additional input (Gate-9 finding 2026-06-24).

**What is NOT closed:** the seesaw scale relation **M_R = κ·M_U** is a **CANDIDATE IN TENSION** (~231× off the corpus κ⁰ baseline; internally inconsistent κ⁰ vs κ¹) and is **OPEN**. Any neutrino quantity not certificate-complete is explicitly labeled **"Pending — not used in claim"** in Appendix K and does NOT support gate closure. Absolute neutrino masses / mass ordering are not pinned.

**(iv) EXACT CAVEAT.** The seesaw construction returns 8 neutrino outputs **GIVEN the SELECTED geometry + E and the declared seesaw assumption**, but the scale-setting M_R is OPEN and several rows are honestly fenced as Pending. The mechanism is stated; the closure is partial.

**(v) EVIDENCE.** GUT.md §6.9 / §7.4 (neutrino output row), Appendix K (per-row status labels, "Pending — not used in claim"); GATE_CLOSURE_MAP M_R = κ·M_U line (CANDIDATE IN TENSION).

---

## GATE 8 — Gauge-coupling unification

**(i) REQUIREMENT.** The three running gauge couplings α_1, α_2, α_3 must meet at a single high scale M_U once finite threshold corrections are included — a genuine unification, not an accident of free knobs.

**(ii) CLOSURE STATUS: PARTIAL / consistency-only.** M_U is **frozen-imported**; the gate demonstrates *consistency*, not an independent prediction of the unification scale.

**(iii) CLOSURE ARGUMENT + kernel.** Inputs: α_i⁻¹(M_Z) (PDG, the comparison/unification *target*, entered as ONE target not generative data), M_Pl normalization, two-loop MS̄ SM beta functions, the KK spectrum of K_gauge under the declared regulator, and the **threshold vector (δ_1, δ_2, δ_3) = (+4.8424, −3.1112, −1.7313)**. Output: unification at **M_U ~ 10¹⁶ GeV** with a residual at the numerical-pipeline floor (**~10⁻⁸ manuscript-body / 9.6×10⁻¹¹ R1-manifest** — the EXACT_GEOMETRY_ANCHOR and GUT.md §6.7 card report ~10⁻⁸ as the body numerical floor, with 9.6×10⁻¹¹ the tighter owner-source / R1.8 closure-target value; both cited per the anchor's dual-citation discipline), inside the PDG band (~10⁻³). R₀ = (2πM_U)⁻¹.

**What is NOT closed:** M_U, R₀, and the threshold δ are **outputs of the declared anchors transported by RG** — they are AXIOM-CLOSED:declared at the root (reduce to the 4-anchor vector), and the unification gate is therefore **consistency-only**: it confirms the couplings *can* be brought to meet under the frozen thresholds, not that the scale is derived from deeper structure. The reproduction harness is **AUDIT pending the heat-kernel CSV mount**.

**(iv) EXACT CAVEAT.** Unification here is **consistency-only with M_U frozen-imported GIVEN the SELECTED geometry and the declared anchors**. It is not an independent prediction of the unification scale; the residual is a self-consistency floor, and the thresholds are frozen (no-hidden-knob audit at G.10). Higher-loop transport would shift the prediction within the published band.

**(v) EVIDENCE.** GUT.md §6.7 (Gate-7 card: δ vector, M_U ~ 10¹⁶, residual ~10⁻⁸ inside ~10⁻³), Appendix G (threshold ledger, G.10 no-hidden-knob audit), R1.8 anchor hashes.

---

## GATE 9 — Proton decay (τ_p)

**(i) REQUIREMENT.** A GUT must not predict a proton lifetime below the Super-Kamiokande bound (τ_p ≳ 10³⁴ yr in the cleanest channels). The historical execution: minimal SU(5) predicted ~10³⁰ yr and the water tanks killed it. The candidate must name its dangerous operator class out loud and show it absent, suppressed, or bounded.

**(ii) CLOSURE STATUS: OPEN** (no numerical lifetime). The manuscript splits the card: **operator level = "Claimed certificate pass"; lifetime = "Diagnostic only."** Per the closure vocabulary, **τ_p is OPEN** — the operator identity is NOT a lifetime prediction.

**(iii) CLOSURE ARGUMENT + kernel.** The E_proton projector layer carries the **sector-projector identity**
> **Π_q M Π_ℓ = 0,**
i.e. the frozen sector projectors Π_q (quark sector) and Π_ℓ (lepton sector) annihilate the mass operator between them — there is **no mediator** coupling quark and lepton sectors (the FCNC / mediator no-go theorem, L.3). Structurally, K_gauge has **no embedding into a single simple GUT group with heavy X/Y mediators**, so the dangerous dimension-six baryon-and-lepton-number-violating operators that killed SU(5) are removed by selection rule on the frozen labeling. The certificate `G10_proton_safety/` verifies this exact projector identity.

**What is OPEN:** there is **no numerical τ_p prediction**. The lifetime is **deliberately not forced** — reported above the Super-K bound and **excluded from the closure claim** (L.4, "Diagnostic only"). A pre-registered falsifier exists (proton-decay observation), observation-blocked until ~2035.

**(iv) EXACT CAVEAT.** The operator-level no-mediator result is **DERIVED-GIVEN-E at the operator level** (Π_q M Π_ℓ = 0 on the frozen labeling); it is a *safety* statement, **NOT a lifetime prediction**. **τ_p is OPEN.** A closure claim resting on the numerical lifetime would itself be a failure mode by the card's own definition. The falsifier is registered, not resolved.

**(v) EVIDENCE.** GUT.md §6.10 (Gate-10 card: "Claimed certificate pass (operator level); Diagnostic only (lifetime)"), §5.9, Appendix L (L.2a declaration, L.2b coverage, L.3 no-mediator theorem, L.4 diagnostic), FCNC no-go hash `fff4b433b7b3`, op-class `551488d06011`, certificate `certificates/G10_proton_safety/`.

---

## ADDITIONAL HAND-CHECKABLE KERNEL — the GR/EM reduction (limit check)

Carried because the prompt's discipline rewards the hand-checkable rungs and reviewers attack the "never reduces to known physics" objection. The ×/⊕/⊗ ledger reproduces the **Einstein–Maxwell field-energy object-for-object**: U_EM,12 = **4.49378 × 10⁸ J** to six significant figures (Appendix O.4/O.7), and explicitly does **not** claim to replace GR. Status: a consistency reduction (PARTIAL — limit check), not a closure gate, but it is the second front-matter "check by hand" rung alongside 3·(1/6)−1/2=0.

---

## PER-GATE STATUS TALLY

| # | Gate | Closure status (this compendium's verdict) |
|---|---|---|
| 1 | Gauge group → SU(3)×U(1)_em | **DERIVED-GIVEN-E** (selection-grade) |
| 2 | Chiral fermion content / no mirrors | **PARTIAL / DERIVED-GIVEN-E** (classical derived; quantum descent AUDIT) |
| 3 | Three generations χ = −3 | **DERIVED-GIVEN-E** (rigid integer index) |
| 4 | Anomaly cancellation | **PARTIAL / AUDIT** (hand-witness + ~10/16 classes; descent AUDIT) |
| 5 | EWSB / Higgs | **PARTIAL** (v, m_h near-hits; hierarchy & ρ≈1 undelivered) |
| 6 | Yukawas / CKM / PMNS / CP | **PARTIAL** ("cert-complete under declared assumptions"; ~4.4σ m_u) |
| 7 | Neutrino masses | **PARTIAL** (O_ν + Type-I seesaw → 8 outputs; M_R OPEN) |
| 8 | Gauge-coupling unification | **PARTIAL / consistency-only** (M_U frozen-imported) |
| 9 | Proton decay τ_p | **OPEN** (operator no-mediator ≠ lifetime) |

**Counts:** DERIVED-GIVEN-E (clean) = 2 (Gates 1, 3) · PARTIAL (incl. PARTIAL/DERIVED-GIVEN-E and PARTIAL/AUDIT and PARTIAL/consistency-only) = 6 (Gates 2, 4, 5, 6, 7, 8) · OPEN = 1 (Gate 9). **No status was ever upgraded.**

---

## HONEST READ

This article does **not** close the GUT problem; it **conditionally closes a derivable spine GIVEN a selected 13D geometry and the SM bundle E**, and is scrupulously honest about where that spine stops. The two genuinely clean results are **Gate 3 (three generations as the rigid topological integer χ = −3)** and **Gate 1 (the gauge group as isometries)** — both DERIVED-GIVEN-E, meaning real computations that nonetheless *consume the geometry as input rather than deriving it*. The single most important discriminator for any reviewer is **selection ≠ derivation**: the shape and E are inputs here, addressed (if at all) only by the separate minimal-shape suite; nothing in Article 1 forces the geometry.

Everything else is honestly partial: anomaly cancellation has a beautiful pencil-and-paper kernel (3·(1/6)−1/2=0 and the Y/Y³ table) but is a **filter on E, not a determiner of E**, and only ~10/16 quantum classes reach certificate (descent AUDIT). Flavor is the load-bearing "is this fitting?" gate and survives by **over-determination (2 anchors → 19+ outputs)**, but only "certificate-complete **under declared assumptions**," with a frankly disclosed **~4.4σ m_u miss**. EWSB delivers v and m_h as **near-hits** but does **not** solve the hierarchy (protection runs out 12 orders short) and does not derive ρ≈1. Unification is **consistency-only** with M_U frozen-imported. And **proton decay is OPEN** — the elegant no-mediator identity Π_q M Π_ℓ = 0 is a safety statement that is explicitly *not* a lifetime prediction, with the falsifier pre-registered and observation-blocked.

The article will survive blind adversarial review **precisely because it pre-routes its own weakest links** (the m_u 4.4σ, the AUDIT descent, the OPEN τ_p, the consistency-only unification) to the top and never relabels a PARTIAL/OPEN as closed. The correct one-line summary: **a strongly over-determined, geometry-conditioned reconstruction of the SM spine — DERIVED-GIVEN-E where it computes, PARTIAL where it validates, OPEN where it cannot (proton lifetime) — with zero promotions and the selection-vs-derivation firewall held throughout.**