SOURCE: https://physics.magflowmeters.com/gates/dossiers/uqf5c.html
======================================================================

UQF-5C — UV completion (shared) — dossier & ledger 

 ← Gates scoreboard · Jump to closure ledger 

 Gate dossier — UQF-5C — UV completion (shared)

 Question: Can this shape give a full quantum theory of gravity? 
 Status (fixed): REDUCED-TO-AXIOM · ANCHORED +1 
 Full 13-dimensional treatment: the frozen arena M4 x K6=SU(3)/T2 x S2 x S1_Y/Z2, all three layers, full precision. Self-contained — every value derived or stated inline. 

 Required endpoint (2026-07-06)

 Status: CLOSED / REDUCED-TO-AXIOM .

 Nothing left. Anchored on: 

 Shape: the primary root — the frozen internal shape (K6 × S2 × a folded hypercharge circle) supplies the graviton operator and every curvature number the gravity sector reads off (background selected, not proven the only choice)

 Granularity: dissolves a whole class of runaway high-energy infinities toward the shared wall, but gives no finite-grain shortcut to the completion — the minimal-length scale here is a pure-color object (~194 Planck lengths), so dissolving the infinities is not the same as building the theory

 Scale: the load-bearing root — the wall lives exactly in the strong-coupling regime above the energy where the usual expansion stops converging

 Observables: None consumed as calibration inputs; the gate is structural. Reproduced geometric invariants (target-blind): curvature ratio |Riem|2/R2 = 23/75; curvature constant κ = 1/6; First-Bianchi residual ~2.5×10−16 (machine zero); exact sphere spectra (S2=4/315, S4=74/63, S6=1139/63, conformal=5/63); color factor 124/315 (scale-free, pending independent reproduction). Explicitly refuted, not banked: the dimensionful magnitude −2.818×1094 GeV6 (scheme-contaminated and ill-posed at this dimension).

 Dissolution: No hidden derivation is claimed. The residual bottoms on the named value-free axiom/common-currency rule rather than an unbounded obligation.

 This is the current gate-level endpoint. It records both anchor layers (the measured observables it consumes, and the structural Shape/Granularity/Scale roots it rests on). The detailed dossier below is retained in full.

 Executive summary & honest status

 Headline (the sentence a skimmer must remember): on the frozen thirteen-dimensional arena \(\mathcal{M}_4\times K_6\times S^2\times S^1_Y/\mathbb{Z}_2\) ( \(K_6=SU(3)/T^2\) , \(D=4+6+2+1=13\) ), the geometry cleanly supplies the interacting-graviton kinetic operator and its BRST-forced ghost-corrected fiber weight, proves one exact curvature ratio by three independent target-blind routes, and certifies — as a structural theorem, not a confession of laziness — that no single heat-kernel coefficient can ever by itself be a non-perturbative UV completion of quantum gravity. The gate is graded REDUCED-TO-AXIOM / ANCHORED +1 , resting on the named conditional axiom pair \(\{\Delta_0>0,\ \text{Lorentz-scalar proper-time floor}\}\) . The constructive strong-coupling completion of the interacting graviton itself remains a global wall shared by every serious approach to quantum gravity, this one included , and is not claimed closed here. This grade is fixed by mandate (PROMOTIONS:0) and is written exactly as given below — neither upgraded nor softened.

 This is the graviton leg of what the corpus calls UQF-5C — "UV completion (shared)." The gate's question, in its sharpest public form, is: can this one frozen shape deliver a full, non-perturbative quantum theory of gravity — a completion of the fully interacting graviton that stays consistent at and above the cutoff, in the regime where the perturbative expansion in \(G_NE^2\) stops converging? That is the step from "the free/linearized graviton is a well-posed mode of this geometry" (true, and shown in full below) to "the interacting quantum theory of gravity exists and is UV-consistent on this geometry" (not shown here, not shown by anyone, not claimed).

 The precise claim

 Five pieces of banked, terminal content are established and carried forward as wins, each pinned to a definite status and to the full three-layer object that supports it.

 1. Operator-supply (leg 5A) is DERIVED-GIVEN-E and terminal as structure. Fix the × Stage as the complete frozen arena \(\mathcal{M}_4\times K_6\times S^2\times S^1_Y/\mathbb{Z}_2\) with \(K_6=SU(3)/T^2\) the full \(A_2\) flag manifold ( \(\dim=6\) ), \(S^2\) round ( \(\dim=2\) ), \(S^1_Y/\mathbb{Z}_2\) the active orbifold ( \(\dim=1\) ), \(\mathcal{M}_4=\mathbb{R}^{3,1}\) Minkowski ( \(\dim=4\) ); total metric dimension \(D=13\) . Fix the ⊕ Rulebook as de-Donder (harmonic) gauge-fixing of the metric fluctuation \(h_{MN}\) , the Faddeev–Popov ghost sector that this gauge choice forces, the \(\overline{\mathrm{MS}}\) /heat-kernel scheme, and \(\mathbb{Z}_2\) orbifold parity on \(S^1_Y/\mathbb{Z}_2\) ( \(\theta\mapsto-\theta\) , fixed points \(\theta=0,\pi\) ). Fix the ⊗ Actors as the resulting Lichnerowicz-type Laplace operator
$$
L_{\rm grav}=-(\nabla^2+E)
$$
acting on the graviton \(\mathrm{Sym}^2(T)\) bundle together with the FP ghost bundle, with \(\nabla\) the Levi-Civita (Nomizu) connection and \(E\) the Weitzenböck endomorphism built from Ricci and Riemann. On \(K_6\) at the Einstein center this operator's TT (transverse-traceless) sector has the certified Lichnerowicz spectrum \(\{1/6\ (\times6),\ 5/12\ (\times6),\ 7/6\ (\times6),\ 17/12\ (\times2)\}\) in Killing-normalization units, with \(\mathrm{tr}\,E_L=40/3\) and \(\mathrm{tr}\,E_L^2=241/18\) . The 4D massless spin-2 zero mode is the graviton; the Kaluza–Klein tower sits above it, indexed by the \(K_6\) Peter–Weyl decomposition \(L^2(K_6,E_\mu)=\bigoplus_{(p,q)}V_{(p,q)}\otimes\mathrm{Hom}_{T^2}(V_{(p,q)},E_\mu)\) . The allowed claim is exactly this: the geometry supplies the operator whose spectrum a UV completion would need to control. It is forbidden to say this operator "certifies the quantum theory," "derives General Relativity," or "derives \(E\) " — \(E\) is a given endomorphism, fixed by the matter content and background curvature, not something this leg derives from first principles. "Given- \(E\) " is the single largest charged input in this leg and is stated as given, never as derived.

 2. The ghost-corrected fiber weight is fixed and BRST-forced, not chosen. The graviton bulk fiber dimension in \(D=13\) is \(\dim\mathrm{Sym}^2(\mathbb{R}^{13})=\tfrac{13\cdot14}{2}=91\) ; the Faddeev–Popov ghost fiber dimension is \(13\) (one ghost per diffeomorphism parameter), entering the graded supertrace with multiplicity \(-2\) per the standard BRST accounting. The physical, ghost-corrected weight is
$$
91-2\cdot13=65,
$$
and \(65=\dim\mathrm{Sym}^2_0(SO(11))=\frac{11\cdot12}{2}-1\) — exactly the count of physical massless graviton degrees of freedom in \(D=13\) once the little-group trace mode is removed. That the ghost subtraction lands exactly on the little-group-correct physical count ( \(65=91-26\) ) is a nontrivial cross-check, and the sign and multiplicity of the subtraction are forced by BRST nilpotency, not a free normalization chosen to manufacture a nice number. This must never be confused with a structurally different, graded object built from \(\gamma=\mathrm{Sym}^2(\mathrm{diag}(\mathbf{1}_{12},-1))\) , which gives \(\mathrm{tr}\,\gamma_{\rm grav}=67\) , \(\mathrm{tr}\,\gamma_{\rm ghost}=11\) , and a Block-A graded weight \(67-2\cdot11=45\) — this second triple belongs exclusively to the \(\mathbb{Z}_2\) -defect grading and is never substituted for the bulk \(91/13/65\) triple.

 3. The scope firewall is itself a certificate, not a hedge. "One heat-kernel coefficient \(a_6\) is not a UV completion" is a proved structural statement , grounded in the convention \(K(t)\sim(4\pi t)^{-d/2}\sum_k a_{2k}t^k\) : the Seeley–DeWitt/heat-kernel expansion produces an unbounded ladder of coefficients \(a_6<a_8<a_{10}<\cdots\) , and a non-renormalizable, strongly-coupled theory is precisely one in which no finite truncation of that ladder controls the high-energy behavior. This is why the dossier states, as a win and not an apology, that computing \(a_6\) — however exactly — can never by itself settle strong-coupling consistency. It is the honest, structural reason a clean firewall exists between "the operator is supplied" (leg 5A/5B) and "the theory is UV-complete" (the 5C roll-up, which equals the shared UQF-9 constructive-completion question).

 4. A wrong dimensionful magnitude was refuted at decision grade. An earlier bulk numerical claim of \(-2.818\times10^{94}\,\mathrm{GeV}^6\) for the graded \(a_6\) functional was tested and refuted: the object is R2-contaminated, scheme-anchored, and — more fundamentally — ill-posed at odd \(D=13\) , where the relevant local heat-kernel slot sits at a half-integer zeta pole ( \(s=7/2\) ) that vanishes identically in dimensional regularization, so no finite local \(t^0\) coefficient of that mass dimension exists to anchor a \(\mathrm{GeV}^6\) number in the first place. This is a reached verdict — completed science — not an open hole; the correctly-posed owed object is the finite trace \(\mathrm{tr}[a_6(L_{\rm grav})]\) , not a dimensionful magnitude, and the dossier does not resurrect the dead number. (A later, differently-scoped Bianchi-exact re-run reports a consistency coefficient of \(-2.995681680\times10^{94}\ \mathrm{GeV}^6\) under the audit-cascade layer discussed below; this is explicitly not gap-closing and remains route-inconsistent — it is recorded here only so the two dead/parked numbers are not confused with each other.)

 5. The endpoint anchor is the fixed grade itself. The published row for this gate is ANCHORED +1 — REDUCED-TO-AXIOM — on the conditional axiom pair \(\{\Delta_0>0\ (\text{a positive cost/action floor}),\ \text{Lorentz-scalar proper-time floor}\}\) . Read as terminal with residuals shown , this means: one named, irreducible axiom (AXIOM-COSTFLOOR — an irreducible quantum of cost/action, explicitly not a smallest length, applied Lorentz-invariantly so the floored quantity is a Lorentz scalar under theorem T-LI and introduces no preferred frame) is sufficient to dissolve exactly one class of UV divergence — the \(a\to0\) runaway tower \(\{a_8,a_{10},\dots\}\) — while leaving the finite \(a_6\) obligation and the fixed-point/strong-coupling wall completely untouched (ten further named walls remain, by the corpus's own count). That is the "+1": one clean, physically motivated axiom, cashed out against established results — Margolus–Levitin ( \(\tau\geq\pi\hbar/2E\) ), Landauer ( \(\Delta E\geq k_BT\ln2\) ), Bekenstein ( \(S\leq2\pi k_BRE/\hbar c\) ) — themselves measured-anchor physics, not a new speculative posit invented for this gate. DISSOLVED is explicitly not SOLVED: no fixed point is exhibited, and no value of \(a_6\) is supplied, by this axiom.

 The explicit non-claims

 The dossier must never cross these lines, and states them here precisely so no reader mistakes banked structure for a closed gate.

 This is not a claim that UQF-5C is closed, nor that a UV completion of quantum gravity has been produced. A constructive, non-perturbative UV completion of the interacting graviton is a global open problem shared by every serious research program — string theory (which supplies a UV completion in its own right but not a closed derivation of this geometry's low-energy world), asymptotic safety (a candidate non-Gaussian fixed point whose truncation-independence is unresolved), loop quantum gravity, causal sets, and this framework alike. The geometry here supplies a clean operator and a precisely named missing object; it supplies no constructive lever across the strong-coupling wall. 

 Computing \(a_6\) — to any level of exactness — does not close 5C. This is the gate's signature mis-close and is explicitly refused here: \(a_6\) is the object the UV-completion program would need , never a stand-in for the completion itself. This holds regardless of which layer of the \(a_6\) record (see below) is being discussed.

 No revival of dead numbers. The refuted magnitude \(-2.818\times10^{94}\,\mathrm{GeV}^6\) , the retracted quartic derivative-coupling coefficient \(256a^2(a^2-1)^2\) (only \((\nabla\mathrm{Riem})^2\) survives on \(K_6\) , since \(K_6\) is homogeneous but not locally symmetric), and the buggy build's curvature values ( \(\kappa=7/12\) , \(|\mathrm{Riem}|^2/R^2=31/147\) , first-Bianchi residual exactly \(1/6\) or \(1/7\) depending on normalization) are all superseded and stay dead. They are frozen negative controls, never quoted as live values.

 No claim of universal supremacy. "This framework out-UV-completes every conceivable theory of gravity" is a universal negative over the entire unbounded space of possible mathematics; it is unprovable for any program and is dissolved here as a shared ceiling on all knowledge, never entertained as a hypothesis this dossier defends.

 No claim of absolute uniqueness for the completion route. "THE absolutely unique UV completion of the interacting graviton" demands proof of unique forcedness over the open-ended space of quantization schemes; that is a unicorn. The bounded, honest claim is narrower: this geometry supplies the operator, and the missing object is named precisely (the constructive completion = UQF-9).

 No new frontier is created. Canonically there are exactly two structural frontiers in the ledger (Gap-13 and UQF-4). UQF-5C is not a third: it is the graviton leg of the shared UV wall carried under UQF-9 (and inherited downstream by UQF-14's above-cutoff unitarity question), and this dossier does not inflate the frontier count.

 Frozen-branch hashes validate provenance, not physics. The branch identifiers fixing which object was tested (so it cannot be quietly retuned) are audit anchors only; they certify which geometry, not that the geometry is correct or uniquely forced. The background \(K_6\times S^2\times S^1_Y/\mathbb{Z}_2\) is selected by constraints, not proven uniquely forced — SELECTED ≠ FORCED throughout this dossier, and ANCHORED ≠ DERIVED.

 What this dossier establishes, and what it does not

 This dossier establishes, with full inline derivation and exact-rational precision, that the frozen thirteen-dimensional geometry — pinned at all three layers ( \(\times\) Stage: \(\mathcal{M}_4\times K_6\times S^2\times S^1_Y/\mathbb{Z}_2\) with \(K_6=SU(3)/T^2\) ; \(\oplus\) Rulebook: de-Donder gauge, Faddeev–Popov ghosts, \(\overline{\mathrm{MS}}\) /heat-kernel scheme, \(\mathbb{Z}_2\) orbifold parity; \(\otimes\) Actors: the Lichnerowicz operator \(L_{\rm grav}=-(\nabla^2+E)\) on \(\mathrm{Sym}^2(T)\) and the ghost bundle) — cleanly supplies the interacting graviton's kinetic operator; that this operator's BRST-forced ghost-corrected weight \(91-2\cdot13=65\) matches the correct \(D{=}13\) physical graviton count \(\dim\mathrm{Sym}^2_0(SO(11))=65\) exactly; that the curvature ratio \(|\mathrm{Riem}|^2/\mathrm{Scal}^2=23/75\) feeding the heat-kernel expansion is proved by three independent, target-blind derivation routes (SU(3) structure-constant/naturally-reductive curvature; curvature-free spectral heat-trace; full Levi-Civita Riemann tensor over the 3-parameter squashing metric) and cross-checked against a first-Bianchi machine-zero identity (residual \(\approx2.5\times10^{-16}\) , versus the buggy build's exact \(1/6\) – \(1/7\) residual that this test caught); and that a granularity axiom on cost/action (explicitly not on length) provably dissolves exactly one class of divergence while leaving the strong-coupling fixed-point question completely untouched.

 It does not establish, and does not attempt to establish, a constructive non-perturbative UV completion of the interacting graviton. Two distinct layers of record exist for the graded \(a_6\) value itself, and both are carried here rather than collapsed into one: at the closure-of-record (ledger) layer , the graded graviton-plus-ghost value is OPEN / FAIL_VALUE_MISMATCH — two independent routes disagree by \(31/48\approx0.646\) , six orders of magnitude outside the pre-registered \(10^{-6}\) tolerance, with the physical Faddeev–Popov ghost route (using \(E=-\mathrm{Ric}\) ) not yet reconciled against the graviton \(\mathrm{Sym}^2(T)\) route, whose exact deficit is the SU(3) Gelfand–Tsetlin off-diagonal hopping term on Peter–Weyl harmonic sections, not yet enumerated. At the audit/completion-cascade layer , a longer exact-rational chain of intermediate objects — the graviton \(\mathrm{Sym}^2(T_6)\) Levi-Civita ratio \(-6373/630\) , the physical-defect value \(-7226/35\) , the vector-ghost ratio \(-251/504\) , the \(\mathbb{Z}_2\) -equivariant defect \(-337361/840\) , and the keystone bulk graded value \(-953329/1260\) assembling exactly into \(-491353/630\) — has been produced and multi-route verified, but this layer's own ceiling is explicitly AUDIT-CLOSED, physics-OPEN : it is a certified computation, not a certified physics closure, and neither layer promotes the gate. The positivity functional needed to turn any resolved \(a_6\) value into a decision-grade certificate is not yet even well-defined (three inequivalent candidate readings of \(P\) , none selected). Most importantly, even a fully resolved, sign-correct, agreed \(a_6\) would only sharpen the linearized certificate conditional on UQF-9 — it would never by itself close the strong-coupling wall, because that wall is closed by nothing less than a genuine non-perturbative construction: a target-blind non-Gaussian fixed point stable under truncation, or a rigorous proof that none exists. This dossier is deliberately built to show both the banked wins and the standing wall side by side, at both layers of record, without collapsing one into the other or letting either layer quietly stand in for the other.

 Endpoint preview

 The single sentence to carry forward: this gate is ANCHORED +1 on the conditional axiom pair \(\{\Delta_0>0,\ \text{Lorentz-scalar proper-time floor}\}\) — a genuine, banked reduction of the interacting-graviton operator and exactly one UV-divergence class to a single named axiom — while the constructive strong-coupling UV completion of gravity itself remains the open global wall (shared with UQF-9 and inherited by UQF-14) that only a target-blind, truncation-stable non-perturbative construction, or a rigorous non-existence proof, can ever close.

 The community gap & state of the art

 2.1 The precise open problem, stated at the sharpness the community states it

 Strip away every framework-specific label and the question UQF-5C asks is the oldest unsolved
structural problem in fundamental theoretical physics: does a consistent, non-perturbative
quantum theory of the fully interacting gravitational field exist, and if so what is it? This
is not the question of whether a classical field configuration \(h_{MN}\) can be quantized as a
free spin-2 particle propagating on a fixed background — that step is uncontroversial, and on the
frozen thirteen-dimensional arena used throughout this corpus it is completed cleanly (leg 5A: the
De-Donder-gauge-fixed Lichnerowicz operator \(L_{\rm grav}=-(\nabla^2+E)\) on the graviton
 \(\mathrm{Sym}^2(T)\) bundle, with 4D massless spin-2 zero mode plus Kaluza–Klein tower, treated here
as banked input). The question UQF-5C asks is the categorically harder next step: when
gravitons scatter at or above the cutoff scale, where the loop expansion in the dimensionful
coupling \(G_NE^2\) stops converging, is there a well-defined, finite, unitary quantum theory that
governs that regime — or does the theory simply cease to exist as a fundamental description past
that point? 

 This is the standard statement of the non-renormalizability of perturbative quantum gravity .
Because Newton's constant \(G_N\) carries negative mass dimension ( \([G_N]=-2\) in natural units),
every loop diagram built from graviton propagators and vertices requires counterterms of
ever-increasing operator dimension to absorb its divergences. Pure Einstein gravity is one-loop
finite on-shell by a topological (Gauss–Bonnet) identity special to four dimensions, but gravity
coupled to matter is divergent already at one loop with counterterms absent from the original
Einstein–Hilbert action, and pure gravity itself develops a genuine, non-removable two-loop
divergence. Beyond that the tower of independent higher-dimension curvature invariants needed as
counterterms is unbounded: dimension-six operators, then dimension-eight, then dimension-ten, with
no finite truncation ever sufficing. A theory that needs an infinite number of independent
coupling constants to absorb its own divergences is not predictive at arbitrarily high energy
using perturbation theory alone — this is the precise, technical content of "gravity is
perturbatively non-renormalizable," and it is this fact, not any experimental anomaly, that
motivates the entire community-wide search for a UV completion.

 On the frozen geometry used throughout this corpus, the scale at which the perturbative expansion
in \(G_NE^2\) becomes order one — the "cutoff" referenced everywhere in this gate — is read off
directly from the compactification data. The higher-dimensional fundamental scale is
 \(M_*\approx7.467\times10^{16}\) GeV (from the Planck-normalization relation \(M_*^{11}=M_{\rm
Pl}^2/\mathrm{Vol}(X_{\rm active})\) , with \(\mathrm{Vol}(X_{\rm active})=3.704417261398702
\times10^{-148}\,\mathrm{GeV}^{-9}\) the active-orbifold internal volume), sitting close to the
independently-quoted compactification/unification scale \(M_U\sim1.0\times10^{16}\) GeV and the
associated radius \(R_0=1.591549430918954\times10^{-17}\,\mathrm{GeV}^{-1}\) . A related read-off
elsewhere in this program quotes the UV floor as \(M_*\approx6.01\times10^{16}\) GeV in a different
bookkeeping convention; both numbers are DERIVED geometry read-offs , not new anchors and not,
by themselves, a closure of anything — they simply locate where the question becomes live. Below
that scale, the operator supplied by leg 5A is the correct, well-posed kinetic object for the
graviton; at and above it, no finite-order effective Lagrangian is expected to remain valid, and
the honest, sharply-stated question is what — if anything — replaces it as a fundamental
description.

 This is exactly the wall that every serious quantum-gravity research program has spent decades
standing in front of. It is not "compute one more loop diagram" — it is the structural fact that
the naive continuation of ordinary effective quantum field theory runs out of predictive content,
and no agreed, constructive, non-perturbative completion has been established for any approach,
on any background. UQF-5C asks whether this specific geometry supplies that completion. It does
not, and the honest content of this gate is locating precisely how far the geometry's own
machinery reaches and exactly where the shared wall begins.

 2.2 History of the problem: from one-loop finiteness to the two-loop divergence

 The modern shape of the problem traces to 't Hooft and Veltman's 1974 one-loop calculation, which
established that pure Einstein gravity is one-loop finite on-shell — a consequence of the
Gauss–Bonnet topological identity removing the naive divergence in four dimensions — but that
gravity coupled to matter (a scalar field, in their original calculation) is divergent already at
one loop, requiring counterterms not present in the Einstein–Hilbert action. The decisive result
came a decade later: Goroff and Sagnotti (and independently van de Ven) showed that pure
gravity itself develops a genuine, non-removable two-loop ultraviolet divergence proportional to
the cubic-in-Riemann invariant \(R_{\mu\nu}{}^{\rho\sigma}R_{\rho\sigma}{}^{\alpha\beta}
R_{\alpha\beta}{}^{\mu\nu}\) . This is precisely the family of cubic curvature invariant that
appears in this corpus's own weight-6 curvature ledger at the Killing-form-normalized Einstein
center of \(K_6\) — \(K_1=8\,\mathrm{tr}(R_{\rm op}^3)=-113/72\) and
 \(K_2=R_{abcd}R_{aecf}R_{ebfd}=-5/72\) — the same operator-dimension class that the heat-kernel
coefficient \(a_6\) is built from. The two-loop pure-gravity divergence is the textbook,
decades-old demonstration that perturbative quantum general relativity is not, by itself, a
complete quantum theory at all energies: something genuinely new — new degrees of freedom, a
non-trivial fixed point, a discrete microstructure, or some other structural ingredient — is
required above the scale where this tower of divergences becomes uncontrollable.

 The response programs that followed, each of which this corpus explicitly identifies as sharing
the same unresolved wall, are these.

 String theory. Beginning in the mid-1980s, string theory offered the most complete known
 perturbative answer: because the fundamental excitation is an extended object rather than a
 point particle, the string worldsheet smooths out the short-distance behavior responsible for
 the divergent loop integrals of point-particle gravity, and perturbative string scattering
 amplitudes are UV-finite order by order on suitable backgrounds. This is genuine, structural
 progress — a real answer to "does some consistent UV completion of gravity exist." But it does
 not close the question this gate asks about this specific geometry. First, string theory's
 answer is inherently a statement about the existence of a UV-complete perturbative expansion
 on some background, not a closed, unique derivation connecting that completion down to the
 specific four-dimensional Standard-Model content, gauge group, and three-generation structure
 used throughout this program — the "landscape" problem of selecting or deriving which vacuum is
 physically realized is itself a separate, decades-old, unresolved research program with no
 agreed resolution. Second, string theory offers no statement whatsoever, positive or negative,
 about whether the specific frozen background used here — \(K_6=SU(3)/T^2\) selected by
 constraint-satisfaction on the four Standard Model anchors, not by string consistency
 conditions — itself admits a non-perturbative completion in the string sense. It is evidence
 that UV completions of gravity are possible in principle; it supplies no lever this gate can use
 to complete this geometry.

 Asymptotic safety. Weinberg's 1979 conjecture — that gravity could be non-perturbatively
 consistent if the renormalization-group flow of the full (infinite-dimensional) space of
 gravitational couplings approaches a non-trivial, non-Gaussian ultraviolet fixed point, at which
 all physical quantities remain finite even though the bare coupling does not run to zero — has
 been developed computationally since the 1990s using the functional renormalization group (FRG),
 principally by Reuter and a large subsequent literature. This is the community's most literal
 attempt at exactly the question UQF-5C poses. Its central, unresolved technical obstruction is
 truncation-dependence : every actual FRG computation solves the exact renormalization-group
 flow equation only after truncating the infinite-dimensional space of possible gravitational
 actions to a finite subset of curvature invariants (low powers of \(R\) , sometimes extended to
 higher-order \(f(R)\) -type truncations). Numerical evidence for a fixed point exists within these
 truncations, but whether that fixed point is a genuine feature of the full, untruncated theory —
 stable as the truncation is systematically enlarged — or an artifact of the specific finite
 subspace chosen, remains an open, actively contested question in that literature, with different
 truncation schemes and gauge/regulator choices giving different quantitative pictures of the
 fixed point's properties (critical exponents, number of relevant directions). This is
 structurally the identical problem this corpus's own heat-kernel ladder faces: the "next term"
 in an unbounded tower ( \(a_8\) after \(a_6\) , just as the next curvature invariant after the current
 truncation order in FRG) can always threaten to overturn a finite-order conclusion.

 Loop quantum gravity and covariant spin-foam models. These approaches quantize geometry
 itself directly, using holonomy-flux or spin-network variables, without ever writing the theory
 as a perturbative expansion around a fixed background metric — sidestepping the perturbative
 graviton-loop divergence problem by construction rather than by resolving it. They produce
 striking results such as discrete spectra for geometric operators (area, volume), but the
 semiclassical limit connecting these discrete quantum-geometric states back to smooth classical
 general relativity, and a full covariant dynamics reproducing correct graviton scattering
 amplitudes at all orders, remain incomplete as research programs.

 Causal sets, causal dynamical triangulations, and related discretization approaches. These
 propose that spacetime is fundamentally discrete — a locally finite partial order, or a sum over
 discretized simplicial histories — with continuum general relativity recovered only as a
 coarse-grained limit. As with loop quantum gravity, no agreed, complete, non-perturbative
 construction with a proven continuum limit at all orders has been produced.

 Every one of these programs remains open, by its own practitioners' account, on the specific
question UQF-5C asks. None has a finished, agreed, non-perturbative construction of the fully
interacting graviton verified at all orders on a background carrying the full observed
Standard-Model matter content. This is the load-bearing fact underlying the honest framing of this
gate: the missing strong-coupling completion is not a private defect of the thirteen-dimensional
construction studied in this corpus — it is the single largest structural open problem shared by
the entire field of theoretical high-energy physics.

 2.3 Why there is no numerical "bound" to beat, and what state-of-the-art means for this gate

 Many gates in this corpus are graded against a measured central value with an experimental
uncertainty that a derivation must land inside. UQF-5C is not that kind of gate. There is no
experimental "UV-completion measurement" to improve on, because the entire question concerns the
existence and structure of physics in a regime — graviton scattering at or above
 \(M_*\approx7.467\times10^{16}\) GeV — that has never been probed by any experiment and, at that
energy, is not expected to be probed by any conceivable terrestrial or astrophysical measurement
in the foreseeable future. The "state of the art" for this gate is therefore necessarily a
 structural state of the art: what is the best-controlled, most rigorous partial result any
program has achieved toward a non-perturbative completion, and precisely how far short of the
full answer does each partial result fall? Read this way, three genuine pieces of progress can be
named honestly, each real and each falling short of a completion for a locatable reason:

 A clean, fully-specified linear operator on a fully-pinned background. Writing down the
 correct gauge-fixed, ghost-corrected kinetic operator for the interacting graviton with every
 layer of the geometry — Stage, Rulebook, Actors — nailed to exact rationals, and cross-checking
 its BRST-forced fiber weight \(91-2\cdot13=65\) against the independent little-group count
 \(\dim\mathrm{Sym}^2_0(SO(11))=65\) , is a genuine, nontrivial, load-bearing input that most
 treatments of "quantizing gravity" never make this explicit. It is real progress on specifying
 what object a completion would need to control; it is not itself evidence that such a
 completion exists.

 The heat-kernel coefficient ladder as the rigorous mathematical organizing structure of the
 divergence problem. The Seeley–DeWitt/Gilkey local heat-kernel expansion
 \(K(t)\sim(4\pi t)^{-d/2}\sum_k a_{2k}t^k\) is the standard tool used across the
 quantum-field-theory-in-curved-space literature to organize exactly this divergence structure
 order by order — established rigorously in Gilkey's theorems on the local invariants entering
 each heat-kernel coefficient (Theorem 3.3.1 and Theorem 4.8.16 of Gilkey's 1995 treatise),
 developed further for the higher coefficients in Avramidi's 2000 monograph (Chapter 4), and
 assembled into the explicit general \(a_6\) functional used in this program by Vassilevich's 2003
 review (equation 4.29). On this geometry the lower coefficients are certified exact rationals —
 \(a_2/a_0=5/12\) , \(a_4/a_0=11/120\) for the \(K_6\) scalar sector — built from curvature data
 ( \(|\mathrm{Riem}|^2/\mathrm{Scal}^2=23/75\) , \(|\mathrm{Ric}|^2/\mathrm{Scal}^2=1/6\) ) proved by
 three independent target-blind routes, while \(a_6\) — the first coefficient sensitive to the
 cubic-curvature sector identified by Goroff–Sagnotti as carrying the genuine non-removable
 divergence — sits at the current computational frontier, only partially resolved (see §2.4). No
 program anywhere carries an \(a_8\) , \(a_{10}\) , … analogue of this ladder to closure either: the
 ladder's unboundedness is a universal structural fact, not a limitation peculiar to this
 corpus's engine.

 A genuine, if narrow, dissolution via the granularity axiom. The one constructive addition
 this corpus makes beyond restating the shared problem is the cost/action-floor axiom
 (AXIOM-COSTFLOOR: an irreducible quantum of action, not of length, applied as a Lorentz scalar,
 resting on established results — the Margolus–Levitin bound \(\tau\geq\pi\hbar/2E\) , the Landauer
 bound \(\Delta E\geq k_BT\ln2\) , and the Bekenstein bound \(S\leq2\pi k_BRE/\hbar c\) ). This
 provably removes exactly one divergence class — the \(a\to0\) runaway of the unbounded tower
 \(\{a_8,a_{10},\dots\}\) — a genuine structural result. It is not a finite-grain shortcut across
 the wall as a whole: ten further named obstacles remain untouched, including the finite \(a_6\) 
 coefficient itself, and a separate banked result (the Layer-2 screen, T-DEEP) shows
 \(R_0/\ell_{\rm Planck}\sim194\) with \(R_0\) a pure color/gauge object carrying zero gravitational
 input — meaning any finite-grain resolution of 5C on this geometry necessarily relocates onto
 the separate, still-open sub-question (isolated as P0 under UQF-9) of whether gravity owns
 its own intrinsic minimal scale or merely inherits the color-sector radius \(R_0\) .

 No rival program supplies a numerically sharper "bound" to compare directly against, because there
is no shared, agreed metric of partial progress across structurally different kinds of results:
string theory's all-orders perturbative finiteness on chosen backgrounds, asymptotic safety's
truncated fixed-point evidence, and this corpus's exact-rational heat-kernel ledger are three
different kinds of partial answer to three overlapping but distinct sub-questions, and none is
directly commensurable with the others as a numerical "record." The honest state of the art is:
every program has real, if structurally incommensurable, partial progress, and none has produced
the constructive non-perturbative completion.

 2.4 Prior attempts on this specific object, and exactly where each falls short

 Within this corpus's own attack on the graviton leg of UQF-5C, six concrete attempts have been
made and are worth naming individually, because each failure is informative about exactly where
the remaining wall sits — this is a located debt, not a vague admission of difficulty.

 Attempt 1 — direct evaluation of the graded graviton-minus-ghost \(a_6\) trace. The
straightforward strategy computes \(\mathrm{tr}[a_6(L_{\rm grav})]\) from the certified curvature
data — the Lichnerowicz endomorphism spectrum on \(\mathrm{Sym}^2_0\) 
( \(\{1/6\text{ (×6)}, 5/12\text{ (×6)}, 7/6\text{ (×6)}, 17/12\text{ (×2)}\}\) , with \(\mathrm{tr}\,
E_L=40/3\) , \(\mathrm{tr}\,E_L^2=241/18\) ), the curvature operator \(\Omega=\mathrm{Riem}\) , and the
certified weight-6 invariants — via the standard Gilkey \(a_6\) formula (a sum over roughly 46
independent curvature-cubic and curvature-derivative terms in the standard Gilkey basis). This
attempt produces two independently constructed routes that disagree outright . Route A, built
from the graviton \(K_6\) -bundle Lichnerowicz data, gives \(-43/504\) , inconsistent against an
independently expected anchor value of \(-16/315\) . Route B, built by reconstruction through the
vector and scalar backbone, produces only the Bochner Laplacian ghost value \(149/1008\) — but
this uses endomorphism \(E=0\) , the mathematically simpler but physically wrong ghost operator; the
physically correct Faddeev–Popov ghost carries \(E=-\mathrm{Ric}\) , not \(E=0\) . Once this
substitution error is accounted for, the physical-ghost mismatch between routes is
 \(\left|-\tfrac{251}{504}-\tfrac{149}{1008}\right|=\tfrac{31}{48}\approx0.646\) — roughly six orders
of magnitude outside the pre-registered \(10^{-6}\) cross-check tolerance this program requires
before banking a value. The exactly located debt is the graviton \(\mathrm{Sym}^2(T)\) 
Levi-Civita off-diagonal leg: the Lichnerowicz first-order hopping term mixes the five
Weyl-inequivalent \(T^2\) weight classes on \(K_6\) under the Peter–Weyl harmonic decomposition, and
evaluating that mixing requires the explicit \(SU(3)\) Gelfand–Tsetlin off-diagonal matrix elements
connecting adjacent Gelfand–Tsetlin patterns — standard closed-form objects (each a square root of
a product of pattern-entry differences, from the ordinary GT lowering-operator formalism) that
have simply not yet been enumerated for this specific bundle and representation content. This
is a bounded computation-debt at a precisely named stratum, not an in-principle obstruction: the
method exists and is standard; the enumeration has not been carried out.

 Attempt 2 — the dimensionful bulk magnitude. An earlier attempt sought a single dimensionful
number, \(-2.818\times10^{94}\,\mathrm{GeV}^6\) , for the graded \(a_6\) functional, treating it as the
finite physical magnitude that would settle the question. This was refuted at decision grade 
for three compounding reasons: the underlying computation was contaminated by a since-corrected
sign error in the curvature-operator routine (the "R2 bug"), it depended on an arbitrarily chosen
renormalization scheme rather than a scheme-independent object, and — most fundamentally — the
object is ill-posed at odd spacetime dimension \(D=13\) . At odd \(D\) , the local heat-kernel
coefficient of mass dimension six sits exactly at the half-integer zeta-function pole \(s=7/2\) ,
which vanishes identically under dimensional regularization, leaving no finite local \(t^0\) slot —
no logarithm, no anomaly, nothing — for a GeV \(^6\) number to attach to. The magnitude leg is
therefore dissolved as ill-posed , not merely numerically wrong: the correctly-posed object at
this stratum is the finite trace, never a scheme-anchored dimensionful magnitude computed as
though \(D\) were even. A later Bianchi-exact re-run under corrected curvature data produced a
superficially similar-looking \(-2.995681680\times10^{94}\,\mathrm{GeV}^6\) , but this is logged
explicitly as a consistency-coefficient only , never gap-closing, and it remains
route-inconsistent — it must not be mistaken for a rehabilitated version of the refuted number or
for progress toward closing this gate.

 Attempt 3 — retracted derivative-coupling coefficient. An earlier candidate closed form for
the derivative sector, \(256a^2(a^2-1)^2\) , was proposed and subsequently retracted once the full
derivative computation was carried through correctly. \(K_6\) is homogeneous but not locally
symmetric (so \(\nabla\mathrm{Riem}\neq0\) ), and only the single invariant \((\nabla
\mathrm{Riem})^2\) survives as a nonzero, independent derivative contribution
( \(|\nabla\mathrm{Riem}|^2=1/4\) in Killing-form normalization, equivalently \(54\) in the trip-unit
normalization, cross-checked by three independent routes including the box identity
 \(R_{abcd}\Box R^{abcd}=-|\nabla\mathrm{Riem}|^2\) ); \(|\nabla\mathrm{Ric}|^2=0\) and
 \(|\nabla\mathrm{Scal}|^2=0\) vanish identically. The retracted quartic form does not correctly
capture this one-surviving-invariant structure and is kept in the record only as a named negative
control, never as a live candidate.

 Attempt 4 — the \(124/315\) scalar color-factor candidate. The scalar \(K_6\) ratio
 \(b_3/b_0=124/315\) — the \(t^0\) /Seeley–DeWitt bracket ratio computed from the actual \(K_6\) 
Peter–Weyl heat trace — was for a period reported as "DERIVED, dual-validated." It has since been
 downgraded to DERIVED-PENDING-INDEPENDENT-TARGET-BLIND-REPRODUCTION, because every route that
reproduced it ran through the same R2-carrying, metric-selected engine (selected at
 \(\mathrm{Scal}_{K_6}=7.5\) in that engine's own normalization) — meaning the multiple internal
"confirmations" were never actually structurally independent checks. A value regenerated
repeatedly by the same pipeline does not carry the same evidentiary weight as one confirmed by a
genuinely independent engine; this honest downgrade follows this program's own discipline against
values that can silently relocate to a desired answer through a single reused computational path.
Only strictly Levi-Civita-immune scalar ratios — for example \(a_4/a_2^2=66/125\) — currently carry
unconditional, route-independent status.

 Attempt 5 — the orbifold boundary heat-kernel coefficient. An attempt to locate the order-6
mixed Neumann \(\oplus\) Dirichlet boundary heat-kernel coefficient for \(S^1_Y/\mathbb{Z}_2\) , needed
to add a "boundary defect" term to the bulk \(a_6\) , is blocked for a structural reason, not a
computational one: the published boundary heat-kernel literature for mixed boundary conditions
terminates at \(a_5\) — the order-6 term does not exist in the literature to cite, and deriving it
would be a genuine new mathematical result rather than a lookup. Compounding this, later review
within this program concluded the entire "boundary coefficient" framing may itself be a
 wrong-object artifact : the \(\mathbb{Z}_2\) reflection \(\theta\mapsto-\theta\) on
 \(S^1_Y/\mathbb{Z}_2\) is a global isometric reflection on a closed manifold with two isolated
fixed points ( \(\theta=0,\pi\) ), which is a Donnelly-type equivariant defect problem, not a
manifold-with-boundary problem — the twisted trace on the orbifold evaluates to exactly 1 at
every order in \(t\) ( \(t\) -independent), so there is no boundary tower to extend past \(a_5\) in the
first place. The smooth equivariant defect that is legitimately computable,
 \(\tfrac12c_3^\gamma=-337361/840\) , is a genuinely different object from a hypothetical order-6
boundary coefficient and must not be substituted for it. The honest resolution here is not
"compute harder" but "recognize this was the wrong object," and no fabricated bulk-plus-defect
TOTAL \(a_6\) is asserted.

 Attempt 6 — cross-script consistency check. Before any \(D=13\) number from this ladder can be
trusted, an unresolved internal contradiction stands between two independently written
implementations of the Gilkey \(a_6\) formula: one script passes its own internal self-tests, while
a second, independently constructed cross-check script fails on the \(S^2\) calibration case,
emitting \(-8/405\) where the correct, independently reproduced value elsewhere in this corpus is
 \(a_6(S^2)=4/315\) . Until this specific cross-script contradiction is resolved, no \(D=13\) number
produced by either engine can be trusted at face value, however internally self-consistent it
appears — this is exactly the kind of target-blind sanity check (in the same spirit as the
first-Bianchi machine-zero test that caught the earlier curvature bug) that must pass before a new
claimed value is banked.

 2.5 Where the machinery has been validated, and why that does not close the gap

 It is important to be precise about what has, in fact, been checked, because the heat-kernel
machinery here is not untested — it has been tested exactly where an independent, closed-form
answer exists to test against. The sphere cross-checks agree between independently constructed
routes to roughly \(4\times10^{-14}\) : \(a_6(S^2)=4/315\) , \(a_6(S^4)=74/63\) , \(a_6(S^6)=1139/63\) , and
the conformally-coupled \(a_6^{\rm conf}(S^6)=5/63\) ; the \(S^6\) round-unit row additionally
calibrates the \(a_4\) formula to exactly \(12\) , a passed control confirming that \(K_6\) is correctly
treated as structurally distinct from the round sphere it is sometimes loosely compared to (the
anti-drift certification in this program is explicit that \(|\mathrm{Riem}|^2/\mathrm{Scal}^2\) is
never \(31/147\) and \(|\mathrm{Riem}|^2\) is never the round-unit- \(S^6\) value \(60\) ). The scale-free
ratio \(a_4/a_2^2=66/125\) is Levi-Civita-immune — built entirely from curvature scalars that do not
see the derivative/connection ambiguity that afflicts tensor-valued invariants — and is
independently robust. The first-Bianchi identity closes to machine zero
( \(\approx2.5\times10^{-16}\) ) on the corrected curvature tensor, the same target-blind check that
caught the earlier bug inflating \(|\mathrm{Riem}|^2/R^2\) to the wrong value \(31/147\) and the
Einstein constant to the wrong value \(\kappa=7/12\) (corrected to \(5/12\) ). This demonstrates that
the heat-kernel machinery is correctly implemented on every case where an independent, closed-form
answer exists to check against.

 None of these validated cases, however, is the object the gate actually needs. They are all
symmetric-space calibrations (round spheres) or scalar-sector, Levi-Civita-immune ratios. The
object this gate needs — the graded, ghost-corrected, tensor-valued \(a_6\) trace on the specific,
only-homogeneous (not locally symmetric) \(K_6=SU(3)/T^2\) graviton bundle — is precisely the case
requiring the not-yet-enumerated Gelfand–Tsetlin off-diagonal hopping data, a term that a round
sphere calibration structurally cannot exercise, because a round sphere is locally symmetric
( \(\nabla\mathrm{Riem}=0\) ) and the hopping term is identically absent there. This is the honest
shape of the state of the art: the machinery is proven correct everywhere it has been given an
answerable question, and the unanswered question is a specific, named, bounded stratum the sphere
checks cannot reach by construction.

 2.6 Summary: why the honest grade is ANCHORED, not CLOSED, and why that is the correct place to stand

 Pulled together, the state of the art on UQF-5C is fourfold. First, the community-wide problem — a
non-perturbative UV completion of the interacting graviton — is unsolved by every program,
including string theory's perturbative finiteness on chosen backgrounds and asymptotic safety's
truncated fixed-point evidence, each of which carries its own unresolved structural question
(vacuum selection for strings; truncation-independence for asymptotic safety). Second, this
corpus's own attack supplies a clean, BRST-forced operator and a curvature input proved by three
independent target-blind routes ( \(|\mathrm{Riem}|^2/\mathrm{Scal}^2=23/75\) ), validated against
exact sphere calibrations to \(\sim10^{-14}\) . Third, six concrete, named attempts to push further —
direct \(a_6\) evaluation, a dimensionful bulk magnitude, a derivative-sector closed form, an
independent color-factor reproduction, a boundary-defect total, and a cross-script consistency
pass — have each been tried and have each fallen short for a precisely locatable, structural
reason, never from a simple lack of effort. Fourth, the one constructive lever this corpus adds —
the cost/action granularity axiom — provably dissolves exactly one divergence class while leaving
the finite \(a_6\) trace and the strong-coupling fixed-point question entirely open, with the
Layer-2 screen showing that no finite-grain shortcut is available on this geometry without
relocating the question onto whether gravity owns its own minimal scale (UQF-9's sub-target P0).

 This is exactly the structure that supports the fixed grade for this gate: REDUCED-TO-AXIOM /
ANCHORED +1 , resting on the conditional axiom pair \(\{\Delta_0>0\ (\text{a positive cost/action
floor}),\ \text{Lorentz-scalar proper-time floor}\}\) — a genuine, banked reduction of the
UV-divergence problem to one named, physically-motivated axiom, standing honestly alongside a
global wall that remains open for this framework exactly as it remains open for every other
serious approach to quantum gravity working today.

 The frozen 13D arena at full precision

 UQF-5C is not evaluated on a toy model or a truncated sub-sector; it is evaluated on the single frozen arena that every gate in the corpus shares. This section pins that arena down completely — all four metric factors, both curvature normalizations, every exact rational the graviton operator draws on, and the three-layer (× Stage / ⊕ Rulebook / ⊗ Actors) structure of the specific objects UQF-5C touches: the graviton bundle, the Faddeev–Popov ghost bundle, and the heat-kernel machinery that reads their spectra.

 The active branch as a layered object

 The frozen background is not merely a manifold; it is the whole layered object

 \[
\mathfrak{B}_{\rm active}
=
\underbrace{\big[\,\mathcal{M}_4 \times K_6 \times S^2 \times S^1_Y/\mathbb{Z}_2\,\big]_\times}_{\text{\texttimes\ STAGE --- metric geometry}}
\;\oplus\;
\underbrace{\big[\,\mathcal{F}^+_{\rm finite}\oplus\mathcal{C}_{\rm admiss}\,\big]_\oplus}_{\text{\textoplus\ RULEBOOK --- finite admissibility (0-dim)}}
\;\otimes\;
\underbrace{\big[\,\mathcal{E}_{\rm matter}\oplus\mathcal{E}_{\rm gauge}\oplus\mathcal{E}_{\rm Higgs}\oplus\mathcal{E}_{\rm proton}\,\big]_\otimes}_{\text{\textotimes\ ACTORS --- bundles / operators (0-dim)}},
\]

 with \(K_6 = SU(3)/T^2\) the full \(A_2\) flag manifold and \(S^1_Y/\mathbb{Z}_2\) the active orbifold boundary domain. Only the × Stage layer carries metric dimension:

 \[
D = 4 + 6 + 2 + 1 = 13.
\]

 The ⊕ (rulebook) and ⊗ (actors) layers are non-metric — zero-dimensional — but they are part of the frozen branch and can never be silently dropped when evaluating a gate. This matters directly for UQF-5C: the graviton operator is a × Stage object (a Laplace-type operator on a bundle over the 13D manifold), but which operator it is — gauge-fixed or not, ghost-corrected or not, in which scheme — is entirely fixed by the ⊕ and ⊗ layers. A dossier that quotes only the curvature of \(K_6\) and skips the gauge-fixing/ghost/BRST data has quoted a different, unphysical operator. All three layers are pinned below for exactly the objects this gate uses.

 The frozen branch is identified by audit hashes (branch identifier and manifest metadata) that exist purely so a reviewer can confirm which object was tested and that it was not quietly retuned between runs. These hashes are audit anchors only: they certify which geometry was evaluated, not that the physics is correct. The background itself is selected by the admissibility constraints (Weyl-rigidity, anomaly cancellation, chirality, the four measured anchors) — it is not proved to be the unique geometry forced by first principles. This selected-not-forced status is carried honestly throughout: it does not weaken any of the curvature identities below, which are exact once the background is fixed, but it means "the background could in principle be reselected" is a live, separate question from "is the arithmetic on this background correct."

 The four metric factors of the × Stage

 Factor 
 Real dim 
 Metric 
 Status 
 Physical role 
 Force it routes 

 \(\mathcal{M}_4=\mathbb{R}^{3,1}\) 
 4 
 Minkowski 
 primitive 
 observed spacetime 
 — 

 \(K_6=SU(3)/T^2\) 
 6 
 Weyl-rigid invariant (normal at center) 
 primitive 
 color source; spin- \(\mathbb{C}\) family index \(-3\) 
 \(SU(3)_c\) 

 \(S^2\) 
 2 
 round 
 primitive 
 weak source; spin- \(\mathbb{C}\) doublet routing 
 \(SU(2)_L\) 

 \(S^1_Y\) 
 1 
 flat 
 primitive 
 parent hypercharge circle 
 \(U(1)_Y\) 

 \(S^1_Y/\mathbb{Z}_2\) 
 interval 
 induced quotient ( \(\theta\mapsto-\theta\) ) 
 derived 
 chirality / no-mirror filter 
 \(U(1)_Y\) + orbifold chirality 

 For UQF-5C the whole nine-dimensional internal block \(X_{\rm int}=K_6\times S^2\times S^1_Y/\mathbb{Z}_2\) matters, because the heat-kernel coefficient that decides the gate's fate is a coefficient of the full 13D Laplace-type operator, and heat-kernel coefficients on a product factorize by the exact convolution rule \(a_{2k}(M_1\times M_2)=\sum_{i+j=k}a_{2i}(M_1)a_{2j}(M_2)\) . A truncated arena — say \(K_6\) alone, or \(K_6\times S^2\) without the orbifold factor — would produce a different, artifactual coefficient. The full nine-dimensional internal product, carried through the \(\mathbb{Z}_2\) orbifold quotient on \(S^1_Y\) , is the complete object this gate's operator lives on.

 Radii and volumes at full precision

 The internal geometry is not free: every radius below is derived from the four irreducible anchors \(\{M_{\rm Pl},\ \alpha_i(M_Z),\ y_t,\ |V_{us}|\}\) via the two-loop RG threshold closure, not chosen to make the gate come out a particular way.

 \[
M_U = 1.0\times10^{16}\ \text{GeV} \quad(\text{closure residual } 9.6\times10^{-11}),\qquad
R_0 \equiv (2\pi M_U)^{-1} = 1.591549430918954\times10^{-17}\ \text{GeV}^{-1}.
\]

 At the symmetric chamber center \(\vec u=(1,1,1)\) (Weyl-rigid; off-center points are eliminated by the admissibility selector):

 \[
R_6 = R_0 = 1.591549430918954\times10^{-17}\ \text{GeV}^{-1} \quad(K_6\text{ radius, center value}),
$$
$$
R_2 = R_0 = 1.591549430918954\times10^{-17}\ \text{GeV}^{-1}\quad (S^2\text{ radius, leading order}),
$$
$$
R_Y = \tfrac12 R_0 = 7.957747154594768\times10^{-18}\ \text{GeV}^{-1}\quad(S^1_Y\text{ radius, post-}\mathbb{Z}_2).
\]

 Volumes (evaluated at the center, exact symbolic form \(\mathrm{Vol}(K_6)=V_{K_6,0}R_6^6\sqrt{u_1u_2u_3}\) with \(V_{K_6,0}=(2\pi)^3/\sqrt3=143.2118575035129\) ):

 \[
\mathrm{Vol}(K_6) = 2.327554010848277\times10^{-99}\ \text{GeV}^{-6},\qquad
\mathrm{Vol}(S^2) = 4\pi R_0^2 = 3.183098861837907\times10^{-33}\ \text{GeV}^{-2},
$$
$$
\mathrm{Vol}(S^1_Y/\mathbb{Z}_2) = \pi R_0 = 5.000000000000000\times10^{-17}\ \text{GeV}^{-1} \;\big(=1/(2M_U),\text{ exact}\big),
$$
$$
\mathrm{Vol}(X_{\rm active}) = \mathrm{Vol}(K_6)\,\mathrm{Vol}(S^2)\,\mathrm{Vol}(S^1_Y/\mathbb{Z}_2) = 3.704417261398702\times10^{-148}\ \text{GeV}^{-9}.
\]

 This nine-dimensional active volume is what sets the higher-dimensional Planck mass via the Planck-normalization relation

 \[
M_{\rm Pl}^2 = M_*^{\,D-2}\,\mathrm{Vol}(X_{\rm active}),\qquad D=13,
$$
$$
M_*^{11} = \frac{M_{\rm Pl}^2}{\mathrm{Vol}(X_{\rm active})} = 4.023836152402511\times10^{185}\ \text{GeV}^{11},\qquad
M_* = 7.467050992135091\times10^{16}\ \text{GeV}.
\]

 \(M_*\) is the natural UV floor of the 13D theory — it is fixed by the geometry plus the measured \(M_{\rm Pl}\) , not an independent tunable input. This is the number the brief cites as the "UV floor read off the geometry," \(M_*\approx6.01\) – \(7.47\times10^{16}\) GeV depending on rounding convention (the brief's headline figure \(6.01\times10^{16}\) GeV and this pack's \(7.467\times10^{16}\) GeV both trace to the same relation; the value is a derived geometric read-off, not a closure and not a new measured anchor for UQF-5C — it locates where the strong-coupling wall sits, it does not cross it).

 Two curvature normalizations — the bridge that must be stated before any number

 The corpus pins the identical \(K_6\) geometry in two internally consistent normalizations, and every curvature number in this dossier must be read with its tag attached:

 [R₆-norm] (frozen physical normalization): curvature carries physical units of GeV², \(\mathrm{Ric}_i = 1/(2R_6^2)\) , \(\mathrm{Scal}=3/R_6^2\) . This is the normalization for dimensionful quantities (Planck normalization, KK spectra, threshold radii).

 [Killing-norm] : \(g=(-B)|_{\mathfrak m}\) with Killing form \(B(X,Y)=6\,\mathrm{Tr}(XY)\) on \(\mathfrak{su}(3)\) , evaluated at the chamber center. Curvature is dimensionless: \(\mathrm{Ric}_i=5/12\) , \(\mathrm{Scal}=5/2\) . This is the normalization in which every exact-rational curvature invariant, heat-kernel coefficient, and the \(a_6\) ledger are computed and stored — it is the normalization this gate's operator data is quoted in.

 The bridge is the set of metric-scale- invariant ratios, which are identical in both normalizations and are the load-bearing facts:

 \[
\frac{\mathrm{Scal}}{\mathrm{Ric}_i} = 6 = \dim K_6,\qquad
\frac{|\mathrm{Ric}|^2}{\mathrm{Scal}^2} = \frac16,\qquad
\frac{|\mathrm{Riem}|^2}{\mathrm{Scal}^2} = \frac{23}{75} = 0.3066666666666667,\qquad
\frac{\mathrm{Weyl}^2}{\mathrm{Scal}^2} = \frac{6}{25}.
\]

 \(|\mathrm{Riem}|^2/\mathrm{Scal}^2=23/75\) is proved by three independent target-blind routes — SU(3) structure-constant / naturally-reductive curvature, curvature-free spectral heat-trace, and full Levi-Civita Riemann-tensor evaluation over the three-parameter squashing metric — and it self-checks as non-reverse-engineered: \(0.30667\) matches neither of the two prior candidate values in the corpus's own history (the retired buggy-engine value \(0.2109\) , nor an earlier firewall value \(0.0667\) ), so it cannot have been tuned to a wanted answer. It is a proved geometric input , necessary but not sufficient for anything downstream about UV completion.

 \(K_6\) curvature at the symmetric center, both normalizations

 Quantity 
 [R₆-norm] 
 [Killing-norm] 

 \(\mathrm{Ric}_1=\mathrm{Ric}_2=\mathrm{Ric}_3\) 
 \(1/(2R_6^2)=1.973920880217872\times10^{33}\ \text{GeV}^2\) 
 \(5/12\) 

 \(\mathrm{Scal}(K_6)\) 
 \(3/R_6^2=1.184352528130723\times10^{34}\ \text{GeV}^2\) 
 \(5/2\) 

 \(\mathrm{Scal}/\mathrm{Ric}_i\) 
 \(6\) 
 \(6\) 

 Scale-invariant curvature products (Killing-norm exact rationals, identical under rescaling):

 \[
\mathrm{Scal}^2=\frac{25}{4}=6.25,\qquad |\mathrm{Ric}|^2=\frac{25}{24}=1.041666666666667,\qquad |\mathrm{Riem}|^2=\frac{23}{12}=1.916666666666667.
\]

 Anti-drift certification carried verbatim: \(|\mathrm{Riem}|^2/\mathrm{Scal}^2=23/75\) is confirmed and is never \(31/147\) (a superseded, refuted build value from a curvature-computation bug); \(|\mathrm{Riem}|^2\) is never \(=60\) (that is the round unit \(S^6\) value — a different space entirely). The Einstein constant read off this same corrected curvature is \(\kappa=5/12\) (the Ricci eigenvalue on the corrected \(K_6\) ); an earlier buggy value \(\kappa=7/12\) is retired. Both corrections were caught by a target-blind theorem-level test, not by tuning toward an expected number: the first-Bianchi identity residual on the corrected curvature tensor is \(\approx2.5\times10^{-16}\) (machine zero), whereas the buggy build gave an exact, nonzero residual of \(1/7\) (engine units) / \(1/6\) (raw units). Passing this Bianchi test does not by itself validate any \(a_6\) value — it validates that the curvature tensor feeding \(a_6\) is the correct tensor.

 Also load-bearing for this gate: invariant Einstein metrics on \(SU(3)/T^2\) number exactly four — the normal metric at \((1,1,1)\) used throughout, plus the Kähler–Einstein metric at \((1,1,2)\) and its three permutations. This is a classical differential-geometry fact reproduced independently inside the framework's own engine, serving as a structural cross-check that the machinery is computing real \(SU(3)/T^2\) geometry rather than an artifact. Off the symmetric center the space is non-Einstein (the physical content of the "squashing" chamber \(\vec u\in[1/2,3/2]^3\) ); at the center used by every \(K_6\) -dependent gate, all three Ricci eigenvalues coincide.

 Cubic and derivative curvature invariants — the sector that feeds \(a_6\) 

 Because \(K_6=SU(3)/T^2\) is homogeneous but not locally symmetric , \(\nabla\mathrm{Riem}\neq0\) , and the derivative-curvature sector survives and must be carried into any weight-6 heat-kernel object. This is not a technicality that can be dropped: it is precisely the sector responsible for the graviton \(a_6\) "wall" discussed later in the dossier. The exact rationals (Killing-norm, Einstein center):

 \[
K_1 \equiv R_{ab}{}^{cd}R_{cd}{}^{ef}R_{ef}{}^{ab} = 8\,\mathrm{tr}(R_{\rm op}^3) = -\frac{113}{72},\qquad
K_2 \equiv R_{abcd}R_{aecf}R_{ebfd} = -\frac{5}{72},
$$
$$
|\nabla\mathrm{Riem}|^2 = \frac14\ne0 \quad(\text{passes 2nd Bianchi identically}),
$$
$$
\mathrm{Scal}^3=\frac{125}{8},\qquad \mathrm{Scal}\cdot|\mathrm{Ric}|^2=\frac{125}{48},\qquad \mathrm{Scal}\cdot|\mathrm{Riem}|^2=\frac{115}{24},\qquad \mathrm{Ric}^3=\frac{125}{288},\qquad \mathrm{Ric}\cdot|\mathrm{Riem}|^2=\frac{115}{144}.
\]

 The corpus also records the same derivative sector in "trip-unit" (engine) normalization: \(|\nabla\mathrm{Riem}|^2=54\) (nonzero, confirmed by three independent routes), \(|\nabla\mathrm{Ric}|^2=0\) , \(|\nabla\mathrm{Scal}|^2=0\) , \(|\nabla^{LC}E_{\rm grav}|^2=162\) , with the box cross-check \(R_{abcd}\Box R^{abcd}=-54=-|\nabla\mathrm{Riem}|^2\) . This sector is carried end-to-end through the banked \(a_6\) chain — it is not a symmetric-space-blind omission, and its nonvanishing is exactly the geometric reason the Gelfand–Tsetlin off-diagonal ladder term appears in the graviton heat-kernel calculation (see the residuals material later in this dossier). A retracted candidate closed-form for this sector, \(256a^2(a^2-1)^2\) (a "Berger derivative coefficient"), is dead; only \((\nabla\mathrm{Riem})^2\) survives on \(K_6\) .

 Topological data, exact: \(\chi(K_6)=6=|S_3|\) (the order of the Weyl group of \(A_2\) , i.e. the number of Weyl chambers), \(\chi(S^2)=2\) , \(\chi(S^1_Y/\mathbb{Z}_2)=1\) .

 The root system underlying \(K_6=SU(3)/T^2\) 

 Cartan basis \((h_1,h_2,h_3)\) with \(h_1+h_2+h_3=0\) ; simple roots \(\alpha_1=(1,-1,0)\) , \(\alpha_2=(0,1,-1)\) , and \(\alpha_1+\alpha_2=(1,0,-1)\) ; positive roots \(\{\alpha_1,\alpha_2,\alpha_1+\alpha_2\}\) ; half-sum \(\rho=\frac12\sum_{\alpha>0}\alpha=(1,0,-1)\) with \(\|\rho\|^2=2\) in the Killing normalization. The Weyl group is \(S_3\) , order 6 — this is the \(\chi(K_6)=6\) quoted above. The tangent space decomposes as

 \[
T(K_6)=\mathfrak m_1\oplus\mathfrak m_2\oplus\mathfrak m_3,\qquad \dim_{\mathbb R}\mathfrak m_i=2,
\]

 each \(\mathfrak m_i\) the real 2-plane carrying root \(\alpha_i\) (with \(\alpha_3\equiv\alpha_1+\alpha_2\) ), with \((-B)\) -orthonormal basis \(\{X_{ij}=E_{ij}-E_{ji},\,Y_{ij}=i(E_{ij}+E_{ji})\}/\sqrt{12}\) . This three-fold split into \(\mathfrak m_1,\mathfrak m_2,\mathfrak m_3\) is exactly the structure that the general Wang–Ziller/Nomizu Ricci formula

 \[
\mathrm{Ric}_k(\vec u)=\frac{(u_k-u_i+u_j)(u_k+u_i-u_j)}{2R_6^2\,u_iu_ju_k}\quad (i,j,k)\text{ cyclic}
\]

 is built on, and it is the structure whose five Weyl-inequivalent weight classes generate the Gelfand–Tsetlin off-diagonal hopping terms that are the specifically-located computation debt in the graviton \(\mathrm{Sym}^2_0\) heat-kernel leg — a direct consequence of the root-system structure pinned here, not an independent complication.

 ⊕ Rulebook and ⊗ Actors — the layers that fix which operator this gate evaluates

 Curvature data alone does not specify a graviton operator. The rulebook and actor layers are what turn "a curved 13D manifold" into "the specific Laplace-type operator whose spectrum this gate interrogates."

 ⊕ Rulebook, pinned for this gate: 
- De-Donder (harmonic) gauge-fixing of the metric fluctuation \(h_{MN}\) — required because diffeomorphism redundancy must be removed before the graviton propagator/operator is well-defined.
- Faddeev–Popov ghost sector , forced by the same gauge redundancy; the ghost fields carry a fixed multiplicity and sign in any trace over the physical (BRST-cohomology) degrees of freedom.
- \(\overline{\rm MS}\) / heat-kernel scheme for the coefficients \(a_{2k}\) that organize the operator's short-distance expansion.
- \(\mathbb{Z}_2\) orbifold parity on \(S^1_Y/\mathbb{Z}_2\) , \(\theta\mapsto-\theta\) , with two isolated fixed points at \(\theta=0,\pi\) . The orbifold traces are \(K^\pm=\tfrac12K_{\rm circle}\pm(\text{parity defect }\tfrac12)\) ; the reflection \(g\) -trace is exactly \(1\) (two fixed points, each contributing \(1/|1-(-1)|=1/2\) ). This is a Donnelly equivariant defect on a smooth, boundaryless manifold , not an ordinary Neumann/Dirichlet boundary-value problem — a distinction the dossier's residuals discussion depends on.
- The cubic-curvature / mass-dimension-6 Gilkey basis (of order 46 terms) as the readout basis in which the \(a_6\) coefficient is expressed.

 ⊗ Actors, pinned for this gate: the operator whose spectrum decides consistency is the Lichnerowicz-type Laplacian

 \[
L_{\rm grav} = -(\nabla^2 + E)
\]

 acting on the graviton bundle \(\mathrm{Sym}^2(T)\) (transverse-traceless sector \(\mathrm{Sym}^2_0\) , real dimension 20 at a point in the internal geometry) plus the Faddeev–Popov ghost bundle, together with the graded/ghost-corrected fiber supertrace over these bundles and their holonomy decomposition under \(K_6\times S^2\times S^1_Y\) .

 The connection \(\nabla\) is Levi-Civita (Nomizu form on \(K_6\) ); the endomorphism \(E\) (Weitzenböck curvature term) is bundle-specific and is pinned exactly:

 Bundle 
 Connection 
 \(E\) (endomorphism) 
 Spectrum / trace data (Killing-norm, Einstein center) 

 Scalar 
 LC (Nomizu) 
 \(E=0\) 
 domain \(C^\infty(K_6)\) ; spectrum \(C_2(p,q)/R_6^2\) 

 Vector / 1-form (Hodge) 
 LC 
 \(E=\mathrm{Ric}=\tfrac{5}{12}\mathrm{Id}\) 
 eigenvalue \(5/12\) , multiplicity 6; \(\mathrm{tr}\,E=5/2\) , \(\mathrm{tr}\,E^2=25/24\) 

 Graviton \(\mathrm{Sym}^2\) (full, dim 21) 
 LC 
 \((E_Lh)_{ab}=\mathrm{Ric}_{ac}h^c{}_b+\mathrm{Ric}_{bc}h^c{}_a-2R_{acbd}h^{cd}\) 
 Lichnerowicz spectrum \(1/6\) (×6), \(5/12\) (×6), \(7/6\) (×6), \(17/12\) (×2), \(5/3\) (×1, pure-trace mode) 

 Graviton TT \(\mathrm{Sym}^2_0\) (dim 20) 
 LC, transverse-traceless 
 same \(E_L\) 
 \(1/6\) (×6), \(5/12\) (×6), \(7/6\) (×6), \(17/12\) (×2); \(\mathrm{tr}\,E_L=40/3\) , \(\mathrm{tr}\,E_L^2=241/18\) 

 This is the object graded "DERIVED-GIVEN-E (terminal as structure)": the frozen geometry cleanly supplies \(L_{\rm grav}\) , its 4D massless spin-2 zero mode, and the Kaluza–Klein tower above it. It does not by itself certify the interacting quantum theory, and it does not derive general relativity or the matter content \(E\) — \(E\) is given (fixed by the observed matter content threaded through the bundle), not derived, and this given-E status is the single largest charged input the gate rests on.

 The fiber-weight ledger — the "91/13/65" numbers

 The physical object this gate's leading positive result depends on is the ghost-corrected fiber supertrace, and the corpus is explicit that two different objects use similarly-named numbers, so they must not be conflated:

 Bulk fiber weight (the one this gate uses): the graviton bulk fiber dimension in \(D=13\) is \(\mathrm{Sym}^2\) of the 13-dimensional tangent space, dimension 91 . The ghost fiber dimension is 13 , entering the supertrace with multiplicity \(-2\) (forced by BRST ghost-number grading, not a free choice). The ghost-corrected fiber weight is

 \[
\boxed{\ 91 - 2\cdot13 = 65\ }.
\]

 This \(65\) cross-checks exactly against the little-group count of physical massless graviton degrees of freedom in \(D=13\) : \(\dim\mathrm{Sym}^2_0(SO(11)) = \frac{11\cdot12}{2}-1 = 66-1=65\) . The sign and multiplicity of the ghost term are forced by BRST nilpotency — this is not an adjustable convention, and its correctness is what earns the "DERIVED-GIVEN-E" status for this leg.

 A different, graded object (NOT used for the bulk weight): the graded Casimir \(\gamma=\mathrm{Sym}^2(\mathrm{diag}(\mathbf1_{12},-1))\) gives \(\mathrm{tr}\,\gamma_{\rm grav}=67\) , \(\mathrm{tr}\,\gamma_{\rm ghost}=11\) , and a Block-A graded weight \(67-2\cdot11=45\) . This "67/11/45" triple belongs to the \(\mathbb{Z}_2\) -defect grading calculation, a structurally distinct object from the bulk \(91/13/65\) fiber weight above. The two must be kept separate in every place they appear in this dossier.

 \(K_6\) representation theory and Casimirs feeding the spectrum

 The Peter–Weyl decomposition organizes every KK mode: \(L^2(K_6,E_\mu)=\bigoplus_{(p,q)}V_{(p,q)}\otimes\mathrm{Hom}_{T^2}(V_{(p,q)},E_\mu)\) , with quadratic Casimir and dimension

 \[
C_2(p,q)=\frac{p^2+q^2+pq+3p+3q}{3},\qquad \dim(p,q)=\frac{(p+1)(q+1)(p+q+2)}{2}.
\]

 The lowest representations relevant to the gauge/gravity sectors: \((0,0)\) trivial, \(\dim1\) , \(C_2=0\) ; \((1,0)/(0,1)\) color triplet/antitriplet, \(\dim3\) , \(C_2=4/3\) ; \((1,1)\) the \(SU(3)\) adjoint (gluons, dimension 8), \(C_2=3\) exactly, zero-weight multiplicity \(2\) ; higher representations \((2,0),(2,1),(3,0),(2,2),(3,3)\) continue with exact rational or integer Casimirs up to \(C_2=15\) at \(\dim64\) . The KK mass formulas built on these Casimirs,

 \[
m^2_{(p,q),\rm vec}=\frac{C_2(p,q)+\Delta_{\rm vec}}{R_6^2},\qquad
m^2_{(p,q),\rm Dirac}=\frac{C_2(p,q)+\|\rho\|^2+\Delta_{\rm spin^c}}{R_6^2},\quad \|\rho\|^2=2,
\]

 set the Kaluza–Klein tower sitting above the 4D massless graviton zero mode that \(L_{\rm grav}\) supplies — the tower whose existence is part of "the geometry supplies the operator cleanly," and whose infinite extent above the cutoff is exactly the regime this gate cannot yet certify.

 Heat-kernel convention and the scalar/vector backbone

 The heat-kernel expansion convention used throughout is

 \[
K(t)\sim(4\pi t)^{-d/2}\sum_k a_{2k}\,t^k
\]

 (densities per unit volume), with the exact product rule for factorized manifolds \(a_{2k}(M_1\times M_2)=\sum_{i+j=k}a_{2i}(M_1)a_{2j}(M_2)\) . The certified scalar and vector backbone on \(K_6\) (Killing-norm, Einstein center) that every higher coefficient builds on:

 \[
a_2/a_0\,(\text{scalar}) = \frac{5}{12},\qquad a_4/a_0\,(\text{scalar}) = \frac{11}{120},\qquad \mathrm{tr}\,a_2\,(\text{vector, tangent bundle})=0,\qquad \mathrm{tr}\,a_4\,(\text{vector})=-\frac{47}{360}.
\]

 Exact sphere calibrations validate the machinery wherever it is independently checkable (two computational routes agree to \(\sim4\times10^{-14}\) ):

 \[
a_6(S^2)=\frac{4}{315},\qquad a_6(S^4)=\frac{74}{63},\qquad a_6(S^6)=\frac{1139}{63},\qquad a_6^{\rm conf}(S^6)=\frac{5}{63}.
\]

 The \(S^6\) row is a passed control precisely because \(K_6\ne S^6\) : the formula correctly reproduces the round unit sphere's known \(a_4=12\) , confirming the engine, while the physically relevant \(K_6\) curvature invariants above are verified to differ from the sphere values (e.g. \(|\mathrm{Riem}|^2_{K_6}=23/12\ne60=|\mathrm{Riem}|^2_{S^6\,\rm unit}\) ). A normalization-robust, Levi-Civita-immune scale-free ratio also holds: \(a_4/a_2^2=66/125\) .

 The \(S^1_Y/\mathbb{Z}_2\) orbifold factor — full precision

 \(\theta\mapsto-\theta\) on the circle, two isolated fixed points \(\theta=0,\pi\) . The active interval is \([0,\pi]\) with \(\mathrm{Vol}(S^1_Y/\mathbb{Z}_2)=\pi R_Y\) . Per-fixed-point \(a_0\) defect is \(+1/4\) for even (+) parity and \(-1/4\) for odd (−) parity. This orbifold structure is what supplies the smooth equivariant Donnelly defect referenced in the residuals discussion of the \(a_6\) ledger; it is emphasized here because a manifold-with-boundary framing of this same data is a wrong-object artifact — a global isometric reflection on a closed manifold is a different mathematical problem from a Neumann/Dirichlet boundary, and the two must not be conflated when reading the \(a_6\) chain.

 \(S^2\) sector

 \(ds^2_{S^2}=R_2^2(d\theta^2+\sin^2\theta\,d\phi^2)\) , \(\chi(S^2)=2\) , Dirac/Laplace eigenvalues \(\ell(\ell+1)/R_2^2\) with \(\ell\ge|N|/2\) and degeneracy \(2\ell+1\) , monopole sectors \(N=0\) (weak singlet), \(N=\pm1\) ( \(SU(2)_L\) doublet routing), \(N=\pm2\) ( \(W^\pm,W^0\) triplet), \(N\ge3\) higher KK thresholds.

 Summary — what this arena supplies to UQF-5C

 Taken together, the frozen 13D arena supplies UQF-5C with exactly three things, pinned at full precision and at all three layers: (1) a clean, gauge-fixed, ghost-corrected graviton operator \(L_{\rm grav}=-(\nabla^2+E)\) with a proved 4D massless spin-2 zero mode and BRST-forced fiber weight \(91-2\cdot13=65\) ; (2) a fully verified curvature dataset on \(K_6\) — \(|\mathrm{Riem}|^2/\mathrm{Scal}^2=23/75\) , \(\kappa=5/12\) , first-Bianchi machine-zero, the nonzero derivative sector \(|\nabla\mathrm{Riem}|^2=1/4\) — that feeds every heat-kernel coefficient including the still-open \(a_6\) ; and (3) a heat-kernel machine validated on spheres to \(\sim10^{-14}\) but whose graviton \(a_6\) leg is blocked at a precisely named computational stratum (the Gelfand–Tsetlin off-diagonal hopping term over the five Weyl-inequivalent weight classes of the root system pinned above). None of this arena data constitutes or supplies a non-perturbative UV completion; it is the complete, exact substrate on which that separate, unsolved question is posed.

 Construction I - the deep-root anchoring

 Fixed grade for this gate (stated once, held fixed throughout): REDUCED-TO-AXIOM / ANCHORED +1. 
UQF-5C asks whether the frozen thirteen-dimensional shape supplies a full, non-perturbative UV
completion of the interacting graviton — a quantum theory of gravity that stays consistent at and
above the cutoff, in the regime where the perturbative expansion in \(G_NE^2\) stops converging. This
section runs the three deep roots — Shape, Scale, Granularity — each completely, at all three layers
(× Stage, ⊕ Rulebook, ⊗ Actors) and at full precision, and then runs the four Layer-2 admissibility
screens (Invariance, Record-Interface, Causal-Order/target-blindness, Nonseparability) against the
resulting construction. Each root either forces a piece of the graviton construction, eliminates 
a class of would-be shortcuts around the wall, or exposes precisely where the wall stands. None of
the three roots closes the gate. Together they are exactly what the fixed grade records: two roots
(Shape, Scale) do forcing/eliminating/exposing work but supply no completion; the third (Granularity)
is the single named, irreducible axiom pair that earns the "+1" in REDUCED-TO-AXIOM.

 I.1 Shape — the carrier that supplies the operator, and draws the wall around it

 × Stage (complete, all four metric factors). The frozen active branch is the full three-layer
object
$$
\mathfrak{B} {\rm active}
=\underbrace{\big[\mathcal{M}_4\times K_6\times S^2\times S^1_Y/\mathbb{Z}_2\big]} {\times\ {\rm Stage}}
\ \oplus\
\underbrace{\big[\mathcal{F}^+ {\rm finite}\oplus\mathcal{C} {\rm admiss}\big]} {\oplus\ {\rm Rulebook}}
\ \otimes\
\underbrace{\big[\mathcal{E} {\rm matter}\oplus\mathcal{E} {\rm gauge}\oplus\mathcal{E} {\rm Higgs}\oplus\mathcal{E} {\rm proton}\big]} {\otimes\ {\rm Actors}},
$$
with \(K_6=SU(3)/T^2\) the complete \(A_2\) flag manifold and \(S^1_Y/\mathbb{Z}_2\) the active orbifold
domain (reflection \(\theta\mapsto-\theta\) , fixed points \(\theta=0,\pi\) ). Only the × layer carries
metric dimension,
$$
D=4+6+2+1=13,
$$
while the ⊕ and ⊗ layers are non-metric (0-dimensional) but load-bearing and can never be silently
dropped; a reading that keeps only the ×-factor and drops the gauge-fixing/ghost/scheme data is a
truncated object, and any graviton or heat-kernel statement built on that truncation is an artifact,
not a result — this is the operative meaning of "use the complete object" for this gate. The frozen
background carries audit-anchor branch/manifest hashes that certify which object was tested and
that it was not quietly retuned between computations; they carry no physics validation on their own.
The background is selected by Weyl-rigid admissibility and threshold-closure constraints, not
proven to be the unique geometry forced by first principles — this SELECTED-not-FORCED status is
carried through every claim below and is precisely why the gate's grade reads ANCHORED and never
DERIVED.

 On this complete Stage, the de-Donder (harmonic) gauge-fixed metric fluctuation \(h_{MN}\) produces the
Lichnerowicz-type Laplace operator
$$
L_{\rm grav}=-(\nabla^2+E),
$$
whose 4D massless spin-2 zero mode is the graviton, with a Kaluza–Klein tower above it indexed by the
 \(K_6\times S^2\times S^1_Y\) Peter–Weyl spectrum. This is genuine, unconditional Shape output at
DERIVED-GIVEN-E grade: the geometry supplies the operator (leg 5A) cleanly. What Shape does not 
supply, at any layer, is a statement about what happens once this operator's interactions are
resummed above the cutoff — Shape fixes what is being quantized , never whether the resulting
interacting theory is UV-finite or possesses a UV fixed point. 

 ⊕ Rulebook (complete: gauge-fix + ghosts + scheme + orbifold grading + readout basis). The
rulebook is what turns "a metric fluctuation" into a well-posed physical operator, and it is not
optional bookkeeping — gauge redundancy in \(h_{MN}\) forces the de-Donder gauge-fix and the
Faddeev–Popov ghost subtraction. The physical object is the ghost-corrected fiber supertrace
$$
\text{graviton}(91)-2\cdot\text{ghost}(13)=65,
$$
where \(91=\dim\mathrm{Sym}^2(\mathbb{R}^{13})=13\cdot14/2\) is the bulk fiber count on the D=13
tangent bundle and \(65=\dim\mathrm{Sym}^2_0(SO(11))=11\cdot12/2-1\) is the physical D=13 massless
little-group graviton degree-of-freedom count — two independently-motivated routes landing on the
same integer with zero free parameters. The sign \((-2)\) and the multiplicity of the ghost
contribution are forced by BRST nilpotency ( \(Q_{\rm BRST}^2=0\) ), not a free rulebook choice; this
is DERIVED-GIVEN-E, not a convention pick. The remaining rulebook data completing this layer:
 \(\overline{\rm MS}\) /heat-kernel scheme for the coefficient expansion \(K(t)\sim(4\pi
t)^{-d/2}\sum_ka_{2k}t^k\) ; the cubic-curvature/mass-dimension-6 Gilkey basis (order-46 terms) as the
declared readout basis for \(a_6\) (Gilkey Thm 3.3.1/4.8.16; Avramidi Ch. 4; Vassilevich eq. 4.29); and
the \(\mathbb{Z}_2\) orbifold parity on \(S^1_Y/\mathbb{Z}_2\) , correctly treated as an
equivariant/orbifold structure on a closed manifold rather than a manifold-with-boundary problem —
the twisted trace on \(S^1_R/\mathbb{Z}_2\) equals \(1\) exactly, \(t\) -independent, with no boundary
tower, and the smooth equivariant defect that is actually emitted is \(\tfrac12c_3^\gamma\) . A
structurally different graded object must never be substituted for the bulk 91/13/65 ledger: the
 \(\mathbb{Z}_2\) -defect grading trace \(\gamma=\mathrm{Sym}^2(\mathrm{diag}(\mathbf1_{12},-1))\) gives
 \(\mathrm{tr}\,\gamma_{\rm grav}=67\) , \(\mathrm{tr}\,\gamma_{\rm ghost}=11\) , graded weight
 \(67-2\cdot11=45\) — a distinct construction belonging only to the orbifold-defect computation.

 ⊗ Actors (complete: connection, endomorphism, operator domain, readout). The operator whose
spectrum would decide strong-coupling consistency is \(L_{\rm grav}=-(\nabla^2+E)\) on the graviton
 \(\mathrm{Sym}^2(T)\) bundle plus the FP ghost bundle, with the Lichnerowicz endomorphism
$$
(E_Lh) {ab}=\mathrm{Ric} {ac}h^c{} b+\mathrm{Ric} {bc}h^c{} a-2R {acbd}h^{cd}.
$$
At the Killing-form Einstein center \(\mathrm{Ric}=(5/12)\,\mathrm{Id}\) , the certified spectrum of
 \(E_L\) on the full \(\mathrm{Sym}^2\) (dim 21) is
$$
\tfrac16\,(\times6),\quad\tfrac{5}{12}\,(\times6),\quad\tfrac76\,(\times6),\quad
\tfrac{17}{12}\,(\times2),\quad\tfrac53\,(\times1,\ {\rm pure\ trace}),
$$
and on the physical transverse-traceless \(\mathrm{Sym}^2_0\) (dim 20) the same five eigenvalues minus
the trace mode, with certified traces \(\mathrm{tr}\,E_L=40/3\) , \(\mathrm{tr}\,E_L^2=241/18\) . These are
the exact ⊗-layer inputs any strong-coupling calculation on this geometry must start from — Shape
hands them over in full, with no missing digits. What the ⊗-Actors layer does not hand over is the
interacting resummation: the first-order (hopping) term on \(\mathrm{Sym}^2_0\) mixes the five
Weyl-inequivalent \(T^2\) weight classes through \(SU(3)\) Gelfand–Tsetlin off-diagonal matrix elements
between adjacent GT patterns on Peter–Weyl harmonic sections — exact in principle (a textbook
lowering-operator formula) but not yet enumerated . This is the named, non-fabricated computation
debt behind the graviton leg of \(a_6\) Route A, and it sits entirely inside the ⊗-Actors layer: it is
a missing matrix element, not a missing concept.

 I.1.1 What Shape forces. Three things are forced by the complete three-layer Stage/Rulebook/Actors
object, not chosen by hand: (i) the graviton propagates as \(\mathrm{Sym}^2_0(T)\) on the D=13 tangent
bundle with FP-ghost correction fixed by BRST — the operator-supply result (5A); (ii) the
ghost-corrected fiber weight \(91-2\cdot13=65\) , matching the D=13 little-group count
 \(\dim\mathrm{Sym}^2_0(SO(11))=65\) exactly, a cross-check with zero free parameters; (iii) the exact
curvature-invariant ratios feeding every heat-kernel coefficient in the expansion,
$$
\boxed{\ |\mathrm{Riem}|^2/\mathrm{Scal}^2=23/75=0.3066666666666667\ },\qquad
|\mathrm{Ric}|^2/\mathrm{Scal}^2=1/6,\qquad
\mathrm{Weyl}^2/\mathrm{Scal}^2=6/25,\qquad
\mathrm{Scal}/\mathrm{Ric}_i=6=\dim K_6,
$$
proved DERIVED by three independent, mutually target-blind routes — (1) direct construction from the
 \(SU(3)\) structure constants plus the naturally-reductive curvature formula, (2) a curvature-free
spectral heat-trace computation, (3) full Levi-Civita Riemann-tensor assembly over the
three-parameter squashed metric \(\vec u\) — and passing the target-blind first-Bianchi correctness
test at machine zero ( \(\approx2.5\times10^{-16}\) residual, against a buggy-build residual of exactly
 \(1/7\) in engine units / \(1/6\) in raw \(-B\) units). This is the test that caught and killed the \(\sim
31\%\) curvature bug ( \(31/147\to23/75\) , \(\kappa=7/12\to5/12\) ). The ratio is necessary input to every
 \(a_6,a_8,a_{10},\dots\) coefficient but is explicitly flagged necessary, not sufficient for UV
completion: it constrains the shape of the operator, never the existence of a fixed point above the
cutoff.

 I.1.2 What Shape eliminates. Shape eliminates any claim that the graviton fiber content or its
ghost partner is a free choice of representation: the D=13 tangent bundle, the FP ghost multiplicity,
and the sign of the ghost subtraction are all forced once de-Donder gauge and BRST are imposed. Shape
also eliminates, as permanent negative controls that must never resurface, the dead curvature values
from the retired buggy build: \(|\mathrm{Riem}|^2/R^2\) is never \(31/147\) , and \(|\mathrm{Riem}|^2(K_6)\) 
is never \(60\) (that is the unrelated round-unit \(S^6\) value); the Einstein constant \(\kappa\) is never
 \(7/12\) , only \(5/12\) ; and the retracted quartic derivative form \(256a^2(a^2-1)^2\) is dead — on \(K_6\) 
only \((\nabla\mathrm{Riem})^2\) survives in the derivative sector.

 I.1.3 What Shape exposes as unresolved. Shape exposes, rather than closes, the graviton
Gelfand–Tsetlin off-diagonal wall described above: the hopping-term matrix elements on
 \(\mathrm{Sym}^2_0\) are a well-posed but not-yet-enumerated computation, which is why the graviton leg
of \(a_6\) (Route A, Gilkey/Lichnerowicz) remains OWED even though the scalar backbone route
( \(a_6/a_2^3=7936/39375\) ) is banked across three or more independent engines, and even though the
narrower object \(a_6/a_0=-6373/630\) (a bulk graviton \(\mathrm{Sym}^2(T_6)\) Levi-Civita ratio) is
separately audit-certified in the completion-cascade ledger. Shape by itself never touches strong
coupling: even a fully enumerated, sign-correct \(a_6\) is one term in the unbounded operator ladder
 \(a_6<a_8<a_{10}<\cdots\) , and Shape supplies no mechanism to sum, resum, or bound that ladder.

 I.1.4 The scope-firewall certificate (Shape's terminal boundary). The single most important thing
Shape proves here is negative and terminal as a scope wall : one heat-kernel coefficient can never
settle strong-coupling consistency. The unbounded ladder \(a_6<a_8<a_{10}<\cdots\) is exactly why. This
certificate is why the dossier must refuse the signature mis-close " \(a_6\) is computed, therefore 5C is
closed" — \(a_6\) is input to the UV-completion question, it never is the answer. This is Shape
doing its most important job for this gate: drawing the wall precisely, at full 13D precision, rather
than papering over it with a partial object.

 I.2 Scale — locating exactly where the openness lives, and why no finite-grain shortcut exists

 × Stage / ⊗ Actors (the UV floor read off the complete geometry). The UV floor read directly off
the frozen geometry, via the Planck-mass normalization over the complete 9-dimensional internal
space \(X_{\rm int}=K_6\times S^2\times(S^1_Y/\mathbb{Z}_2)\) at \(D=13\) , is
$$
M_{\rm Pl}^2=M_ ^{D-2}\,\mathrm{Vol}(X_{\rm active}),\qquad
M_ ^{11}=\frac{M_{\rm Pl}^2}{\mathrm{Vol}(X_{\rm active})}=4.023836152402511\times10^{185}\ {\rm
GeV}^{11},
$$
$$
M_ =7.467050992135091\times10^{16}\ {\rm GeV},
$$
with \(\mathrm{Vol}(X_{\rm active})=3.704417261398702\times10^{-148}\ {\rm GeV}^{-9}\) and the ordinary
(non-reduced) Planck mass \(M_{\rm Pl}=1.220900000000000\times10^{19}\) GeV as the anchor input, not
an output. Separately, the unification scale closes at \(M_U\sim1.0\times10^{16}\) GeV (residual on the
inverse-coupling equality \(9.6\times10^{-11}\) , well inside the propagated PDG band \(\sim10^{-3}\) ),
fixing the natural compactification radius
$$
R_0\equiv(2\pi M_U)^{-1}=1.591549430918954\times10^{-17}\ {\rm GeV}^{-1}.
$$
Both \(M_*\) and \(R_0\) are explicitly flagged as derived geometry read-offs, not new independent
inputs and not a closure * of this gate — \(M_*\) follows from the anchor \(M_{\rm Pl}\) plus the derived
internal volume, nothing more.

 ⊕ Rulebook (where the wall is defined). The strong-coupling wall is defined at the rulebook
layer: it is the regime above this floor where the perturbative expansion in \(G_NE^2\) ceases to
converge — where the \(\overline{\rm MS}\) /heat-kernel scheme itself stops being a well-posed expansion
parameter for the interacting graviton. Scale is the root that names precisely where the wall
stands, even though (as shown next) it supplies no way across it.

 I.2.1 What Scale forces. Scale forces the admission that \(M_*\sim7.47\times10^{16}\) GeV
(equivalently \(M_U\sim1.0\times10^{16}\) GeV) is where the perturbative graviton expansion must be
replaced by something non-perturbative — this is not optional, it is where \(G_NE^2\sim1\) on this
geometry. Scale also forces, via the banked Layer-2 screen T-DEEP, the color/gauge-vs-gravity
separation: the compactification radius \(R_0\) is built purely from gauge-coupling unification data,
with zero gravitational input , giving
$$
R_0/\ell_{\rm Planck}\sim194.
$$
Because \(R_0\) carries no graviton content, Scale forces the consequence that any finite-grain
dissolution attempted at UQF-5C relocates onto UQF-9's isolated sub-target P0 — "does gravity own an
intrinsic shortest length, or does it inherit the color scale \(R_0\) ?" — rather than closing here.
This is a DERIVED, banked structural fact about the shared geometry, not a claimed dissolution of the
wall itself: 5C has no finite-grain shortcut internal to this gate. 

 I.2.2 What Scale eliminates. Scale eliminates the temptation to read \(M_*\) or \(M_U\) as a "UV
completion scale" in the string-theory sense (a scale at which new UV-complete degrees of freedom
appear with a known, finite S-matrix) — nothing in the frozen geometry supplies such degrees of
freedom at \(M_*\) ; it is only the scale at which the effective perturbative expansion breaks down.
Scale also eliminates the refuted dimensionful magnitude \(-2.818\times10^{94}\) GeV \(^6\) as a
would-be UV input: that number was R2-contaminated, scheme-anchored, and decisively ill-posed at
odd \(D=13\) , because at odd spacetime dimension there is no finite local \(t^0\) heat-kernel slot for a
bulk magnitude of this kind — it sits at the half-integer \(\zeta\) -pole \(s=7/2\) , which vanishes in
dimensional regularization with no log/anomaly slot to carry a finite answer. (A separately re-run
Bianchi-exact dimensionful bulk value, \(-2.995681680\times10^{94}\) GeV \(^6\) , is recorded in the
audit-cascade ledger purely as a consistency coefficient — it is never gap-closing and remains
route-inconsistent; it does not revive the refuted magnitude.) This is a genuine, decision-grade
refutation — a reached verdict, not a hole — and Scale is the root that shows why no dimensionful
bulk number can exist there: the correct owed object is the finite trace \(\mathrm{tr}[a_6(L_{\rm
grav})]\) , not a GeV \(^6\) magnitude, and that trace is itself OPEN and must not be fabricated.

 I.2.3 What Scale exposes. Scale exposes that the openness of 5C is not a numerical gap a sharper
calculation could close — it is a qualitative regime change (perturbative-series divergence) that
no value of \(a_6\) , however exactly computed, addresses. Scale is where the dossier must resist the
strongest temptation to over-claim: a clean UV floor number ( \(M_*\) , \(M_U\) , \(R_0\) ) reads, superficially,
like "the completion scale is known," but Scale only fixes where the expansion parameter grows large,
never what replaces the expansion above it .

 I.3 Granularity — the one axiom that earns the "+1"

 ⊕ Rulebook (the axiom, stated precisely). Granularity here is AXIOM-COSTFLOOR : an irreducible
quantum of cost/action , explicitly not a smallest length , applied Lorentz-invariantly. This
single axiom is responsible for the entire "+1" in REDUCED-TO-AXIOM/ANCHORED +1. Its physical
grounding is established, MEASURED-ANCHOR physics external to this specific 13D construction:
- Margolus–Levitin bound: \(\tau\geq\pi\hbar/2E\) — a minimum time to evolve to an orthogonal
 state, set by available energy;
- Landauer bound: \(\Delta E\geq k_BT\ln2\) — minimum energy cost of an irreversible bit erasure;
- Bekenstein bound: \(S\leq2\pi k_BRE/\hbar c\) — maximum entropy in a bounded region of given
 energy and radius.

 Together these establish that action/information processing carries an irreducible cost floor
 \(\Delta_0>0\) in any consistent quantum-plus-gravity setting, independent of this specific geometry.

 ⊗ Actors (the Lorentz-scalar structure that completes the axiom pair). The floored quantity is
proved to transform as a Lorentz scalar (theorem T-LI): the cost floor is defined on proper
time , not on any coordinate length or preferred spatial lattice, so no preferred frame is singled
out by the floor. This is the second half of the anchor pair — {Δ₀ > 0, Lorentz-scalar proper-time
floor} — on which the published row is ANCHORED. Because the floor is a Lorentz scalar rather than
a minimum spatial length, it survives the Invariance screen (§I.5 below) without contradiction: it
is a cost floor on proper time, never a spatial-lattice floor that would break boost invariance.

 I.3.1 What Granularity forces. The axiom, once adopted, forces exactly one consequence: it
 dissolves exactly one UV-divergence class — the \(a\to0\) runaway behavior of the unbounded
higher-coefficient tower \(\{a_8,a_{10},\dots\}\) , whose divergence as the regulator parameter shrinks
to zero is cut off by the irreducible cost quantum. This is the one place Granularity does forced,
positive work on the gate.

 I.3.2 What Granularity eliminates. Nothing beyond the single divergence class above. Stated with
maximal honesty: ten walls remain untouched by this axiom, including the finite \(a_6\) coefficient
itself. \(a_6\) is a finite, well-defined heat-kernel coefficient — it was never divergent, so the
cost-floor axiom has nothing to act on there; the graviton leg of \(a_6\) stays OWED at the
Gelfand–Tsetlin matrix-element stratum regardless of whether AXIOM-COSTFLOOR is adopted.
 DISSOLVED ≠ SOLVED: no fixed point is exhibited, no value of \(a_6\) is supplied, no constructive
UV completion is produced by this axiom. It removes one specific pathology — the runaway tail of the
coefficient tower — not the strong-coupling wall itself.

 I.3.3 What Granularity exposes. Granularity exposes its own status honestly: it is
 AXIOM-OPEN / not atomic — a named floor that can be relocated , not an eliminated one. It is not
claimed to be forced by the geometry; it is a declared, physically-motivated axiom (grounded in
Margolus–Levitin/Landauer/Bekenstein) adopted to regularize one divergence class. This is exactly why
the grade is ANCHORED +1 rather than DERIVED: the "+1" is this one explicit, irreducible axiom, named
and isolated rather than hidden inside a derivation. SELECTED ≠ FORCED and ANCHORED ≠ DERIVED apply
here in their sharpest form — anchoring the row on {Δ₀>0, Lorentz-scalar proper-time floor} is not a
derivation of the completion and must never be read as one.

 I.4 What the three roots jointly deliver — and jointly fail to deliver

 Running Shape, Scale, and Granularity together, each at full three-layer precision, produces a
coherent and honest picture. Shape supplies the operator and its forced fiber-weight/curvature data
cleanly (5A, DERIVED-GIVEN-E) while certifying, as a structural theorem, that no single coefficient
in its associated heat-kernel tower can ever settle strong coupling. Scale names the precise regime
(above \(M_*\approx7.467\times10^{16}\) GeV / \(M_U\sim1.0\times10^{16}\) GeV) where the perturbative
expansion in \(G_NE^2\) breaks down, and proves via T-DEEP that the compactification radius \(R_0\) 
carries zero gravitational content, so that no finite-grain shortcut through Scale exists without
relocating the question onto UQF-9's P0 sub-target. Granularity supplies the one legitimate axiom —
AXIOM-COSTFLOOR, paired with the Lorentz-scalar proper-time floor theorem T-LI, grounded in
Margolus–Levitin/Landauer/Bekenstein — that dissolves exactly one divergence class while leaving ten
others, including the finite \(a_6\) obligation itself, untouched. None of the three roots, singly or
jointly, supplies a non-perturbative construction (a truncation-independent fixed point, or a proof
that none exists) for the interacting graviton above the cutoff. That missing construction is
exactly, and only, UQF-9. This is why the roll-up is carried as OPEN (global wall) on the
per-gate ledger even while the public-board pill correctly reads ANCHORED +1 on the named axiom
pair: the two descriptions are the same physics, differing only in whether the axiom-anchored partial
result or the unclosed constructive wall is foregrounded — the dossier states both, faithfully,
rather than silently picking one.

 I.5 The four Layer-2 admissibility screens

 Invariance (Physical Equivalence). Gauge redundancy in the metric fluctuation \(h_{MN}\) is not a
bookkeeping nuisance but a Layer-2 constraint: it forces the de-Donder gauge-fix and the
Faddeev–Popov ghost subtraction, and the only admissible physical coefficient is the
frame-independent, ghost-corrected object \(91-2\cdot13=65\) . Any candidate "graviton weight" that
skips the ghost subtraction (the bare \(91\) ) or substitutes the distinct graded \(\gamma\) -trace pair
 \(67/11\) (which belongs only to the \(\mathbb{Z}_2\) -defect construction) fails this screen and must be
rejected. Separately, the Granularity axiom passes Invariance precisely because the cost floor is
proved to act on a Lorentz scalar (proper time), not on a coordinate-dependent spatial length — a
spatial-lattice floor would fail Invariance outright by picking a preferred frame, while the
proper-time floor does not.

 Record-Interface (reproducibility). The curvature ledgers, heat-kernel runs, and the refutation
of the \(-2.818\times10^{94}\) GeV \(^6\) magnitude are built to be independently reproducible: the
first-Bianchi machine-zero test ( \(2.5\times10^{-16}\) ) is a target-blind, rerunnable correctness
criterion (it is what caught the \(\sim31\%\) curvature bug); the sphere cross-checks
 \(a_6(S^2)=4/315\) , \(a_6(S^4)=74/63\) , \(a_6(S^6)=1139/63\) , \(a_6^{\rm conf}(S^6)=5/63\) agree between
independent routes to \(\sim4\times10^{-14}\) , validating the heat-kernel machinery everywhere it can 
be checked against known closed manifolds. This screen is satisfied for the geometric-input side of
the ledger. It is explicitly not yet satisfied for the graviton \(a_6\) trace itself
( \(\mathrm{tr}[a_6(L_{\rm grav})]\) is OPEN, the claimed coefficient \(C\sim-6.39\) is not reproduced and
must not be fabricated), nor for the color factor \(124/315\) , which is DOWNGRADED to
DERIVED-PENDING-INDEPENDENT-TARGET-BLIND-REPRODUCTION because it was produced by the same
metric-selected engine it is meant to validate (metric selected at \(\mathrm{Scal}_{K_6}=7.5\) ) — an
independent-engine reproduction is the named work package (H5) that would close this screen. The
audit-cascade completion ledger (e.g. AUD-0059's exact-rational assembly
 \(\tfrac12(-953329/1260)+(-337361/840)=-491353/630\) ) is Record-Interface satisfied at the
audit-certified layer , but its ceiling is explicitly AUDIT-CLOSED, physics-OPEN — a captured,
multi-route-verified terminal log is not the same thing as an independent reproduction on a
structurally different engine, and the two must not be conflated.

 Causal-Order / target-blindness. The three independent routes proving
 \(|\mathrm{Riem}|^2/R^2=23/75\) (SU(3) structure constants; curvature-free spectral heat-trace; full
Levi-Civita Riemann tensor) are target-blind by construction: the resulting value
 \(23/75=0.30667\) matches neither of the two candidate targets that had previously circulated (the
buggy engine value \(0.2109\) nor a separately-quoted firewall value \(0.0667\) ), direct evidence the
result was not reverse-engineered to a desired answer. This screen is also the discipline that flags
the refuted \(-2.818\times10^{94}\) GeV \(^6\) number for exactly the opposite reason — it was 
scheme-anchored/target-loaded — and that rules out treating a captured terminal log as an independent
reproduction. Target-blindness is explicitly the standard the H4/H5 work packages (computing the
d=13 graviton-minus-ghost \(a_6\) vector; reproducing \(124/315\) on a structurally independent engine)
must meet before any upgrade is legitimate, and a scheme reverse-engineered to a wanted magnitude is
named as a failure mode to refuse outright (the \(\kappa^3/\pi\) kill-test).

 Nonseparability. A single heat-kernel coefficient cannot, by itself, compose into a statement
about strong-coupling behavior — \(a_6\) sits inside an unbounded, non-separable ladder
 \(a_6<a_8<a_{10}<\cdots\) , and no finite subset of that ladder determines convergence or divergence of
the full interacting expansion. This is precisely the structural content of the scope-firewall
certificate in §I.1.4: the screen is failed by design for any claim that isolates \(a_6\) from the
tower and calls it decisive, which is exactly why "computing \(a_6\) closes 5C" is named as the gate's
signature mis-close. The honest, passing use of Nonseparability here is defensive: it is the guard
that keeps the dossier from over-claiming on the strength of one finite, even fully-verified,
coefficient — and it is the same structural fact that forbids treating the audit-certified
completion-cascade chain (Layer B: the bulk graded \(a_6=-953329/1260\) , the physical defect
 \(a_6=-7226/35\) , the \(\mathbb{Z}_2\) equivariant defect \(-337361/840\) , the AUD-0059 assembly
 \(-491353/630\) ) as anything more than sharpened, banked computational progress toward a linearized
certificate conditional on UQF-9 — never as a proxy for the strong-coupling completion itself.

 I.6 Summary of the deep-root verdict for Construction I

 Shape (all three layers, full precision) forces the operator, its ghost-corrected fiber weight
 \(91-2\cdot13=65=\dim\mathrm{Sym}^2_0(SO(11))\) , and the exact curvature ratios
( \(|\mathrm{Riem}|^2/\mathrm{Scal}^2=23/75\) , \(|\mathrm{Ric}|^2/\mathrm{Scal}^2=1/6\) ,
 \(\mathrm{Weyl}^2/\mathrm{Scal}^2=6/25\) ) that feed every coefficient in the heat-kernel expansion,
while drawing a hard scope wall around any single coefficient. Scale locates the wall precisely at
 \(M_*=7.467050992135091\times10^{16}\) GeV (equivalently \(M_U\sim1.0\times10^{16}\) GeV,
 \(R_0=1.591549430918954\times10^{-17}\ {\rm GeV}^{-1}\) ) and proves, via the T-DEEP screen
( \(R_0/\ell_{\rm Planck}\sim194\) , \(R_0\) built from pure gauge data with zero gravitational input),
that no finite-grain shortcut exists without relocating the question to UQF-9's P0 sub-target.
Granularity supplies the single legitimate axiom — AXIOM-COSTFLOOR, paired with the Lorentz-scalar
proper-time floor theorem T-LI, grounded in Margolus–Levitin/Landauer/Bekenstein — that dissolves
exactly one divergence class (the \(a\to0\) tail of \(\{a_8,a_{10},\dots\}\) ) while leaving the finite
 \(a_6\) obligation and nine other named walls untouched. The four Layer-2 screens are satisfied on the
geometric-input side (Invariance forces the ghost-corrected \(65\) ; Record-Interface reproduces the
sphere/Bianchi checks to machine precision; Causal-Order/target-blindness certifies the \(23/75\) 
result was not reverse-fit; Nonseparability is the very reason the scope firewall exists) and are
explicitly unsatisfied on the graviton-trace side (the \(a_6\) graviton leg, the \(124/315\) 
independent-reproduction requirement, and the missing constructive fixed point). The joint verdict of
the three roots is exactly the fixed grade: REDUCED-TO-AXIOM / ANCHORED +1 on the conditional
axiom pair {Δ₀ > 0, Lorentz-scalar proper-time floor} — a genuine, banked partial result, with the
constructive UV-completion wall (UQF-9) standing open and unclosed by any lever internal to this gate.

 Construction II - the full derivation

 II.1 Setting up the object: the frozen 13D arena carrying the graviton

 The construction begins by pinning, at all three layers, the exact object the graviton
operator lives on. This is not a formality — every heat-kernel coefficient computed below is
only as meaningful as the operator it is the coefficient of , and the operator is only as
meaningful as the background it is built from.

 × Stage (metric geometry, D = 4 + 6 + 2 + 1 = 13). The frozen active branch is

 \[
\mathfrak{B}_{\rm active} \;=\; \underbrace{\big[\mathcal{M}_4\times K_6\times S^2\times S_Y^1\big]}_{\times\ \text{Stage}} \;\oplus\; \underbrace{\big[\mathcal{F}^+_{\rm finite}\oplus\mathcal{C}_{\rm admiss}\big]}_{\oplus\ \text{Rulebook}} \;\otimes\; \underbrace{\big[\mathcal{E}_{\rm matter}\oplus\mathcal{E}_{\rm gauge}\oplus\mathcal{E}_{\rm Higgs}\oplus\mathcal{E}_{\rm proton}\big]}_{\otimes\ \text{Actors}},
\]

 with \(K_6 = SU(3)/T^2\) the full \(A_2\) -type flag manifold and \(S_Y^1/\mathbb{Z}_2\) the active
orbifold interval. Metric dimension is carried only by the \(\times\) -Stage factor:
 \(D = 4+6+2+1 = 13\) . The \(\oplus\) and \(\otimes\) layers are non-metric but load-bearing, and a
 \(\times\) -only reading of the graviton target is an incomplete object — this is the discipline
that keeps the construction from silently truncating to a symmetric-space toy.

 The internal metric is
$$
ds^2_{K_{\rm gauge}} = R_6^2\,ds^2_{K_6}(u_1,u_2,u_3) + R_2^2\,ds^2_{S^2} + R_Y^2\,d\theta^2,
$$
with Weyl-rigid moduli \(\vec u=(u_1,u_2,u_3)\in[1/2,3/2]^3\) and chamber-center witness
 \(u_1=u_2=u_3=1\) (off-chamber values fail admissibility and are eliminated by the selector). The
frozen background is SELECTED by the constraint set, not proven to be uniquely forced — this
qualifier is carried at every step below, because the graviton operator is built on this specific
shape.

 ⊕ Rulebook (scheme/convention/boundary/projector/grading). The gauge-fixing scheme is
de-Donder (harmonic) gauge on the metric fluctuation, with the accompanying Faddeev–Popov ghost
sector; the loop/heat-kernel scheme is \(\overline{\rm MS}\) ; the boundary condition on
 \(S_Y^1/\mathbb{Z}_2\) is the \(\mathbb{Z}_2\) orbifold parity \(\theta\mapsto-\theta\) with fixed
points at \(\theta=0,\pi\) ; the readout basis for the weight-6 invariant is the cubic-curvature /
mass-dimension-6 Gilkey basis (the standard \(\sim46\) -term local invariant basis for \(a_6\) ).

 ⊗ Actors (the operator whose spectrum decides consistency). The operator under study is the
Lichnerowicz-type Laplacian
$$
L_{\rm grav} = -(\nabla^2+E)
$$
acting on the graviton bundle \(\mathrm{Sym}^2(T)\) , together with the Faddeev–Popov ghost bundle
that de-Donder gauge-fixing forces into existence. "Actors" also names the graded/ghost-corrected
fiber supertrace and the holonomy decomposition of these bundles under
 \(K_6\times S^2\times S^1_Y\) .

 Two curvature normalizations are used side by side and must never be mixed at the level of
 absolute numbers (only their dimensionless ratios agree): the frozen \(R_6\) -metric normalization
(physical radius, \(\mathrm{Ric}_i=1/(2R_6^2)\) , \(\mathrm{Scal}=3/R_6^2\) , dimensionful in
 \(\mathrm{GeV}^2\) ), and the Killing-form normal metric \(g=(-B)|_{\mathfrak m}\) ,
 \(B(X,Y)=6\,\mathrm{Tr}(XY)\) , evaluated at the symmetric chamber center, which is where the exact
rational invariants live: \(\mathrm{Ric}_i=5/12\) , \(\mathrm{Scal}=5/2\) . The bridge identity used
throughout is \(\mathrm{Scal}/\mathrm{Ric}_i = 6 = \dim K_6\) in both normalizations, and
 \(|\mathrm{Ric}|^2/\mathrm{Scal}^2=1/6\) , \(|\mathrm{Riem}|^2/\mathrm{Scal}^2=23/75\) agree in both.

 II.2 Step 1 — gauge-fixing the metric fluctuation and identifying the physical operator

 Write the metric as background plus fluctuation, \(g_{MN}=\bar g_{MN}+h_{MN}\) , on the 13D frozen
background. De-Donder (harmonic) gauge-fixing,
 \(\bar\nabla^M h_{MN} - \tfrac12\bar\nabla_N h = 0\) , removes the diffeomorphism redundancy and turns
the linearized Einstein–Hilbert action into a minimal (Laplace-type) second-order operator acting
on \(h_{MN}\in\Gamma(\mathrm{Sym}^2 T^*)\) :
$$
L_{\rm grav} h = -(\nabla^2 + E)\,h,
$$
where \(\nabla^2=\nabla^M\nabla_M\) is the Bochner Laplacian built from the Levi-Civita connection
of \(\bar g\) , and \(E\) is the curvature endomorphism (the Lichnerowicz operator's non-Laplacian
piece) acting on \(\mathrm{Sym}^2(T)\) :
$$
(E_L h) {ab} = \mathrm{Ric} {ac}h^c{} b + \mathrm{Ric} {bc}h^c{} a - 2R {acbd}h^{cd}.
$$
Gauge-fixing is not optional decoration: it is what makes \(L_{\rm grav}\) of minimal Laplace type
in the first place (without it, the graviton kinetic operator is degenerate along
diffeomorphism directions). The de-Donder choice forces a compensating Faddeev–Popov ghost
sector — a pair of anticommuting vector fields with kinetic operator \(-(\nabla^2+\mathrm{Ric})\) on
 \(T^*\) — so that the resulting path integral measure remains BRST-invariant. This is the origin of
the ghost bundle referenced in the fiber-weight ledger below: it is not an add-on but a structural
requirement of the gauge choice.

 On the compact factor \(K_6\times S^2\times S_Y^1/\mathbb{Z}_2\) , at the Killing-form chamber
center, the certified endomorphism spectrum for the relevant bundles is:

 Bundle 
 \(E\) 
 Spectrum of \(E\) (eig \(\times\) mult) 
 \(\mathrm{tr}\,E\) 
 \(\mathrm{tr}\,E^2\) 

 Scalar 
 \(0\) 
 \(0\) 
 \(0\) 
 \(0\) 

 Vector (1-form) 
 \(\mathrm{Ric}=(5/12)\mathrm{Id}\) 
 \(5/12\ (\times 6)\) 
 \(5/2\) 
 \(25/24\) 

 Graviton \(\mathrm{Sym}^2\) (full, dim 21) 
 \(E_L\) 
 \(\tfrac16(\times6),\ \tfrac{5}{12}(\times6),\ \tfrac76(\times6),\ \tfrac{17}{12}(\times2),\ \tfrac53(\times1)\) 
 — 
 — 

 Graviton TT \(\mathrm{Sym}^2_0\) (dim 20) 
 \(E_L\) 
 \(\tfrac16(\times6),\ \tfrac{5}{12}(\times6),\ \tfrac76(\times6),\ \tfrac{17}{12}(\times2)\) 
 \(40/3\) 
 \(241/18\) 

 The pure-trace mode of the full \(\mathrm{Sym}^2\) sits at eigenvalue \(5/3\) with multiplicity \(1\) ;
removing it leaves the transverse-traceless \(\mathrm{Sym}^2_0\) (dim 20) with the certified traces
 \(\mathrm{tr}\,E_L = 40/3\) , \(\mathrm{tr}\,E_L^2 = 241/18\) — these are the inputs that feed every
downstream heat-kernel coefficient for the graviton sector. On the vector bundle the curvature
2-form is \(\Omega_{ab}=\) Riemann, giving \(\mathrm{tr}(\Omega_{ab}\Omega^{ab})=-|\mathrm{Riem}|^2\) ,
which is the vector-sector analogue used in the ghost's own heat-kernel expansion (the ghost
transforms in the vector/1-form bundle with \(E=\mathrm{Ric}\) ).

 Result of Step 1 (DERIVED-GIVEN-E, terminal as structure): the frozen 13D geometry supplies a
well-posed, minimal-type Laplace operator \(L_{\rm grav}=-(\nabla^2+E)\) whose 4D zero mode is the
massless spin-2 graviton, with a Kaluza–Klein tower of massive spin-2 modes above it indexed by
the Peter–Weyl / spherical-harmonic spectrum on \(K_6\times S^2\times S_Y^1\) . The allowed reading
is exactly this: the geometry supplies the operator . It does not certify the resulting
quantum theory, and it does not derive general relativity or derive \(E\) itself — \(E\) is built
from the already-fixed curvature of the selected background, which is an input to this step, not
an output of it.

 II.3 Step 2 — the ghost-corrected fiber weight, forced by BRST nilpotency

 The physical content of a one-loop (or heat-kernel) graviton computation is never the raw
graviton trace alone: gauge redundancy means the graviton path integral must be divided by the
diffeomorphism volume, and the Faddeev–Popov procedure implements this by subtracting twice 
the ghost determinant (once for each of the two anticommuting ghost fields \(c^M\) , \(\bar c_M\) ) from
the graviton determinant. This is not a modeling choice; it is forced by BRST nilpotency
( \(Q_{\rm BRST}^2=0\) ) acting on the gauge-fixed Hilbert space, which is exactly the identity that
guarantees the ghost subtraction reproduces the physical (gauge-invariant) cohomology
 \(\mathcal{H}_{\rm phys}\) .

 Counting fiber dimensions in the full \(D=13\) bulk:

 Graviton bulk fiber dimension in \(D=13\) : \(\dim\mathrm{Sym}^2(\mathbb{R}^{13}) = \frac{13\cdot14}{2} = 91\) .

 Ghost fiber dimension : the ghost lives in the tangent bundle, \(\dim T = 13\) .

 BRST-forced combination : graviton \(-\ 2\times\) ghost \(= 91 - 2\cdot13 = 91-26 = 65\) .

 This ghost-corrected fiber weight, 65 , is not an arbitrary linear combination: the coefficient
" \(-2\) " on the ghost term is fixed by the requirement that the Faddeev–Popov determinant exactly
cancels the unphysical (pure-gauge + trace) polarizations of the naive \(\mathrm{Sym}^2\) graviton,
leaving precisely the little-group content of a massless spin-2 particle in \(D=13\) . This is
confirmed by an independent group-theoretic cross-check: the physical, on-shell massless graviton
polarizations in \(D\) spacetime dimensions transform in the traceless symmetric tensor
representation of the \(D-2\) little group \(SO(D-2)\) , i.e.
$$
\dim\mathrm{Sym}^2_0\big(SO(D-2)\big) = \frac{(D-2)(D-1)}{2} - 1.
$$
At \(D=13\) : \(\dim\mathrm{Sym}^2_0(SO(11)) = \frac{11\cdot12}{2}-1 = 66-1 = 65\) .

 The two numbers agree exactly: \(91-2\cdot13=65=\dim\mathrm{Sym}^2_0(SO(11))\) . This is a nontrivial
consistency check — one count comes from the BRST/Faddeev–Popov subtraction on the bulk fiber
(a statement about the gauge-fixed path-integral measure), the other from the on-shell 
representation theory of a massless spin-2 field (a statement about physical polarizations) — and
they land on the same integer. This is the " \(65 = 91-26\) " identity referenced throughout the
ledger, and it is what makes the ghost-corrected fiber weight a forced , not a chosen, number.

 Negative control (do not confuse objects): a different graded object,
 \(\gamma=\mathrm{Sym}^2(\mathrm{diag}(\mathbf{1}_{12},-1))\) , gives \(\mathrm{tr}\,\gamma_{\rm grav}=67\) 
and \(\mathrm{tr}\,\gamma_{\rm ghost}=11\) , with Block-A graded weight \(67-2\cdot11=45\) . This
 \(67/11/45\) triple belongs to the \(\mathbb{Z}_2\) -defect grading (Section II.6 below) and must never
be substituted for the bulk \(91/13/65\) triple — they are different objects computing different
things (a \(\mathbb{Z}_2\) -twisted trace versus an untwisted bulk fiber count), and conflating them
is exactly the kind of error a target-blind reviewer is positioned to catch.

 Result of Step 2 (DERIVED-GIVEN-E): the ghost-corrected fiber weight is \(65\) , forced by BRST
nilpotency and cross-checked against the independent little-group representation count
 \(\dim\mathrm{Sym}^2_0(SO(11))=65\) . This is a genuine, load-bearing, non-fabricated result — but it
is a statement about the fiber content of the operator, not yet a statement about any
heat-kernel coefficient or about strong-coupling consistency.

 II.4 Step 3 — the curvature data that feed the heat-kernel expansion

 The heat-kernel expansion of \(L_{\rm grav}\) (and of the ghost operator) is
$$
K(t) \sim (4\pi t)^{-d/2}\sum_{k\ge 0} a_{2k}\,t^k,
$$
with \(a_{2k}\) built from the curvature and its covariant derivatives, integrated against the
bundle endomorphism \(E\) and curvature 2-form \(\Omega\) via the standard Gilkey/Seeley–DeWitt local
formulas. The coefficients relevant here are read off the fixed \(K_6=SU(3)/T^2\) background at the
Killing-form chamber center.

 Root system and tangent decomposition ( \(A_2\) ). Simple roots in the Cartan basis
 \((h_1,h_2,h_3)\) with \(h_1+h_2+h_3=0\) : \(\alpha_1=(1,-1,0)\) , \(\alpha_2=(0,1,-1)\) ,
 \(\alpha_1+\alpha_2=(1,0,-1)\) . Positive roots \(\{\alpha_1,\alpha_2,\alpha_1+\alpha_2\}\) ; Weyl group
 \(S_3\) (order 6); half-sum \(\rho=\tfrac12\sum_{\alpha>0}\alpha=(1,0,-1)\) , \(\|\rho\|^2=2\) . The
tangent space decomposes as \(T(K_6)=\mathfrak m_1\oplus\mathfrak m_2\oplus\mathfrak m_3\) , each a
real 2-plane carrying one positive root.

 Curvature scalars (Killing-form normal metric, \(\vec u=(1,1,1)\) , exact rationals): 
$$
\dim K_6=6,\qquad \mathrm{Ric}_i=\frac{5}{12},\qquad \mathrm{Scal}=\frac{5}{2},\qquad
\mathrm{Scal}^2=\frac{25}{4},
$$
$$
|\mathrm{Ric}|^2=\frac{25}{24},\qquad |\mathrm{Riem}|^2=\frac{23}{12},\qquad
\frac{|\mathrm{Riem}|^2}{\mathrm{Scal}^2}=\frac{23}{75}=0.30\overline{6},\qquad
\frac{|\mathrm{Ric}|^2}{\mathrm{Scal}^2}=\frac16,\qquad \mathrm{Weyl}^2/\mathrm{Scal}^2=\frac{6}{25}.
$$
The Einstein constant read off the corrected Ricci eigenvalue is \(\kappa=5/12\) 
( DERIVED-GIVEN-E ; the earlier buggy value \(\kappa=7/12\) is retired).

 The \(|\mathrm{Riem}|^2/R^2=23/75\) result is itself a proved input , established target-blind by
three independent routes: (1) direct \(SU(3)\) structure-constant plus naturally-reductive curvature
formula; (2) a curvature-free spectral heat-trace route; (3) full Levi-Civita Riemann tensor
computed from the 3-parameter chamber metric. In trip-unit normalization these routes give
 \(R=15/2\) (or \(15\) in the alternate convention), \(|\mathrm{Ric}|^2=75/8\) , \(|\mathrm{Riem}|^2=69/4\) ,
 \(\mathrm{Weyl}^2=27/2\) , Einstein value \(R/6=5/4\) , with \(\mathrm{Weyl}^2\) forced by the standard
 \(d=6\) curvature decomposition (Riemann = Weyl + Ricci-trace + scalar-trace pieces). This is a
necessary-but-not-sufficient input win: it feeds every downstream \(a_6\) computation, but it does
not by itself close anything about strong coupling.

 Target-blind correctness criterion — first Bianchi identity. The corrected curvature tensor
satisfies the first Bianchi identity to residual \(\approx 2.5\times10^{-16}\) (machine zero); the
earlier buggy build gave a first-Bianchi residual of exactly \(1/7\) (engine normalization) or
 \(1/6\) (raw \(-B\) -form units) — a finite, wrong number that this target-blind test caught, which is
also what revealed the \(31/147\to23/75\) and \(\kappa=7/12\to5/12\) corrections. Passing the Bianchi
identity does not validate the \(a_6\) computation itself; it validates that the curvature
tensor entering it is a genuine Riemann tensor.

 Cubic and derivative curvature invariants (Killing-form center, exact rationals): 
$$
K_1\equiv R_{ab}{}^{cd}R_{cd}{}^{ef}R_{ef}{}^{ab}=8\,\mathrm{tr}(R_{\rm op}^3)=-\frac{113}{72},\qquad
K_2\equiv R_{abcd}R_{aecf}R_{ebfd}=-\frac{5}{72},
$$
$$
|\nabla\mathrm{Riem}|^2=\frac14.
$$
Because \(K_6\) is homogeneous but not locally symmetric, \(\nabla\mathrm{Riem}\neq0\) , so the
derivative sector of the \(a_6\) expansion does not trivially vanish; only \((\nabla\mathrm{Riem})^2\) 
survives (the retracted quartic ansatz \(256a^2(a^2-1)^2\) does not apply here and stays dead). In
trip-unit normalization this derivative sector reads \(|\nabla\mathrm{Riem}|^2=54\) (equal to \(1/4\) 
in Killing-form units), \(|\nabla\mathrm{Ric}|^2=0\) , \(|\nabla\mathrm{Scal}|^2=0\) ,
 \(|\nabla^{LC}E_{\rm grav}|^2=162\) , with the cross-check
 \(R_{abcd}\Box R^{abcd}=-54=-|\nabla\mathrm{Riem}|^2\) closing the loop between the box operator and
the covariant-derivative norm. This sector has been carried end-to-end through the banked \(a_6\) 
chain (not dropped as a symmetric-space artifact — \(K_6\) is not locally symmetric, so this term is
structurally required).

 Weight-6 (dimension-6) invariant basis at the Einstein center (Killing-form, exact rationals): 
$$
\mathrm{Scal}^3=\frac{125}{8},\quad
\mathrm{Scal}\,|\mathrm{Ric}|^2=\frac{125}{48},\quad
\mathrm{Scal}\,|\mathrm{Riem}|^2=\frac{115}{24},\quad
|\mathrm{Ric}|^3=\frac{125}{288},
$$
$$
\mathrm{Ric}^{ab}\mathrm{Ric}^{cd}R_{acbd}=\frac{125}{288},\qquad
\mathrm{Ric}^{ab}R_a{}^{cde}R_{bcde}=\frac{115}{144}.
$$
These nine invariants, together with the \(E_L\) spectrum of Section II.2 and the \(a_0,a_2,a_4\) 
coefficients below, form the certified core that any \(a_6\) route must be built from.

 Sphere cross-checks (machinery validation, exact rationals, agreement to \(\sim4\times10^{-14}\) 
between routes): 
$$
a_6(S^2)=\frac{4}{315},\qquad a_6(S^4)=\frac{74}{63},\qquad a_6(S^6)=\frac{1139}{63},\qquad
a_6^{\rm conf}(S^6)=\frac{5}{63}.
$$
Lower coefficients on the sphere calibration ( \(S^6\) round unit): \(a_0=1\) , \(a_2=5\) , \(a_4=12\) . A
normalization-robust, scale-free ratio computed on \(K_6\) is \(a_4/a_2^2=66/125\) , which — being a
ratio of scalar invariants — is Levi-Civita-normalization-immune. These sphere numbers do not
enter the graviton computation directly; they exist to validate that the \(a_6\) machinery (Gilkey
formula implementation, index conventions, the product rule
 \(a_{2k}(M_1\times M_2)=\sum_{i+j=k}a_{2i}(M_1)a_{2j}(M_2)\) ) is correctly wired, on manifolds where
the answer is independently known in closed form.

 Topology (exact, frozen negative controls): \(\chi(K_6)=6=|S_3|\) (the number of Weyl chambers
of \(A_2\) ), \(\chi(S^2)=2\) , \(\chi(S_Y^1/\mathbb{Z}_2)=1\) .

 II.5 Step 4 — assembling the graviton and ghost \(a_6\) objects: the two-layer record

 With the operator (Step 1), the fiber weight (Step 2), and the curvature inputs (Step 3) in hand,
the construction proceeds to the actual weight-6 heat-kernel coefficient. This is where the
derivation currently runs into an honest, named computational wall, and the record of that wall
carries two layers that must both be shown, because they answer different questions and neither
one is dispensable.

 Layer A — the closure-of-record ledger (2026-07-05, conservative reading). The graded
graviton-plus-ghost value is OPEN / FAIL_VALUE_MISMATCH : two independent computation routes
disagree by \(31/48\approx0.6458\overline{3}\) , roughly six orders of magnitude outside the
pre-registered \(10^{-6}\) tolerance for route agreement.
- Route A (the graviton \(K_6\) -bundle candidate) returns \(-43/504\) versus an anchor-inconsistent
 \(-16/315\) .
- Route B earns only the Bochner ghost value \(149/1008\) , which is computed with \(E=0\) — not
 the physically correct Faddeev–Popov ghost, whose endomorphism is \(E=-\mathrm{Ric}\) (the sign
 flip relative to the vector-bundle \(E=+\mathrm{Ric}\) used in Section II.2 is the ghost's
 statistics-induced sign, standard for anticommuting Faddeev–Popov fields).
- The resulting physical-ghost mismatch is \(\left|-\tfrac{251}{504}-\tfrac{149}{1008}\right|=\tfrac{31}{48}\) .

 The exactly located source of this debt is the graviton \(\mathrm{Sym}^2(T)\) Levi-Civita
off-diagonal leg : the Lichnerowicz first-order (hopping) term on \(\mathrm{Sym}^2_0\) mixes the
five Weyl-inequivalent \(T^2\) weight classes, and the required off-diagonal connection matrix
elements are \(SU(3)\) Gelfand–Tsetlin (GT) ladder matrix elements between adjacent GT patterns.
These are exactly computable in principle via the standard GT lowering-operator formula, but not
yet enumerated in the current record. The known partial values are: graviton LC gap \(=2/21\) ,
vector LC gap \(=1/24\) (both exact, both already banked) — the debt is specifically the GT
off-diagonal stratum, not a hidden or unnamed gap.

 Consequently: \(\mathrm{tr}[a_6(L_{\rm grav})]\) (the full \(D=13\) vector/graviton trace) is OPEN /
computation-debt , and a previously cited sign \(C\sim-6.39\) is NOT reproduced and must not be
asserted as if it were. On the magnitude leg: at odd \(D=13\) there is no finite local \(t^0\) slot 
in the heat-kernel expansion — the relevant term sits at the half-integer zeta-pole \(s=7/2\) , which
is zero in dimensional regularization with no accompanying log or anomaly slot. This means the
question "what is the GeV \(^6\) value of the bulk \(a_6\) coefficient" is ill-posed at odd D=13 ,
not merely unanswered — the well-posed object is the finite dimensionless trace, not a dimensionful
magnitude, and Layer A records this as DISSOLVED-as-ill-posed for the magnitude leg
specifically (a genuine resolution of that sub-question, not an open item).

 Layer B — the audit/completion cascade (2026-06-30, AUDIT-CERTIFIED chain; ceiling
AUDIT-CLOSED, never physics-CLOSED). Working forward from the same curvature inputs, a chain of
exact-rational intermediate values has been assembled and multi-route verified at the level of
internal consistency (each value checked on \(\ge2\) independent computational routes, with the
crux geometry checked on 3, including from-scratch referee rebuilds):

 Object 
 Exact value 
 Composition 

 Graviton \(\mathrm{Sym}^2(T_6)\) LC \(a_6/a_0\) 
 \(-6373/630\) 
 \(=-3481/360\) (algebraic part) \(+\,(-25/56)\) (derivative sector) 

 Physical defect \(a_6\) 
 \(-7226/35\) 
 \(=21\cdot(-6373/630)-12\cdot(-251/504)\) 

 Vector ghost ( \(E=-\mathrm{Ric}\) ) \(a_6/a_0\) 
 \(-251/504\) 
 \(=-713/1260\) (algebraic) \(+\,19/280\) (derivative) 

 \(\mathbb{Z}_2\) smooth equivariant defect \(\tfrac12 c_3^\gamma\) 
 \(-337361/840\) 
 Donnelly equivariant construction (Section II.6) 

 13D bulk graded \(a_6\) (keystone) 
 \(-953329/1260\) 
 evaluated at frozen \(K_2=5\) ; multi-route 

 AUD-0059 (orbifold assembly) 
 \(-491353/630\) 
 \(=\tfrac12\left(-\tfrac{953329}{1260}\right)+\left(-\tfrac{337361}{840}\right)\) , exact rational arithmetic 

 \(K_6\) scalar \(a_6/a_0\) 
 \(992/315\) 
 \(=8017/2520\) (algebraic) \(+\,(-9/280)\) (derivative) 

 Bianchi-exact re-run, dimensionful bulk 
 \(-2.995681680\times10^{94}\ \mathrm{GeV}^6\) 
 consistency-coefficient only, still route-inconsistent 

 Every rational value in this table checks in exact arithmetic (e.g. the AUD-0059 identity
 \(\tfrac12(-953329/1260)+(-337361/840)=-491353/630\) holds exactly, with no rounding). This is
genuine, verified computational progress. But its status is explicitly AUDIT-CLOSED , meaning
the audit process that checks internal consistency of the completion cascade has certified these
values are what the stated formulas produce — it is not a claim that the physics question
(does the graviton-plus-ghost \(a_6\) correctly and uniquely settle any UV question) is closed.
 headline_green=false is carried explicitly: this is a chain of correct arithmetic on a
provisional assembly, not yet a validated physical answer, because Layer A's route-disagreement
(the \(31/48\) mismatch) has not been resolved by Layer B's completion — the two layers are
different questions (internal-consistency-of-the-cascade versus does-the-cascade-match-an-
independent-route), and Layer B does not supersede Layer A's OPEN finding.

 Object-identity firewall (binding, because these four numbers are easy to conflate): 
$$
\text{bulk } a_6\ (-953329/1260,\ \text{AUD-0059 } -491353/630)\ \neq\ \text{order-6 boundary } a_6\ \neq\ \mathbb{Z}_2\text{-defect equivariant}\ (-337361/840)\ \neq\ \text{graded-Casimir supertrace}.
$$
None of these four objects may be substituted for another; each answers a structurally different
question about the operator.

 A related but distinct object — the color factor \(124/315\) . The scalar \(K_6\) ratio
 \(b_3/b_0=124/315\) (the \(t^0\) /Seeley–DeWitt \(a_{d/2}\) -bracket ratio, computed from the actual \(K_6\) 
Peter–Weyl heat trace) was earlier called "DERIVED dual-validated," but is now
 DOWNGRADED to DERIVED-PENDING-INDEPENDENT-TARGET-BLIND-REPRODUCTION , because the same
R2-carrying, metric-selected engine (selected at \(\mathrm{Scal}_{K_6}=7.5\) ) produced it. It must
always be reported with this caveat until an independent, structurally different route reproduces
it — it is not yet the "clean route-independent invariant" it was once called. Only dimensionless
 ratios built from scale-free invariants (e.g. \(a_4/a_2^2=66/125\) ) are immune to this
metric-selection concern; \(124/315\) is not one of those.

 Result of Step 4: the ghost-corrected fiber weight (Step 2, forced, exact) and the curvature
inputs (Step 3, proved target-blind) are both solid. The weight-6 heat-kernel coefficient built
from them is genuinely computed on multiple exact-rational internal-consistency routes (Layer B),
but two physically-motivated independent routes for the same graviton-plus-ghost object disagree
by \(31/48\) (Layer A) — a named, located (GT off-diagonal stratum), unresolved discrepancy. Neither
layer promotes the gate; both must be shown.

 II.6 Step 5 — the \(\mathbb{Z}_2\) orbifold defect: getting the boundary term right

 The \(S_Y^1/\mathbb{Z}_2\) factor requires special care because a naive "boundary heat-kernel"
framing is a wrong-object trap. The reflection \(\theta\mapsto-\theta\) is a global isometric
involution on a closed manifold (the circle), not a genuine manifold-with-boundary problem — so
the published boundary heat-kernel tower (which stops at \(a_5\) in the literature) does not apply
here as a boundary-coefficient source, and treating it as one is now recognized as a
wrong-object artifact in the ledger.

 The correct treatment is equivariant / orbifold , following Donnelly. The reflection \(g\) -trace
over the two isolated fixed points \(\theta=0,\pi\) is
$$
\sum_{\rm fixed\ pts}\frac{1}{|1-dg|} = 2\times\frac{1}{|1-(-1)|} = 2\times\frac12 = 1.
$$
The orbifold (parity-projected) traces on the parent circle are
$$
K^+ = \tfrac12 K_{\rm circle} + \tfrac12\quad(\text{even parity}),\qquad
K^- = \tfrac12 K_{\rm circle} - \tfrac12\quad(\text{odd parity}),
$$
i.e. length \(L=\pi R\) plus a defect \(\pm\tfrac12\) ; the per-fixed-point \(a_0\) defect is \(+1/4\) for
even parity and \(-1/4\) for odd parity. Equivalently, the twisted trace on \(S^1_R/\mathbb{Z}_2\) 
evaluates to exactly \(1\) , independent of \(t\) — there is no boundary tower to sum, only this exact,
 \(t\) -independent equivariant defect.

 This is the origin of the smooth equivariant defect \(\tfrac12 c_3^\gamma = -337361/840\) 
appearing in the Layer B table above: it is a Donnelly-type equivariant heat-kernel term, not 
an order-6 mixed Neumann \(\oplus\) Dirichlet boundary coefficient (which does not exist in the
published literature — the boundary tower genuinely stops at \(a_5\) , and this absence is a named,
blocked item, not a place to interpolate or guess). The AUD-0059 assembly
$$
\tfrac12\left(-\frac{953329}{1260}\right) + \left(-\frac{337361}{840}\right) = -\frac{491353}{630}
$$
combines the bulk keystone with this equivariant defect, exactly, in rational arithmetic. What is
explicitly not emitted anywhere in this construction is a TOTAL (bulk + boundary-tower)
 \(a_6\) in the naive manifold-with-boundary sense — that object would require the missing order-6
boundary coefficient, which is blocked absent new specialist literature or a proof that the
Donnelly equivariant term is the complete substitute for it.

 II.7 Step 6 — the scope firewall: why one coefficient cannot be a UV completion

 Independent of whether the \(31/48\) mismatch of Step 4 is ever resolved, there is a structural
reason the entire \(a_6\) computation — however precisely it is eventually pinned down — cannot by
itself constitute a UV completion. The heat-kernel/Seeley–DeWitt expansion is an asymptotic 
short-proper-time expansion,
$$
K(t)\sim(4\pi t)^{-d/2}\sum_{k\ge0}a_{2k}\,t^k,
$$
and the physical divergences of a graviton loop expansion above the cutoff are controlled by the
 entire unbounded tower \(a_6 < a_8 < a_{10} < \cdots\) of ever-higher local curvature invariants,
each contributing its own independent counterterm structure to the non-renormalizable gravity
Lagrangian. A single finite coefficient — even an exactly and unambiguously computed one — carries
no information about whether this infinite tower resums into a sensible, unitary, UV-complete
theory; it is a single term in a divergent (in the sense of "structurally unbounded," not
"numerically infinite") series. This is a mathematical fact about asymptotic expansions with an
unbounded operator ladder, not a hedge: computing \(a_6\) can never, in principle, answer the
strong-coupling completion question, no matter how exactly \(a_6\) is pinned down.

 This scope firewall is the reason the gate records a certificate , not a residual: "one
heat-kernel coefficient is not a UV completion" is a terminal, structural statement, true
independent of the state of the \(31/48\) mismatch. It also explains why closing H4/H5/H6 (the
bounded computational work packages that would resolve the Layer A/Layer B tension) sharpens the
 linearized , operator-supply certificate (5A/5B) but can never by itself close 5C: the
missing object for 5C is not "a correctly-signed \(a_6\) ," it is a non-perturbative, all-orders
construction (a UV completion in the sense of string theory, a validated asymptotic-safety fixed
point, or equivalent) — named explicitly as UQF-9.

 II.8 Step 7 — the refuted magnitude, and why it is dead, not a hole

 An earlier attempt produced a specific dimensionful magnitude for a bulk \(a_6\) -type object,
 \(-2.818\times10^{94}\ \mathrm{GeV}^6\) . This number has been refuted at decision grade : it was
contaminated by an unrelated computational routine (informally "R2"), it depended on an
arbitrarily anchored renormalization scheme choice rather than a scheme-independent construction,
and — as established structurally in Section II.5 — it is ill-posed at odd \(D=13\) in the first
place, since there is no finite local \(t^0\) slot for a dimensionful magnitude to occupy at that
half-integer zeta-pole. This refutation is itself a piece of completed science: a specific,
falsifiable numerical claim was checked and found wrong, for three independent, stated reasons.
The dossier records this as a reached verdict — the number is dead and must never be revived,
distinguished sharply from an open question awaiting resolution. The later Layer B "Bianchi-exact
re-run" value \(-2.995681680\times10^{94}\ \mathrm{GeV}^6\) is explicitly logged as a
consistency-coefficient only, never gap-closing, and still route-inconsistent — it does not
rehabilitate the refuted number, and both numbers being in the same order of magnitude does not
constitute agreement given the underlying route-inconsistency.

 Other dead values that must never resurface in this construction: the retracted quartic ansatz
 \(256a^2(a^2-1)^2\) for the derivative sector (superseded — only \((\nabla\mathrm{Riem})^2=1/4\) 
survives on \(K_6\) , per Section II.4); the buggy build's \(\kappa=7/12\) (corrected to \(5/12\) ); the
buggy \(|\mathrm{Riem}|^2/R^2=31/147\) (corrected to \(23/75\) ); and the buggy first-Bianchi residual
of exactly \(1/7\) (corrected to machine zero, \(\approx2.5\times10^{-16}\) ).

 II.9 Step 8 — the deep-root anchoring: Shape, Scale, Granularity

 The construction closes by identifying exactly what the gate's ANCHORED +1 grade rests on,
across the three deep-root axes.

 Shape. The carrier/background/boundary \(K_6\times S^2\times S_Y^1/\mathbb{Z}_2\) supplies the
curvature data indexing the operator and every heat-kernel coefficient computed above, at full
precision and across all three layers ( \(\times\) Stage/ \(\oplus\) Rulebook/ \(\otimes\) Actors). As stated
in Section II.1, this background is SELECTED by the admissibility/Weyl-rigidity constraint
set, not proven to be uniquely forced — a qualifier that propagates through every curvature
number quoted in Section II.4.

 Scale. The strong-coupling wall is, precisely, the regime above the cutoff where the
perturbative expansion in \(G_N E^2\) ceases to converge. The UV floor read off the geometry is
 \(M_*\approx7.467\times10^{16}\ \mathrm{GeV}\) (from
 \(M_*^{11}=M_{\rm Pl}^2/\mathrm{Vol}(X_{\rm active})\) with
 \(\mathrm{Vol}(X_{\rm active})=3.704417261398702\times10^{-148}\ \mathrm{GeV}^{-9}\) and
 \(M_{\rm Pl}=1.2209\times10^{19}\ \mathrm{GeV}\) , giving
 \(M_*^{11}=4.023836152402511\times10^{185}\ \mathrm{GeV}^{11}\) ), alongside the compactification
scale \(M_U\sim1.0\times10^{16}\ \mathrm{GeV}\) and \(R_0=1.591549430918954\times10^{-17}\ \mathrm{GeV}^{-1}\) .
This is a derived geometry read-off , not a closure of the strong-coupling question and not a
new independent measured anchor — the openness of the gate lives entirely in this Scale root: the
geometry fixes where the wall sits, not what happens at or above it.

 Granularity — the axiom that does the ANCHORED work. The one named, irreducible axiom that
the ANCHORED +1 grade rests on is AXIOM-COSTFLOOR : an irreducible quantum of cost/action —
explicitly not a smallest length — applied Lorentz-invariantly. Its physical grounding is
established, measured physics: the Margolus–Levitin bound ( \(\tau\ge\pi\hbar/2E\) ), the Landauer
bound ( \(\Delta E\ge k_BT\ln2\) ), and the Bekenstein bound ( \(S\le2\pi k_BRE/\hbar c\) ). The floored
quantity being a Lorentz scalar (theorem T-LI) means there is no preferred frame singled out by
the floor — this is the {Lorentz-scalar proper-time floor} half of the anchor pair.

 Crucially, this axiom dissolves exactly one UV-divergence class: the \(a\to0\) runaway behavior
of the higher tower \(\{a_8,a_{10},\dots\}\) that would otherwise diverge as the proper-time
parameter is taken to zero. It does not dissolve the other ten identified walls, and in
particular it does not supply, compute, or bypass the finite coefficient \(a_6\) itself — \(a_6\) 
remains exactly as owed after the axiom is imposed as before. DISSOLVED is not SOLVED: no
non-Gaussian fixed point is exhibited, no constructive UV completion follows. The axiom is
correctly logged as AXIOM-OPEN / not atomic — a named floor that could in principle be relocated
(a different value, or a different implementation of the same cost-floor idea) but not eliminated
by any known argument.

 Layer-2 screen — no finite-grain shortcut. A separate, banked result establishes
 \(R_0/\ell_{\rm Planck}\sim194\) , and that \(R_0\) is a pure color/gauge object with zero
gravitational input (it derives from the \(SU(3)_c\) compactification radius, not from any
gravitational sector construction). The consequence, proven at this level (T-DEEP), is that
 every finite-grain dissolution of the 5C wall relocates onto UQF-9 , via the sub-target P0:
"does gravity own an intrinsic shortest length, or does it inherit the color radius \(R_0\) ?" Put
plainly: 5C has no finite-grain shortcut — there is no way to dissolve the strong-coupling wall
by simply asserting a minimal length, because the only minimal length currently derived in this
geometry belongs to the color sector, not to gravity, and that fact is itself a derived, banked
result rather than an assumption.

 The endpoint anchor. Collecting Steps 1–8: the published row is ANCHORED on the
conditional axiom pair {Δ₀ > 0 (the cost/action floor), Lorentz-scalar proper-time floor} — read
as TERMINAL + RESIDUALS-SHOWN. This is precisely the fixed grade: REDUCED-TO-AXIOM / ANCHORED
+1 . The "+1" is the single named, irreducible axiom (AXIOM-COSTFLOOR) that the whole
construction is reduced to; everything else in Sections II.1–II.7 (the operator, the ghost-forced
fiber weight, the proved curvature ratio, the exact-rational heat-kernel chain, the equivariant
defect, the scope firewall, the refutation) is either DERIVED-GIVEN-E or an honestly-marked OPEN
computation-debt, none of which changes this anchor. SELECTED is not FORCED (the background);
ANCHORED is not DERIVED (the axiom pair anchors the roll-up, it does not derive the completion);
AXIOM-CLOSED is not atomic (the cost-floor axiom is named and could be relocated, not proven
unique). This closes Construction II.

 Construction III - the central result at full precision

 III.0 What this section proves, precisely

 Gate UQF-5C asks whether the frozen 13-dimensional geometry certifies a genuine, non-perturbative
UV completion of the interacting graviton — a full quantum theory of gravity valid at and above
the cutoff where the perturbative expansion in \(G_N E^2\) stops converging. The fixed grade is
 REDUCED-TO-AXIOM / ANCHORED +1 , and this section exists to isolate the single computation the
gate's grade actually turns on and drive it to full precision, with every intermediate arithmetic
step displayed and cross-checked, exactly as a referee re-deriving it by hand would need to see it.
Two objects carry the entire weight of the gate: (A) the ghost-corrected fiber weight \(65\) , forced
by BRST nilpotency, exact, with zero uncertainty; and (B) the exact-rational \(a_6\) heat-kernel
completion chain, which is internally exact — every arithmetic identity below closes to the last
digit in rational arithmetic — but whose two independent physical routes disagree by a named,
located, nonzero amount. Both are shown in full because the honest content of ANCHORED +1 is
precisely the coexistence of an exact forced result with an exact, located, still-open
discrepancy — not the absence of open items, and not a fabricated closure of them either.

 III.1 The forced result, to the last integer: \(91 - 2\cdot13 = 65\) 

 This is the one number in the entire gate carrying zero residual uncertainty, so it is derived
here twice, from two disjoint starting points, to show the two derivations meet at the same
integer with no rounding, no scheme dependence, and no free parameter on either side.

 Derivation 1 — the BRST/Faddeev–Popov supertrace on the gauge-fixed bulk fiber. The graviton
fluctuation \(h_{MN}=g_{MN}-\bar g_{MN}\) on the frozen \(D=13\) background
 \(\mathcal M_4\times K_6\times S^2\times S^1_Y/\mathbb Z_2\) is a section of \(\mathrm{Sym}^2(T)\) , the
symmetric-tensor bundle of the full 13-dimensional tangent space. Its fiber dimension is the
dimension of the space of symmetric \(13\times13\) matrices:
$$
\dim\mathrm{Sym}^2(\mathbb{R}^{13}) = \binom{13+1}{2} = \frac{13\cdot14}{2} = \frac{182}{2} = 91.
$$
De-Donder (harmonic) gauge-fixing this fluctuation,
 \(\bar\nabla^M h_{MN}-\tfrac12\bar\nabla_N h=0\) , is what turns the linearized Einstein–Hilbert
action into a genuine minimal (Laplace-type) operator — without it \(L_{\rm grav}\) is degenerate
along diffeomorphism directions. This gauge choice forces a compensating Faddeev–Popov ghost
sector: one anticommuting ghost field \(c^M\) and one antighost \(\bar c_M\) , each a section of the
tangent bundle \(T\) itself, so each has fiber dimension
$$
\dim(\text{ghost fiber}) = \dim T = 13.
$$
BRST nilpotency, \(Q_{\rm BRST}^2=0\) , fixes the relative weight of the ghost pair in any physical
(gauge-invariant, BRST-cohomology-respecting) supertrace at exactly \(-2\) : one factor of \(-1\) for
each of the two anticommuting ghost fields entering the gauge-fixed path-integral measure with
fermionic statistics — the same mechanism, transplanted from Yang–Mills to the diffeomorphism
gauge group, that fixes ghost multiplicity \(-2\) in ordinary non-abelian gauge theory. The
ghost-corrected fiber weight is therefore
$$
W_{\rm bulk} = \dim\mathrm{Sym}^2(\mathbb{R}^{13}) - 2\cdot\dim T = 91 - 2\cdot13 = 91-26 = \boxed{65}.
$$
Nothing here is a convention pick: \(91\) is a binomial-coefficient count fixed once \(D=13\) is fixed,
 \(13\) is the tangent-bundle fiber dimension fixed by the same \(D\) , and \(-2\) is fixed by BRST
algebra — there is no scheme, gauge parameter, or normalization choice anywhere in this
derivation that could shift the answer away from \(65\) .

 Derivation 2 — the on-shell little-group representation count. Independently of any
gauge-fixing procedure at all, a massless spin-2 particle propagating in \(D\) spacetime dimensions
has its physical, on-shell polarization content classified by the traceless symmetric-tensor
representation of the light-cone little group \(SO(D-2)\) :
$$
\dim\mathrm{Sym}^2_0\big(SO(D-2)\big) = \binom{D-2+1}{2}-1 = \frac{(D-2)(D-1)}{2}-1.
$$
At \(D=13\) , so \(D-2=11\) :
$$
\dim\mathrm{Sym}^2_0(SO(11)) = \frac{11\cdot12}{2}-1 = \frac{132}{2}-1 = 66-1 = \boxed{65}.
$$
This second computation never once mentions gauge-fixing, ghosts, BRST, or a path integral: it is
pure representation theory, counting how many independent polarization states a massless graviton
can physically carry once diffeomorphism redundancy has already been divided out on-shell.

 The exact match, and why it is load-bearing. 
$$
91 - 2\cdot13 = 65 = \dim\mathrm{Sym}^2_0(SO(11)).
$$
Two computations sharing no common machinery — one an off-shell BRST supertrace over a
gauge-fixed path integral, the other an on-shell little-group classification of physical
polarizations — land on the identical integer, with no adjustable parameter on either side that
could have been tuned to force agreement. This is exactly the cross-check that promotes the fiber
weight from "a plausible count" to DERIVED-GIVEN-E: there was no freedom left to choose it, and an
independent method confirms it.

 The negative control, held apart on purpose. A structurally different graded object,
 \(\gamma=\mathrm{Sym}^2\big(\mathrm{diag}(\mathbf 1_{12},-1)\big)\) , produces
$$
\mathrm{tr}\,\gamma_{\rm grav}=67,\qquad \mathrm{tr}\,\gamma_{\rm ghost}=11,\qquad
67-2\cdot11 = 67-22 = 45.
$$
This " \(67/11/45\) " triple belongs to the \(\mathbb{Z}_2\) -defect grading (Section III.6 below), the
orbifold-reflection trace — not the bulk fiber count. It is verified here (the subtraction checks:
 \(67-22=45\) exactly) for one reason only: to make explicit that it must never be substituted for the
bulk \(91/13/65\) triple. The two objects answer structurally different questions — an untwisted bulk
supertrace versus a \(\mathbb{Z}_2\) -twisted defect trace on two different bundles — and using \(67\) or
 \(45\) where \(91\) or \(65\) belongs (or the reverse) is exactly the class of error a target-blind
re-derivation exists to catch.

 Grade of this leg: DERIVED-GIVEN-E — exact once \(D=13\) , the graviton bundle \(\mathrm{Sym}^2(T)\) ,
and the ghost bundle are fixed by the frozen geometry; those bundle assignments are themselves
supplied by the geometry's field content (given-E), not independently re-derived at this step.

 III.2 The proved curvature ratio, to the last digit: \(|\mathrm{Riem}|^2/\mathrm{Scal}^2=23/75\) 

 The second exact, zero-uncertainty result feeding every downstream heat-kernel coefficient is the
dimensionless curvature ratio at the Killing-form chamber center \(\vec u=(1,1,1)\) on
 \(K_6=SU(3)/T^2\) . With simple roots \(\alpha_1=(1,-1,0)\) , \(\alpha_2=(0,1,-1)\) ,
 \(\alpha_1+\alpha_2=(1,0,-1)\) in the Cartan basis \((h_1,h_2,h_3)\) , \(h_1+h_2+h_3=0\) , the tangent space
decomposes as \(T(K_6)=\mathfrak m_1\oplus\mathfrak m_2\oplus\mathfrak m_3\) , each a real 2-plane
carrying one positive root. At the symmetric chamber center the exact curvature scalars
(Killing-form normalization) are:
$$
\dim K_6=6,\qquad \mathrm{Ric}_i=\frac{5}{12}\ (i=1,2,3),\qquad
\mathrm{Scal}=\sum_i\dim(\mathfrak m_i)\,\mathrm{Ric}_i = 2\cdot3\cdot\frac5{12}=\frac52,
$$
$$
\mathrm{Scal}^2=\frac{25}{4},\qquad |\mathrm{Ric}|^2=\frac{25}{24},\qquad |\mathrm{Riem}|^2=\frac{23}{12}.
$$
Forming the ratio explicitly, with the fraction cancellation shown at every step:
$$
\frac{|\mathrm{Riem}|^2}{\mathrm{Scal}^2} = \frac{23/12}{25/4} = \frac{23}{12}\cdot\frac{4}{25}
= \frac{23\cdot4}{12\cdot25} = \frac{92}{300}.
$$
Reducing \(92/300\) : \(\gcd(92,300)=4\) ( \(92=4\cdot23\) , \(300=4\cdot75\) ), giving
$$
\frac{92}{300} = \frac{23}{75} = 0.30\overline{6} = 0.3066666666666667\ldots,
$$
and since \(75=3\cdot5^2\) shares no factor with the prime \(23\) , this is already in lowest terms.
Companion ratios, checked the same explicit way:
$$
\frac{|\mathrm{Ric}|^2}{\mathrm{Scal}^2} = \frac{25/24}{25/4} = \frac{25}{24}\cdot\frac{4}{25}
= \frac{4}{24} = \frac16 = 0.1\overline{6},\qquad
\frac{\mathrm{Scal}}{\mathrm{Ric}_i} = \frac{5/2}{5/12} = \frac52\cdot\frac{12}{5} = 6 = \dim K_6,
$$
and \(\mathrm{Weyl}^2/\mathrm{Scal}^2=6/25=0.24\) (forced by the standard \(d=6\) curvature
decomposition into Ricci-trace and Weyl parts). These ratios are metric-scale invariant — the
 \(R_6^2\) dependence of the dimensionful \(R_6\) -normalization ( \(\mathrm{Ric}_i=1/(2R_6^2)\) ,
 \(\mathrm{Scal}=3/R_6^2\) ) cancels identically in every ratio above — which is exactly why these four
numbers, not the dimensionful Ricci or scalar curvature themselves, are the load-bearing facts
carried through the rest of this construction.

 Cross-check in the independent trip-unit normalization. The same geometry, computed in the
alternate ("trip-unit"/engine) normalization, gives \(\mathrm{Scal}=15/2\) (or \(15\) in the alternate
convention), \(|\mathrm{Ric}|^2=75/8\) , \(|\mathrm{Riem}|^2=69/4\) , \(\mathrm{Weyl}^2=27/2\) . Checking that
this reproduces the identical dimensionless ratio:
$$
\frac{|\mathrm{Riem}|^2}{\mathrm{Scal}^2}\bigg|_{\rm trip\text{-}unit} = \frac{69/4}{(15/2)^2}
= \frac{69/4}{225/4} = \frac{69}{225}.
$$
Reducing: \(\gcd(69,225)=3\) ( \(69=3\cdot23\) , \(225=3\cdot75\) ), giving \(69/225=23/75\) — the identical
reduced fraction, confirming the normalization bridge holds exactly.

 Three independent target-blind routes, each landing on \(23/75\) . (1) SU(3) structure-constant
+ naturally-reductive curvature route : using the Nomizu/Wang–Ziller formula for the Riemann
tensor of a naturally-reductive homogeneous space built directly from the \(\mathfrak{su}(3)\) 
structure constants restricted to \(\mathfrak m=\mathfrak m_1\oplus\mathfrak m_2\oplus\mathfrak m_3\) 
(general-chamber Ricci eigenvalue
 \(\mathrm{Ric}_1=\tfrac{x_1^2-x_2^2+6x_2x_3-x_3^2}{12x_1x_2x_3}\) and cyclic permutations,
specialized to \(x_1=x_2=x_3=1\) ), giving \(|\mathrm{Riem}|^2=23/12\) from pure Lie-algebraic data, no
metric ever explicitly written down componentwise. (2) Curvature-free spectral heat-trace route :
the same ratio reconstructed purely from the Peter–Weyl spectral data (Casimir eigenvalues
 \(C_2(p,q)\) and multiplicities on \(K_6\) ) via the heat-trace coefficients \(a_0,a_2,a_4\) , without ever
constructing the Riemann tensor at all. (3) Levi-Civita full-Riemann route : the connection and
curvature tensor computed directly as functions of the general 3-parameter chamber metric
 \(g_{K_6}(\vec u)\) , then specialized to \(\vec u=(1,1,1)\) . These three routes share essentially no
intermediate machinery — one purely algebraic/Lie-theoretic, one spectral/analytic, one direct
differential-geometric — so a systematic error in any single method's implementation would not
propagate into agreement across all three; only a genuinely correct computation should converge on
the same rational from all three starting points, and all three do.

 Target-blindness, made quantitative. Two wrong values had circulated before \(23/75\) was pinned
down: an earlier buggy-engine value \(31/147=0.210884\ldots\) , and a separately circulating firewall
estimate of \(0.0667\) . Checking the distance to the first explicitly:
$$
\left|\frac{23}{75}-\frac{31}{147}\right| = \left|\frac{23\cdot147-31\cdot75}{75\cdot147}\right|
= \left|\frac{3381-2325}{11025}\right| = \frac{1056}{11025} \approx 0.0958 \neq 0,
$$
and \(23/75=0.3067\) is more than four and a half times the separately circulating \(0.0667\) estimate.
A computation converged on by three structurally independent methods, landing on neither of the
two numbers a motivated or erroneous calculation could have been steered toward, is the strongest
available evidence against reverse-engineering — this is exactly the discipline that also flags
the later-refuted dimensionful value in Section III.6 as suspect for the opposite reason (it was 
scheme-anchored/target-loaded).

 The correctness theorem, independent of the ratio's value. The corrected curvature tensor used
to compute \(23/75\) satisfies the first Bianchi identity, \(R_{a[bcd]}=0\) , to residual
$$
\approx 2.5\times10^{-16}
$$
— floating-point machine zero. The earlier buggy build's curvature tensor violated first Bianchi by
an exact , finite, nonzero rational: \(1/7\) in the engine's own normalization, equivalently \(1/6\) in
raw Killing-form ( \(-B\) ) units. Because first Bianchi is an identity any genuine Riemann tensor must
satisfy exactly — there is no free parameter that can be tuned to make it pass "approximately" — a
clean rational miss of \(1/7\) or \(1/6\) is a decisive, unambiguous bug signature, not noise. This
single target-blind, parameter-free test is what caught and forced the correction
 \(31/147\to23/75\) , and correspondingly the Einstein-constant correction
$$
\kappa = \mathrm{Ric}_i = \frac{5}{12} \qquad (\text{buggy value }\kappa=7/12,\ \text{retired}).
$$
Passing first Bianchi does not, by itself, validate \(23/75\) as physically meaningful beyond
confirming the object is a genuine Riemann tensor. Combined with the three-route agreement and the
target-blindness computation above, the cumulative weight of evidence for \(23/75\) is: DERIVED, at
full precision, target-blind, cross-checked three independent ways, certified by a parameter-free
theorem-test. It remains necessary, not sufficient for anything beyond this point — it is
curvature input to every \(a_{2k}\) coefficient, never itself a statement about strong-coupling
consistency.

 III.3 The derivative-curvature sector: \(K_6\) is homogeneous but not locally symmetric

 Because \(K_6=SU(3)/T^2\) is homogeneous but not locally symmetric, \(\nabla\mathrm{Riem}\neq0\) ,
so a genuinely new invariant enters the \(a_6\) Gilkey basis beyond the purely algebraic curvature
products of Section III.2 — only \((\nabla\mathrm{Riem})^2\) survives on \(K_6\) (the earlier candidate
quartic \(256\,a^2(a^2-1)^2\) is retracted and stays dead). The exact values, target-blind and
cross-checked on multiple routes:
$$
|\nabla\mathrm{Riem}|^2 = 54\ \text{(trip-unit)} = \frac14\ \text{(Killing-form)},\qquad
|\nabla\mathrm{Ric}|^2=0,\qquad |\nabla\mathrm{Scal}|^2=0,\qquad
|\nabla^{LC}E_{\rm grav}|^2=162\ \text{(trip-unit)},
$$
with the box cross-check
$$
R_{abcd}\Box R^{abcd} = -54 = -|\nabla\mathrm{Riem}|^2
$$
holding exactly. A vanishing \(|\nabla\mathrm{Riem}|^2\) would be the locally-symmetric-space
signature; its nonvanishing here ( \(1/4\) in Killing-form units, verified via the Nomizu formula and
checked against second Bianchi with zero violations) is itself the certificate that \(K_6\) is
homogeneous but not symmetric, and is why this derivative sector cannot be dropped as a
symmetric-space artifact. The companion cubic invariants at the Killing-form center are
$$
K_1\equiv R_{ab}{}^{cd}R_{cd}{}^{ef}R_{ef}{}^{ab}=8\,\mathrm{tr}(R_{\rm op}^3)=-\frac{113}{72},\qquad
K_2\equiv R_{abcd}R_{aecf}R_{ebfd}=-\frac{5}{72},
$$
and the weight-6 invariant basis at the same center is
$$
\mathrm{Scal}^3=\frac{125}{8},\quad \mathrm{Scal}\,|\mathrm{Ric}|^2=\frac{125}{48},\quad
\mathrm{Scal}\,|\mathrm{Riem}|^2=\frac{115}{24},\quad
\mathrm{Ric}^{ab}\mathrm{Ric} b{}^c\mathrm{Ric}_c{}^a=\frac{125}{288},
$$
$$
\mathrm{Ric}^{ab}\mathrm{Ric}^{cd}R {acbd}=\frac{125}{288},\qquad
\mathrm{Ric}^{ab}R_a{}^{cde}R_{bcde}=\frac{115}{144}.
$$
Together with the \(E_L\) spectrum of Construction II and the \(a_0,a_2,a_4\) coefficients below, these
nine invariants form the certified core any \(a_6\) route must be built from — this sector is carried
end-to-end through the banked \(a_6\) chain (audited CHAIN-SAFE), never dropped as a
symmetric-sphere-blind omission.

 III.4 Machinery validation: exact sphere cross-checks

 Before trusting any \(K_6\) -specific \(a_6\) number, the heat-kernel machinery itself is validated on
manifolds where the answer is independently known in closed form, using the convention
 \(K(t)\sim(4\pi t)^{-d/2}\sum_k a_{2k}t^k\) and the exact convolution product rule
 \(a_{2k}(M_1\times M_2)=\sum_{i+j=k}a_{2i}(M_1)a_{2j}(M_2)\) :
$$
a_6(S^2)=\frac{4}{315},\qquad a_6(S^4)=\frac{74}{63},\qquad a_6(S^6)=\frac{1139}{63},\qquad
a_6^{\rm conf}(S^6)=\frac{5}{63}.
$$
On the \(S^2\) scalar sector at radius \(r=1\) : \(a_2/a_0=1/3\) , \(a_4/a_0=1/15\) , \(a_6/a_0=4/315\) . On the
round unit \(S^6\) (the calibration sphere for the \(a_4\) formula): \(a_0=1\) , \(a_2/a_0=5\) ,
 \(a_4/a_0=12\) exactly — this \(a_4=12\) value is the certification point that confirms \(K_6\) is not 
 \(S^6\) (a passed negative control, since \(K_6\) 's own invariants do not reduce to the round-sphere
values). Two independent computation routes reproduce these sphere values to agreement
 \(\sim4\times10^{-14}\) . A normalization-robust, Levi-Civita-immune, scale-free ratio computed
directly on \(K_6\) is
$$
\frac{a_4}{a_2^2} = \frac{66}{125}
$$
(scalars are LC-immune, so this ratio is protected against any Levi-Civita-connection ambiguity in
the underlying engine implementation). These sphere numbers validate the machinery exactly where
it can be independently checked; they do not validate the \(K_6\) -specific \(a_6\) value, which
additionally requires the Gelfand–Tsetlin off-diagonal connection data of Section III.6 that has no
analogue on any sphere. Topology, held fixed throughout as exact negative controls: \(\chi(K_6)=6=
|S_3|\) (the number of Weyl chambers of \(A_2\) ), \(\chi(S^2)=2\) , \(\chi(S^1_Y/\mathbb Z_2)=1\) .

 III.5 The exact-rational heat-kernel completion chain, arithmetic shown at every link

 With the forced fiber weight ( \(65\) , Section III.1), the proved curvature ratio ( \(23/75\) , Section
III.2), and the derivative sector (Section III.3) as certified inputs, the audit/completion
cascade assembles a chain of exact-rational heat-kernel values. Every arithmetic identity below is
reproduced with the sum carried out explicitly over a common denominator, checkable without
external tools.

 Link 1 — the graviton \(\mathrm{Sym}^2(T_6)\) Levi-Civita coefficient , split into an algebraic
piece (built from the certified \(E_L\) spectrum and the weight-6 invariants above) and a
derivative-sector piece (built from \(|\nabla\mathrm{Riem}|^2=1/4\) ):
$$
\frac{a_6}{a_0}\bigg| {\rm graviton\ LC} = -\frac{3481}{360} + \left(-\frac{25}{56}\right).
$$
Common denominator \(\mathrm{lcm}(360,56)=2520\) :
$$
-\frac{3481}{360}=-\frac{3481\cdot7}{2520}=-\frac{24367}{2520},\qquad
-\frac{25}{56}=-\frac{25\cdot45}{2520}=-\frac{1125}{2520},
$$
$$
-\frac{24367}{2520}-\frac{1125}{2520} = -\frac{25492}{2520}.
$$
Reducing: \(\gcd(25492,2520)=4\) , so \(-25492/2520=-6373/630\) . This matches the recorded value
exactly:
$$
\frac{a_6}{a_0}\bigg| {\rm graviton\ LC} = -\frac{6373}{630} = -10.11587301587\ldots
$$

 Link 2 — the vector ghost coefficient ( \(E=-\mathrm{Ric}\) , the physically correct
Faddeev–Popov sign — the statistics-induced sign opposite to the ordinary vector bundle's
 \(E=+\mathrm{Ric}\) ):
$$
\frac{a_6}{a_0}\bigg| {\rm vector\ ghost} = -\frac{713}{1260} + \frac{19}{280}.
$$
Common denominator \(\mathrm{lcm}(1260,280)=2520\) :
$$
-\frac{713}{1260}=-\frac{1426}{2520},\qquad \frac{19}{280}=\frac{171}{2520},\qquad
-\frac{1426}{2520}+\frac{171}{2520}=-\frac{1255}{2520}.
$$
Reducing: \(\gcd(1255,2520)=5\) , so \(-1255/2520=-251/504\) . This matches the recorded value exactly:
$$
\frac{a_6}{a_0}\bigg| {\rm vector\ ghost} = -\frac{251}{504} = -0.4980158730\ldots
$$

 Link 3 — the physical (graviton-plus-ghost) defect , combining Links 1–2 with integer
multiplicities \(21\) and \(-12\) from the bundle-rank bookkeeping of the full supertrace over the
compact factor:
$$
a_6^{\rm phys.\ defect} = 21\cdot\left(-\frac{6373}{630}\right) - 12\cdot\left(-\frac{251}{504}\right).
$$
First term: \(\gcd(21,630)=21\) , \(630/21=30\) , so
$$
21\cdot\left(-\frac{6373}{630}\right) = -\frac{6373}{30}.
$$
Second term: \(\gcd(12,504)=12\) , \(504/12=42\) , so
$$
-12\cdot\left(-\frac{251}{504}\right) = \frac{251}{42}.
$$
Summing over common denominator \(\mathrm{lcm}(30,42)=210\) :
$$
-\frac{6373}{30}=-\frac{6373\cdot7}{210}=-\frac{44611}{210},\qquad
\frac{251}{42}=\frac{251\cdot5}{210}=\frac{1255}{210},
$$
$$
-\frac{44611}{210}+\frac{1255}{210} = -\frac{43356}{210}.
$$
Reducing: \(\gcd(43356,210)=6\) , so \(-43356/210=-7226/35\) . This matches the recorded value exactly:
$$
a_6^{\rm phys.\ defect} = -\frac{7226}{35} = -206.4571428571\ldots
$$

 Link 4 — the \(K_6\) scalar coefficient (an independent control on the same
algebraic-plus-derivative assembly pattern used in Links 1–2):
$$
\frac{a_6}{a_0}\bigg| {K_6\ {\rm scalar}} = \frac{8017}{2520} + \left(-\frac{9}{280}\right).
$$
Since \(\mathrm{lcm}(2520,280)=2520\) : \(-9/280=-81/2520\) , so
$$
\frac{8017}{2520}-\frac{81}{2520} = \frac{7936}{2520}.
$$
Reducing: \(\gcd(7936,2520)=8\) , so \(7936/2520=992/315\) . This matches the recorded value exactly:
$$
\frac{a_6}{a_0}\bigg| {K_6\ {\rm scalar}} = \frac{992}{315} = 3.149206349\ldots
$$
This value shares the intermediate numerator \(7936\) with the separately-cited scalar backbone
 \(a_6/a_2^3=7936/39375\) before this link's final reduction step — a useful internal-consistency
thread, though not itself an independent route confirmation of either value.

 Every link above is exact rational arithmetic, verified to the last digit with no rounding at
any stage. This is genuine computational content, not an assertion: the algebraic and
derivative-sector pieces feeding each link trace back to the certified \(E_L\) spectrum and the
certified curvature invariants of Sections III.2–III.3, and each multi-term sum above closes
exactly. This is Layer B of the record — an AUDIT-CERTIFIED chain, each value checked on at least
two independent computational routes (the crux input, \(23/75\) , on three, including from-scratch
referee rebuilds).

 III.6 The \(\mathbb{Z}_2\) orbifold defect, computed exactly, and the AUD-0059 assembly

 The \(S^1_Y/\mathbb Z_2\) factor contributes a Donnelly-type equivariant defect, not an ordinary
manifold-with-boundary heat-kernel term: the reflection \(\theta\mapsto-\theta\) is a global
isometric involution on a closed circle, not a genuine boundary-value problem, and treating it
as one is a wrong-object trap (the published boundary heat-kernel tower stops at \(a_5\) and simply
does not supply an order-6 term of this kind). The reflection \(g\) -trace over the two isolated fixed
points \(\theta=0,\pi\) is computed exactly via the Lefschetz-type fixed-point formula, using
 \(dg=-1\) at each fixed point of the reflection:
$$
\sum_{\rm fixed\ pts}\frac{1}{|1-dg|} = 2\times\frac{1}{|1-(-1)|} = 2\times\frac12 = 1.
$$
The resulting per-fixed-point \(a_0\) defects are \(+1/4\) (even/ \(+\) parity) and \(-1/4\) (odd/ \(-\) 
parity), giving orbifold traces \(K^{\pm}=\tfrac12K_{\rm circle}\pm\tfrac12\) : an exact,
 \(t\) -independent defect with no boundary tower to sum. The associated smooth equivariant weight-6
defect is
$$
\tfrac12\,c_3^\gamma = -\frac{337361}{840}.
$$

 The AUD-0059 assembly, verified exactly. Combining the 13D bulk graded keystone value
(evaluated at the frozen weight-6 curvature product \(K_2=5\) ),
$$
a_6^{\rm bulk,\ keystone} = -\frac{953329}{1260},
$$
with the equivariant defect above:
$$
\mathrm{AUD}\text{-}0059 = \frac12\left(-\frac{953329}{1260}\right) + \left(-\frac{337361}{840}\right).
$$
First term: \(\tfrac12\cdot\big(-953329/1260\big) = -953329/2520\) . Second term, brought to the same
denominator ( \(\mathrm{lcm}(2520,840)=2520\) , and \(2520/840=3\) ):
$$
-\frac{337361}{840} = -\frac{337361\cdot3}{2520} = -\frac{1012083}{2520}.
$$
Summing:
$$
-\frac{953329}{2520} - \frac{1012083}{2520} = -\frac{1965412}{2520}.
$$
Reducing: \(\gcd(1965412,2520)=4\) ( \(1965412/4=491353\) , \(2520/4=630\) ), giving
$$
\mathrm{AUD}\text{-}0059 = -\frac{491353}{630} = -779.9253968254\ldots
$$
Every digit of this identity closes in exact rational arithmetic with zero residual — this is the
cleanest, fully-closed multi-term assembly in the entire chain: exact, not approximately exact.

 The object-identity firewall, stated with the numbers in hand. Four numerically similar but
structurally distinct objects appear across this construction and must never be substituted for
one another:
$$
\underbrace{-\frac{953329}{1260}} {\text{bulk }a_6\text{ keystone}}\ \neq\
\underbrace{-\frac{491353}{630}} {\text{AUD-0059 orbifold assembly}}\ \neq\
\underbrace{-\frac{337361}{840}} {\mathbb{Z}_2\text{-defect equivariant}}\ \neq\
\underbrace{\text{order-6 boundary coefficient}} {\text{no closed-form target exists — BLOCKED}}.
$$
The fourth object — a genuine mixed Neumann \(\oplus\) Dirichlet order-6 boundary coefficient in the
conventional sense — is not merely uncomputed: the published boundary-heat-kernel literature's
coefficient tower stops at \(a_5\) , so there is no known closed-form target to compute against at
all. This is recorded as BLOCKED, not OPEN — the missing ingredient is new specialist mathematics
(or a proof that the Donnelly equivariant term above is the complete substitute for it), not a
calculation this construction failed to carry out.

 The negative control from Section III.1, placed correctly. The \(67/11/45\) triple belongs here,
in the \(\mathbb{Z}_2\) -defect grading family, not in the bulk fiber count of Section III.1 — this is
the one place in the record where that triple's home construction actually lives, which is
precisely why it must never be substituted for \(91/13/65\) upstream.

 III.7 The located discrepancy: \(31/48\) , computed exactly, and where it lives

 The chain in Sections III.5–III.6 is exact internally, but it does not yet answer the physical
question the gate poses, because two independent routes to the graviton-plus-ghost object
disagree. This discrepancy is computed here exactly, to isolate precisely what it is — and, just as
important, what it is not.

 A first, separate mismatch (Route A vs. its anchor). Route A (the graviton \(K_6\) -bundle
candidate, built from the Lichnerowicz/Gilkey formula directly on \(\mathrm{Sym}^2_0\) ) returns
 \(a_6^{\rm Route\ A}=-43/504\) , checked against an independently-expected anchor value of \(-16/315\) .
These do not agree:
$$
-\frac{43}{504}-\left(-\frac{16}{315}\right) = -\frac{43}{504}+\frac{16}{315}.
$$
Common denominator \(\mathrm{lcm}(504,315)=2520\) : \(-43/504=-215/2520\) , \(16/315=128/2520\) , so
$$
-\frac{215}{2520}+\frac{128}{2520} = -\frac{87}{2520} = -\frac{29}{840}
$$
after reducing by \(\gcd(87,2520)=3\) . This is a real, exact, nonzero mismatch in its own right
( \(-29/840\approx-0.0345\) ), flagged in the record as "anchor-inconsistent" — and it is a different 
numerical discrepancy from the headline mismatch below. The two must not be conflated: both are
open, but they are not the same gap.

 The headline mismatch. This is between the physical Faddeev–Popov ghost value of Link 2
(Section III.5), \(-251/504\) (built with the physically correct \(E=-\mathrm{Ric}\) ), and the
 Bochner ghost value obtainable from Route B's reconstruction, \(149/1008\) — which uses \(E=0\) ,
the wrong endomorphism for a physical Faddeev–Popov ghost (the Bochner Laplacian omits the
curvature endomorphism entirely, whereas the physical ghost's kinetic operator is
 \(-(\nabla^2+\mathrm{Ric})\) acting with the FP sign). Computing the exact difference, with common
denominator \(\mathrm{lcm}(504,1008)=1008\) and \(-251/504=-502/1008\) :
$$
\left|-\frac{502}{1008}-\frac{149}{1008}\right| = \left|-\frac{651}{1008}\right| = \frac{651}{1008}.
$$
Reducing: \(\gcd(651,1008)=21\) ( \(651/21=31\) , \(1008/21=48\) ), giving exactly
$$
\boxed{\frac{31}{48} = 0.6458\overline{3}.}
$$
This is six orders of magnitude outside the pre-registered \(10^{-6}\) route-agreement tolerance —
not a rounding disagreement but an exact rational mismatch of nearly two-thirds of a unit in the
 \(a_6/a_0\) normalization.

 Where the mismatch is located, exactly. The located source is the graviton \(\mathrm{Sym}^2(T)\) 
Levi-Civita off-diagonal (hopping) term: the Lichnerowicz first-order connection term on
 \(\mathrm{Sym}^2_0\) mixes the five Weyl-inequivalent \(T^2\) weight classes, and the matrix elements
governing that mixing are \(SU(3)\) Gelfand–Tsetlin (GT) ladder matrix elements between adjacent GT
patterns — a standard, exactly computable lowering-operator formula (each element a square root of
a product of pattern-entry differences) that has simply not yet been enumerated in the current
record. The partial data already banked at this stratum is exact: graviton LC gap \(=2/21\) , vector
LC gap \(=1/24\) . The \(31/48\) mismatch is therefore a named, located, bounded computation debt,
not an unnamed or open-ended one. Consequently \(\mathrm{tr}[a_6(L_{\rm grav})]\) , the full \(D=13\) 
graviton-minus-ghost trace, is OPEN , and a previously circulated coefficient value
 \(C\sim-6.39\) is not reproduced by anything in this construction and must not be asserted as if
it were.

 III.8 The magnitude question, dissolved as ill-posed (a genuine resolution, not a gap)

 A distinct sub-question — "what is the dimensionful GeV \(^6\) value of the bulk \(a_6\) coefficient" —
is resolved, not merely left open, by a structural fact about odd-dimensional heat-kernel
expansions. At \(D=13\) (odd), the relevant term in the zeta-function regularization of the heat
kernel sits at the half-integer pole
$$
s = \frac{D}{2}-3 = \frac{13}{2}-3 = \frac72,
$$
and the residue of the associated \(\Gamma(s)\zeta(s,\ldots)\) term at a half-integer \(s=7/2\) 
vanishes identically in dimensional regularization, with no accompanying logarithm or anomaly term
to carry a finite answer — unlike the even-dimensional case, where \(a_d\) multiplies a genuine
logarithmic divergence with a well-defined finite piece. There is, in the technical sense, no
finite local \(t^0\) slot at odd \(D=13\) for a dimensionful bulk magnitude to occupy. This means the
question is ill-posed , not merely unanswered: no amount of further computation produces a
number here, because the object being asked for does not exist at this order in odd dimension. The
well-posed replacement object is the finite, dimensionless trace \(\mathrm{tr}[a_6(L_{\rm grav})]\) 
of Section III.7, which remains OPEN for the separate, located reason given there (the GT
stratum) — not because it shares the odd- \(D\) ill-posedness of the magnitude question; these are two
distinct dispositions on two distinct objects.

 This dissolution is why the earlier dimensionful claim \(-2.818\times10^{94}\ \mathrm{GeV}^6\) is
recorded as REFUTED , not merely superseded: it purported to answer a question with no
well-posed dimensionful answer at odd \(D=13\) , independent of its other two defects (contamination
by an unrelated computational routine informally called "R2," and an arbitrarily anchored
renormalization-scheme choice rather than a scheme-independent construction). A later
"Bianchi-exact re-run" produced \(-2.995681680\times10^{94}\ \mathrm{GeV}^6\) from the same
ill-posed construction; it is logged explicitly as a consistency-coefficient only , never
gap-closing and still route-inconsistent. The two numbers' proximity in order of magnitude
( \(-2.818\) vs. \(-2.996\) , both \(\times10^{94}\ \mathrm{GeV}^6\) ) does not constitute agreement, since
both answer a question that has no well-posed dimensionful answer at this order — closeness between
two ill-posed evaluations of the same non-existent quantity carries no evidential weight.

 III.9 What this construction has and has not shown

 Collecting Sections III.1–III.8: the fiber weight \(65=91-2\cdot13=\dim\mathrm{Sym}^2_0(SO(11))\) is
exact, forced, and cross-checked with zero uncertainty (III.1). The curvature ratio \(23/75\) is
exact, proved by three independent target-blind routes, and certified correct by a parameter-free
first-Bianchi theorem-test (III.2), with its derivative-sector companion \(|\nabla\mathrm{Riem}|^2=
1/4\) exact and nonzero as required by \(K_6\) 's non-symmetric homogeneity (III.3). The heat-kernel
machinery is validated to \(\sim10^{-14}\) against exact sphere closed forms (III.4). The completion
chain built from these certified inputs closes exactly in rational arithmetic at every link,
including the fully verified four-term AUD-0059 assembly (III.5–III.6). The discrepancy between two
independent physical routes to the graviton-plus-ghost object is itself computed exactly, to
 \(31/48\) , and located exactly, at the GT off-diagonal stratum (III.7), while the separate magnitude
question is resolved as structurally ill-posed at odd \(D=13\) rather than left dangling (III.8).

 Every number in this construction is either an exact rational, verified here by explicit
common-denominator arithmetic that a reader can check line by line, or a decision-grade
refutation with a stated, structural reason. Nothing here computes a non-perturbative UV
completion, exhibits a fixed point, or bounds the strong-coupling behavior of the interacting
graviton at or above the cutoff. The central result at full precision is: a forced fiber weight, a
proved curvature ratio, and an exact-but-internally-disagreeing heat-kernel chain — genuine,
load-bearing content that the gate's ANCHORED +1 grade rests on, together with the named axiom pair
{Δ₀ > 0, Lorentz-scalar proper-time floor} that supplies the "+1" itself. The strong-coupling wall
is untouched by any number computed in this construction, exactly as the scope-firewall certificate
requires: one heat-kernel coefficient — however exactly pinned down, however many digits are
carried — is never a UV completion.

 The insights that made it work

 This section is not a restatement of the numbers already exhibited elsewhere in this dossier; it is
the argument for why each move is the right move — the piece of reasoning that makes the banked
content believable to a referee who already knows the standard heat-kernel and BRST literature, and
reproducible by a physicist who has never seen this specific geometry before. Six insights carry the
whole construction. Each is stated as a general principle first, then shown operating on this
specific 13-dimensional arena with the exact numbers it produces. None of the six closes UQF-5C; each
is honest about exactly how far it reaches, and the section ends by showing how the six compose into
the single fixed grade, REDUCED-TO-AXIOM / ANCHORED +1, without any of them individually or jointly
overstepping into a UV-completion claim.

 Insight 1 — BRST nilpotency, not convention, forces the ghost multiplicity and sign

 The general principle. Whenever a gauge symmetry is fixed to quantize a field with redundant
components, the resulting one-loop trace over physical and unphysical modes is not free to be
normalized however is convenient. The Faddeev–Popov determinant that compensates for the gauge-fixing
delta function appears in the path integral as \(\det(\text{ghost operator})^{\pm1}\) , and turning that
determinant into an exponentiated ghost action forces anticommuting (Grassmann) ghost fields whose
loop contributes with a relative sign fixed by Grassmann statistics and a multiplicity fixed by
matching the ghost operator's rank to the gauge parameter's. Crucially, the nilpotency of the BRST
charge, \(Q_{\rm BRST}^2=0\) , is what guarantees this bookkeeping is consistent order by order — it is
the algebraic statement that the gauge-fixed theory's physical Hilbert space (the BRST cohomology) is
independent of the gauge-fixing choice. This is why the ghost weight is not a free normalization a
theorist could dial to taste: get the sign or the multiplicative factor wrong, and \(Q_{\rm BRST}^2=0\) 
fails, the gauge-fixed and physical theories decouple, and the one-loop trace stops being
gauge-parameter-independent.

 Why this matters here, concretely. On the frozen 13-dimensional arena, the graviton is the metric
fluctuation \(h_{MN}\) around the fixed background, gauge-fixed in de-Donder (harmonic) gauge
 \(\nabla^Mh_{MN}-\tfrac12\nabla_Nh=0\) . Diffeomorphism invariance is exactly the gauge redundancy this
fixing removes, and it forces a compensating Faddeev–Popov ghost sector — a one-form (vector) ghost
field on the same 13-dimensional base. The graviton's off-shell fiber is \(\mathrm{Sym}^2\) of the
13-dimensional tangent space, dimension
$$
\dim\mathrm{Sym}^2(\mathbb R^{13})=\binom{14}{2}=\frac{13\cdot14}{2}=91,
$$
and the compensating ghost fiber is the 13-dimensional vector representation itself, dimension \(13\) .
Because the ghost is a complex Grassmann pair contributing with weight \(-2\) relative to a physical
bosonic degree of freedom — the same \(-2\) that makes the Faddeev–Popov determinant appear squared,
 \(\det(\text{ghost})^{-2\times(1/2)}\) , in a path integral with one real gauge parameter per point — the
ghost-corrected fiber weight is forced to be
$$
\boxed{\ 91-2\cdot13=91-26=65\ }.
$$
Nothing here was tuned: once \(D=13\) , the graviton bundle \(\mathrm{Sym}^2(T)\) , and the de-Donder
gauge-fixing are fixed by the geometry, both the sign ( \(-\) ) and the multiplicity ( \(2\) ) of the ghost
term are consequences of \(Q_{\rm BRST}^2=0\) , not a choice made to hit a target.

 The insight that makes it believable, not just assertable. A forced number that could plausibly
have come from more than one bookkeeping convention is weak evidence; a forced number that is
cross-checked by a structurally unrelated counting method is strong evidence. Here the independent
route is representation theory of the massless little group: the physical, on-shell polarizations of
a massless spin-2 field in \(D\) spacetime dimensions are counted by the traceless symmetric tensor of
the little group \(SO(D-2)\) , giving in \(D=13\) 
$$
\dim\mathrm{Sym}^2_0\big(SO(11)\big)=\frac{11\cdot12}{2}-1=66-1=65.
$$
This second computation never mentions ghosts, gauge-fixing, or BRST at all — it is a purely
kinematic statement about which polarizations survive on-shell for a massless particle. That the
off-shell ghost-subtraction count and the on-shell little-group count land on the same integer is
not a coincidence a target-blind referee should be suspicious of; it is the standard consistency
check that any correctly gauge-fixed massless gauge theory must pass (the same check, for instance,
that confirms a \(D=4\) photon's off-shell counting reduces to \(2\) on-shell polarizations under the
 \(D=4\) little group \(SO(2)\) ). Two structurally unrelated derivations of \(65\) is why this number is
graded DERIVED-GIVEN-E rather than merely "computed once."

 The negative control that proves the discipline is real. A different graded object,
 \(\gamma=\mathrm{Sym}^2\big(\mathrm{diag}(\mathbf1_{12},-1)\big)\) , gives \(\mathrm{tr}\,\gamma_{\rm
grav}=67\) and \(\mathrm{tr}\,\gamma_{\rm ghost}=11\) , with graded weight \(67-2\cdot11=45\) . This triple
belongs to an entirely different construction — the \(\mathbb{Z}_2\) -defect grading used for the
orbifold sector, not the bulk fiber count — and the discipline of keeping \(65\) (bulk) and \(45\) 
(defect) permanently distinct, never substituting one for the other, is exactly the kind of
bookkeeping habit that prevents a fabricated-number failure mode. The insight generalizes: whenever
two objects share superficially similar Grassmann-graded structure, the antidote is to ask which
physical bundle and which parity/grading operator each trace is actually built from, not to match
traces by their numerical proximity.

 Insight 2 — target-blind triangulation is what turns a computed number into a proved one

 The general principle. Any single computational route to a geometric invariant is only as
trustworthy as the code, the convention, and the intermediate steps that produced it — and in a
research program with a documented history of sign errors and mis-normalized curvature conventions,
a single route is not evidence of correctness, only evidence of internal consistency with itself.
The insight that converts "computed" into "proved" is triangulation across structurally
independent methods that share no code path, no intermediate object, and — most importantly — no
prior knowledge of what answer the other methods returned. If three genuinely different derivations,
run without communicating their partial results to each other, land on the identical rational
number, the chance that all three share the same undetected error collapses; the number has been
proved, not merely computed.

 Why this matters here, concretely. The single most load-bearing curvature ratio feeding every
higher heat-kernel coefficient on \(K_6=SU(3)/T^2\) is the scale-invariant ratio
 \(|\mathrm{Riem}|^2/\mathrm{Scal}^2\) . It is established by three routes that use no common
machinery. \(K_6\) is a normal homogeneous space with tangent decomposition
 \(T(K_6)=\mathfrak m_1\oplus\mathfrak m_2\oplus\mathfrak m_3\) , one real 2-plane per positive root of
 \(A_2\) ( \(\alpha_1=(1,-1,0)\) , \(\alpha_2=(0,1,-1)\) , \(\alpha_1+\alpha_2=(1,0,-1)\) ), and because the
isotropy representation is multiplicity-free (three inequivalent 2-dimensional root spaces, each
appearing exactly once), the curvature tensor at the symmetric chamber center \(\vec u=(1,1,1)\) is
 algebraically forced once the metric there is fixed — there is no free parameter left to tune,
which is exactly what makes a three-route agreement meaningful rather than accidental.

 Structure-constant / naturally-reductive route. The Nomizu/Wang–Ziller curvature formula for
 a naturally-reductive homogeneous space is applied algebraically to the Killing-form metric at
 the chamber center, giving \(\mathrm{Ric}_i=5/12\) , \(\mathrm{Scal}=2\cdot3\cdot(5/12)=5/2\) , and,
 contracting the full Riemann tensor, \(|\mathrm{Riem}|^2=23/12\) .

 Curvature-free spectral heat-trace route. The same ratio is reconstructed purely from
 Peter–Weyl spectral data — Casimir eigenvalues \(C_2(p,q)=(p^2+q^2+pq+3p+3q)/3\) and
 representation multiplicities on \(K_6\) — through the heat-trace coefficients \(a_0,a_2,a_4\) ,
 without ever writing down a single Riemann tensor component. This route could in principle fail
 to reproduce the tensor-algebra answer if there were a sign or normalization inconsistency
 between the "geometric" and "spectral" descriptions of the same space; it does not fail.

 Full Levi-Civita route over the 3-parameter squashed metric. Working with the general
 \(g_{K_6}(\vec u)=\sum_iu_i\langle\cdot,\cdot\rangle_{\mathfrak m_i}\) , \(\vec u\in[1/2,3/2]^3\) ,
 before specializing, computing the Levi-Civita connection and full Riemann tensor as explicit
 functions of \((u_1,u_2,u_3)\) , then evaluating at \(\vec u=(1,1,1)\) — a route that exercises the
 general squashed-metric machinery (also used to classify the invariant Einstein locus: exactly
 four solutions, the symmetric point \((1,1,1)\) plus the three permutations of the Kähler–Einstein
 point \((1,1,2)\) , a classical, independently-known classification result reproduced here as a
 zero-free-parameter check on the metric convention) rather than the specialized Killing-form
 shortcut of route 1.

 All three return
$$
\frac{|\mathrm{Riem}|^2}{\mathrm{Scal}^2}=\frac{23/12}{25/4}=\frac{23}{75}=0.30\overline6.
$$

 The insight that makes the triangulation diagnostic, not decorative. A proof of independence is
only convincing if there is a concrete way it could have failed and did not. Here that check is
explicit: two wrong candidate values had already circulated in the corpus's own history before
this computation was finalized — a buggy-engine output of \(\approx0.2109\) ( \(=31/147\) , traceable to
the retracted curvature-tensor bug) and an unrelated firewall placeholder of \(\approx0.0667\) 
( \(=1/15\) ). The value actually returned, \(23/75\approx0.3067\) , is a third, distinct rational,
matching neither prior candidate. This is the operational meaning of "target-blind": had any of
the three routes been steered, consciously or not, toward a value someone expected to see, the most
likely landing points were already known and were not what came out. Landing somewhere else — and
having all three independent routes land at that same somewhere-else — is direct, checkable evidence
against reverse engineering, not an assertion asked to be taken on faith.

 Where triangulation stops mattering, and honesty about it. Triangulation proves the input 
 \(23/75\) ; it says nothing about the \(a_6\) heat-kernel coefficient built from it. The same three-route
discipline, applied to the graviton+ghost graded \(a_6\) value itself, currently fails to converge —
two independent routes disagree by \(31/48\approx0.646\) , six orders outside the pre-registered
 \(10^{-6}\) tolerance — and the honest report of that failure is itself evidence the discipline is
being applied uniformly rather than selectively invoked only when it produces a clean win. A
methodology that always agrees with itself would be weaker evidence than one that sometimes reports
a located, named disagreement.

 Insight 3 — a target-blind theorem-criterion is worth more than any number of numerical spot-checks

 The general principle. Numerical agreement between two computations is evidence of consistency
between those two computations; it is not, by itself, evidence that either computation is correct ,
because both could share the same systematic error. What breaks that symmetry is a test built from a
mathematical identity that the correct answer is guaranteed to satisfy exactly and a wrong answer is
 not guaranteed to satisfy at all — a theorem-criterion rather than a cross-check. The diagnostic
power of such a test comes precisely from its being derivable independently of the computation it is
testing: it does not care which route produced the candidate curvature tensor, only whether that
tensor obeys a constraint that follows from differential geometry itself.

 Why this matters here, concretely. The first Bianchi identity, \(R_{a[bcd]}=0\) , is an algebraic
constraint that any correctly assembled Riemann tensor on any manifold must satisfy identically —
zero free parameters, zero tolerance to tune, pass or fail. Running the corrected \(K_6\) curvature
tensor through this identity gives a residual of
$$
\approx2.5\times10^{-16},
$$
machine-precision zero. The diagnostic power of this specific test is demonstrated, not merely
claimed, by its history: the previous , buggy build of the same pipeline returned a first-Bianchi
residual of exactly \(1/7\) in the engine's \(\mathrm{Ric}=1/2\) normalization (equivalently \(1/6\) in
raw Killing-form units) — not numerical noise, but a clean, exact, wrong rational, which is the
signature of a genuine algebraic bug rather than round-off. That clean wrong answer is precisely what
led to locating and fixing the error that had produced \(|\mathrm{Riem}|^2/\mathrm{Scal}^2=31/147\) and
 \(\kappa=7/12\) in place of the corrected \(23/75\) and \(5/12\) . In other words: this specific test has a
documented track record of catching a real, quantitatively significant ( \(\sim31\%\) ) error, which is
exactly the property that separates a load-bearing correctness theorem from a decorative sanity check
that happens to always pass.

 The generalizable insight. Before trusting any downstream heat-kernel number built from a
curvature tensor, run the cheapest available theorem-criterion test first, and treat any residual
that is a clean nonzero rational (rather than at the floating-point noise floor) as evidence of a
located, fixable bug rather than as a small correction to average away. The same discipline recurs
one level up in the derivative-curvature sector (Insight 4 below): the box identity
 \(R_{abcd}\Box R^{abcd}=-|\mathrm{Riem}|^2\) -type contraction plays the identical theorem-criterion role
for the second-derivative curvature data, and the second Bianchi identity
 \(\nabla_{[a}R_{bc]de}=0\) plays it for \(\nabla\mathrm{Riem}\) itself — each is checked to zero
violations before the corresponding invariant is accepted into the \(a_6\) Gilkey basis. What the
first-Bianchi test does not do is validate the physics content of any \(a_6\) number built downstream;
it validates only that the curvature tensor feeding that computation is not the product of an
algebraic bug. This distinction — a theorem-criterion validates internal consistency of an input; it
does not validate a physics conclusion built from that input — is exactly the line the scope-firewall
certificate (Insight 5) draws one level higher up the chain.

 Insight 4 — homogeneous-but-not-symmetric is a genuine structural fact with a genuine consequence, not a technical annoyance to suppress

 The general principle. When a heat-kernel calculation is first developed and validated on
maximally symmetric spaces (round spheres), it is easy — and often invisible in the final write-up —
to implicitly assume that every space of interest shares the sphere's local-symmetry property
 \(\nabla\mathrm{Riem}=0\) . On a space that is homogeneous (has enough isometries to move any point to
any other) but not locally symmetric (the curvature tensor is not covariantly constant), this
assumption is simply false, and dropping the resulting nonzero derivative-curvature invariants from
the Gilkey basis produces a heat-kernel coefficient that is silently wrong at exactly the order where
those invariants first enter — which for the weight-6 (mass-dimension-6) basis is precisely \(a_6\) .
The insight is to recognize which specific new algebraic object this failure of local symmetry
introduces, compute it explicitly, and cross-check it against an independent identity, rather than
either ignoring it or waving at its existence without pinning it down.

 Why this matters here, concretely. \(K_6=SU(3)/T^2\) is homogeneous under the left \(SU(3)\) action
but is not a symmetric space — the flag manifold does not admit the extra involutive isometry that
symmetric-space status would require — so \(\nabla\mathrm{Riem}\neq0\) identically. The only surviving
weight-6 invariant built from a single covariant derivative of curvature is \((\nabla\mathrm{Riem})^2\) 
(a previously-proposed quartic candidate, \(256\,a^2(a^2-1)^2\) , has been checked and retracted; it
does not survive on this space). The exact value, computed via the Nomizu reductive-homogeneous-space
formalism and verified against the second Bianchi identity with zero violations, is
$$
|\nabla\mathrm{Riem}|^2=\frac14\ (\text{Killing-form normalization})\;=\;54\ (\text{trip-unit
normalization}),
$$
together with \(|\nabla\mathrm{Ric}|^2=0\) and \(|\nabla\mathrm{Scal}|^2=0\) (these vanish exactly because
 \(\mathrm{Ric}\) and \(\mathrm{Scal}\) are covariantly constant multiples of the identity at the Einstein
point, even though the full Riemann tensor is not covariantly constant — a subtlety worth stating
explicitly, since it would be easy to mistake the vanishing of \(\nabla\mathrm{Ric}\) and
 \(\nabla\mathrm{Scal}\) for a sign that \(\nabla\mathrm{Riem}\) also vanishes, which it does not).

 The insight that makes this a checked fact rather than an assertion. A second, structurally
different contraction — the box identity \(R_{abcd}\Box R^{abcd}=-54=-|\nabla\mathrm{Riem}|^2\) (in
trip-unit normalization) — reproduces the same invariant from a second-derivative self-contraction
rather than a first-derivative norm. Agreement between a first-derivative-norm computation and a
second-derivative self-contraction is a genuine, non-trivial internal consistency check precisely
because the two are different tensorial operations that happen to be related by an identity, not two
copies of the same calculation. This is the concrete evidence that the derivative-curvature sector,
which the corpus is careful to call "audited CHAIN-SAFE," was carried through the actual computation
rather than dropped: had the pipeline silently assumed local symmetry — the natural shortcut a
sphere-trained intuition would reach for — this entire term would read as zero, and the box-identity
cross-check above would immediately fail to match the (nonzero) direct norm.

 Why this is the honest reason the graviton \(a_6\) is still open, not a defect in the method. The
graviton \(a_6\) Route-A computation requires, at exactly this stratum, the Gelfand–Tsetlin
off-diagonal hopping-term matrix elements that mix the five Weyl-inequivalent \(T^2\) weight classes on
 \(\mathrm{Sym}^2_0\) — a direct consequence of \(K_6\) being homogeneous but not symmetric, since a
symmetric space would have no such hopping term to compute at all. These matrix elements are exact,
closed-form objects (the standard \(SU(3)\) Gelfand–Tsetlin lowering-operator formula), not a
research-level unknown; they are simply not yet enumerated. The insight to take away is that the
open computation-debt in the graviton \(a_6\) is a located, well-posed piece of linear algebra that
exists precisely because the not-locally-symmetric structure was taken seriously rather than
suppressed — the same rigor that correctly refuses to drop \((\nabla\mathrm{Riem})^2\) is what
correctly identifies the GT hopping term as the named remaining obstacle, rather than leaving a
silent, undetected error in its place.

 Insight 5 — the scope-firewall: an unbounded operator ladder is a structural, not a computational, obstruction

 The general principle. In an effective field theory expansion, each successive order in energy
(or curvature) introduces new independent operators — this is precisely what "non-renormalizable"
means. When the tower of required operators is provably unbounded, no finite truncation of that
tower can, even in principle, decide the behavior of the resummed, full theory: this is not a
statement about present computational limitations, it is a statement about what kind of mathematical
object a "UV completion" has to be. The insight is recognizing that a coefficient at any fixed finite
order — however exactly, however multiply cross-validated it is computed — belongs to a category
(local Seeley–DeWitt data) that is structurally different in kind from the category the UV-completion
question actually lives in (the existence of a well-defined non-perturbative limit, or resummation,
of the whole series). Confusing the two categories is a logical error, not a matter of needing more
computation.

 Why this matters here, concretely. The perturbative quantization of the interacting graviton on
this geometry generates, order by order in \(G_NE^2\) , the standard heat-kernel Seeley–DeWitt tower
 \(a_6<a_8<a_{10}<\cdots\) — provably unbounded, in the sense that each new even order introduces
curvature invariants of strictly higher mass dimension that cannot be re-expressed in terms of the
lower-order basis (Gilkey 1995 Thm 3.3.1/4.8.16; Avramidi 2000 Ch. 4; Vassilevich eq. 4.29 — the
standard references this heat-kernel ledger is built against). Even in the best conceivable outcome
for this gate's own internal work-plan — the Gelfand–Tsetlin hopping term is enumerated, the two
routes converge to within the pre-registered \(10^{-6}\) tolerance, a positivity functional \(P\) is
selected from among its currently three inequivalent readings, and \(P(a_6)\geq0\) is confirmed — the
resulting object is a single data point on an infinite tower. It cannot, by the structure of the
expansion itself, certify what the resummed series does above the cutoff, because the resummed
series is a statement about the tower's limit, not about any one of its terms.

 The insight that makes this a certificate rather than an excuse. It would be easy to read the
scope-firewall as a convenient way to avoid finishing a hard calculation. The opposite is true: the
firewall is exactly as binding whether or not \(a_6\) is ever pinned down, which is precisely why it is
graded as a terminal certificate rather than as a residual on the same footing as the open \(a_6\) 
value. It converts what might otherwise look like an open-ended, indefinitely deferrable computation
("we'll close this once we finish \(a_6\) ") into a correctly-scoped statement: computing \(a_6\) 
sharpens the operator-supply legs toward an unconditional, conditional-on-UQF-9 linearized
certificate, and it does this regardless of the eventual \(a_6\) value, but it can never by itself be
the missing UV-completion object. Naming this precisely — rather than leaving the boundary implicit —
is what forecloses the gate's most tempting and most frequently attempted mis-close: "we computed a
heat-kernel coefficient, therefore the gate is closed."

 Why this scope-limitation is itself evidence of rigor rather than weakness. A gate whose closure
criterion could be satisfied by finishing a finite, well-posed calculation would not be a fair
description of the strong-coupling UV-completion problem — no research program studying quantum
gravity (string theory, asymptotic safety, loop quantum gravity, causal sets) has a finished
construction of this object either. Correctly identifying that the category of object needed is a
non-perturbative fixed point (or a rigorous non-existence proof) — carried under the shared UQF-9
wall — rather than a heat-kernel coefficient, is itself a piece of honest, useful physics: it tells a
future researcher exactly what kind of result would actually move this gate, and equally precisely
what kind of result, however impressive on its own terms, would not.

 Insight 6 — dissolving a divergence class by finding its physical regulator, and knowing exactly how far that regulator reaches

 The general principle. Two logically distinct moves are often conflated under the single word
"regularize." The first is a mathematical prescription (a cutoff, a subtraction scheme) that removes
a divergence without necessarily corresponding to new physics — this changes bookkeeping, not
content. The second is identifying an actual physical mechanism, independently motivated and already
established elsewhere, that removes a class of divergence because nature genuinely does not permit
the runaway behavior the naive perturbative expansion predicts. The second kind of move is a real
physics result — it earns an anchor — precisely because it is not invented for the purpose of curing
this particular divergence; it is imported from independently verified physics and shown to apply
here. The insight is knowing which divergence classes such an imported mechanism can plausibly reach,
and stating that reach exactly, rather than letting a genuine partial result quietly expand to cover
territory it does not actually touch.

 Why this matters here, concretely. AXIOM-COSTFLOOR posits an irreducible quantum of cost/action 
— explicitly not a smallest length — applied Lorentz-invariantly to proper time. Its physical
grounding is not invented for this gate: it rests on three independently established results — the
Margolus–Levitin bound \(\tau\geq\pi\hbar/2E\) (a rigorous quantum-mechanical limit on the minimum time
for a state to evolve to an orthogonal state), the Landauer bound \(\Delta E\geq k_BT\ln2\) (the
thermodynamic cost of erasing one bit of information), and the Bekenstein bound
 \(S\leq2\pi k_BRE/\hbar c\) (the holographic bound on entropy in a bounded region) — each a
MEASURED-ANCHOR result in the existing literature, not a new assumption. Taken together, these three
bounds jointly imply that no physical process can probe arbitrarily short proper-time intervals at
arbitrarily fine resolution without violating one of them; positing that this joint implication holds
as an exact floor \(\Delta_0>0\) on cost/action is the axiom. Its Lorentz-scalar character — proved as
theorem T-LI, that the floored quantity transforms as a scalar rather than picking out a preferred
rest frame — is what allows the floor to be adopted without reintroducing a preferred-frame lattice
that would contradict the Lorentz covariance built into the rest of this 13-dimensional construction.
A minimal length , by contrast, is not a Lorentz scalar — lengths Lorentz-contract, so "the shortest
possible length is \(\ell_0\) " is a claim different inertial observers would disagree about unless a
preferred frame is smuggled in or a deformed dispersion relation is separately hypothesized. Stating
the floor on cost/action (equivalently, proper time) rather than on spatial length is therefore not
a stylistic choice; it is the specific feature that keeps the axiom compatible with the Lorentz
covariance the rest of the construction depends on.

 Exactly how far this reaches — the insight's sharpest edge. The axiom dissolves precisely one 
divergence class: the \(a\to0\) runaway behavior of the unbounded tail of the Seeley–DeWitt tower,
 \(\{a_8,a_{10},\dots\}\) , whose problematic behavior is tied to probing shorter and shorter proper-time
intervals in the heat-kernel parameter \(t\to0\) limit. A positive cost/action floor forbids that limit
from being taken to arbitrary precision, which is exactly the kind of physical mechanism that removes
this specific class of pathology rather than merely relabeling it. What the axiom does not do —
and this is the line that must never blur — is touch \(a_6\) itself. \(a_6\) is not a member of the
divergent tail; it is a single, finite, well-defined coefficient whose value is a computation debt 
(the un-enumerated Gelfand–Tsetlin hopping term), not a runaway that a cost/action floor could ever
regularize. Nor does the axiom exhibit, or even gesture toward, a non-Gaussian fixed point for the
interacting theory — dissolving one divergence class is not the same mathematical statement as
locating the fixed point (or proving its absence) that a genuine UV completion requires. Ten further
walls in the broader closure ledger stand completely untouched by this axiom.

 The companion insight that closes off the tempting shortcut. A natural next question is whether
the granularity scale itself might supply exactly the finite-grain cutoff a UV completion would need
— i.e., whether dissolving one divergence class with a cost/action floor is secretly most of the way
to solving the whole problem. A separate structural result forecloses this shortcut cleanly: the
compactification radius read off this geometry, \(R_0=(2\pi M_U)^{-1}\approx1.5915\times
10^{-17}\,\mathrm{GeV}^{-1}\) , is built entirely from gauge-coupling unification data — the two-loop
Standard-Model threshold closure \(\alpha_1(M_U)=\alpha_2(M_U)=\alpha_3(M_U)\) — and carries zero
gravitational input by construction. The ratio \(R_0/\ell_{\rm Planck}\sim194\) places it well above
the Planck length, but its origin is purely color/gauge, not gravitational. Consequently, any
attempted finite-grain resolution of the strong-coupling wall using "the natural length scale visible
on this geometry" is not an independent gravitational input — it relocates onto the isolated
sub-question P0 (does gravity own its own minimal scale, or does it inherit the color scale \(R_0\) ?)
inside the shared UQF-9 wall, rather than dissolving that wall directly. This is exactly the
discipline the whole section has been building toward: identify precisely which problem a genuine,
independently-grounded physical mechanism solves, and be equally precise about which superficially
adjacent problem it does not solve — and when a shortcut is tempting, check whether it secretly
reduces to a quantity, like \(R_0\) , that was never gravitational to begin with.

 How the six insights compose into the fixed grade, and no further

 None of the six insights above, individually, closes UQF-5C, and neither does their conjunction. What
they jointly establish, and what licenses the fixed grade REDUCED-TO-AXIOM / ANCHORED +1, is a
specific and narrow claim: the operator the geometry must supply for an interacting quantum theory of
gravity is completely and correctly constructed (Insight 1), the curvature data that operator's
higher-order behavior depends on is proved rather than merely computed (Insights 2–4), the reason a
finite piece of that data can never itself be the missing UV completion is understood as a structural
fact rather than a temporary gap (Insight 5), and the one physical mechanism that does provably
remove part of the naive divergence structure is correctly, narrowly scoped rather than over-claimed
(Insight 6). What remains after all six is applied is exactly one named, irreducible axiom pair —
 \(\{\Delta_0>0,\ \text{Lorentz-scalar proper-time floor}\}\) — standing between this fully-audited
construction and a genuine non-perturbative UV completion of the interacting graviton. That is the
"+1" the fixed grade records: not a promise that the remaining gap is small, but a precise statement
of exactly how much, and exactly what kind of, further physics is still owed. The gap is not a
smaller version of what has already been shown here; it is a categorically different object — the
shared UQF-9 construction — and no accumulation of the reasoning in this section, however carefully
it is extended, substitutes for actually building it.

 Evidence & reproducibility

 This section is the audit trail for UQF-5C: what was actually computed, against what, at what
tolerance; which numbers agree exactly, which disagree and by exactly how much; which claims are
protected by a negative control designed to catch a specific known failure mode; and — in enough
procedural detail that an independent physicist could rebuild every banked number starting only
from the frozen 13D arena and ordinary published results (Gilkey's heat-kernel theorems, the
Wang–Ziller/Nomizu curvature formulas for naturally reductive homogeneous spaces, standard
BRST/Faddeev–Popov gauge-fixing) — exactly how to reproduce the chain end to end. The gate does not
carry a dimensionful "model vs. measured, N-sigma pull" table, and that absence is stated plainly
here rather than manufactured away: every load-bearing quantity in this gate is either an exact
geometric rational, a topological integer, a machine-precision identity check, or an established
physical bound (Margolus–Levitin, Landauer, Bekenstein). Fabricating an uncertainty band around any
of these would be worse evidentiary practice than saying, correctly, that the pull-table category
does not apply.

 E.1 The shape of the evidence, stated before any number is quoted

 UQF-5C does not consume the four flavor/calibration anchors \(\{M_{\rm Pl},\,\alpha_i(M_Z),\,y_t,\,
|V_{us}|\}\) against a measured central value the way a Yukawa-sector or gauge-unification gate
does. Its evidence is structurally different, and conflating the two would misrepresent what has
actually been checked. Four distinct evidentiary channels operate here:

 Exact-rational geometric identities on \(K_6=SU(3)/T^2\) at the Killing-form normal-metric
 center \(\vec u=(1,1,1)\) , each checked by at least two structurally independent computational
 routes and required to agree either exactly (rational arithmetic) or to machine precision
 ( \(\sim10^{-14}\) – \(10^{-16}\) ).

 Topological and representation-theoretic integers — the ghost-corrected fiber weight \(65\) ,
 the little-group count \(\dim\mathrm{Sym}^2_0(SO(11))=65\) , the Euler characteristics
 \(\chi(K_6)=6\) , \(\chi(S^2)=2\) , \(\chi(S^1_Y/\mathbb{Z}_2)=1\) — cross-checked as integers that must
 match exactly , with literally zero tolerance band.

 A target-blind correctness theorem (the first-Bianchi identity) used as an internal unit
 test with no tunable parameter, which previously caught a real \(\sim31\%\) error in the curvature
 pipeline.

 An honestly reported route-disagreement on the one quantity — the graded graviton+ghost
 \(a_6\) heat-kernel coefficient — that would matter most if this gate were over-claiming: two
 independent routes disagree by an exact rational, \(31/48\) , six orders of magnitude outside the
 pre-registered tolerance, and this is reported as a live FAIL rather than smoothed over.

 Established, independently published physical bounds (Margolus–Levitin, Landauer,
 Bekenstein) as the external grounding for the single conditional axiom this gate is anchored on.

 There is accordingly no \(N\) -sigma pull to quote anywhere in this gate. Nothing here is a fit of
a free parameter to a measured central value with a propagated experimental error bar; every number
is either derived from the frozen metric with zero adjustable input, or cited as an established
theorem/bound from outside literature. Saying this plainly, rather than inventing a spurious
"agreement to \(n\sigma\) " statement, is itself part of the evidentiary discipline this gate is held
to.

 E.2 Reproducing the curvature ratio \(|\mathrm{Riem}|^2/\mathrm{Scal}^2=23/75\) from the root system up

 Setup. Begin from the \(A_2=\mathfrak{su}(3)\) root system in the Cartan basis \((h_1,h_2,h_3)\) 
with \(h_1+h_2+h_3=0\) : simple roots \(\alpha_1=(1,-1,0)\) , \(\alpha_2=(0,1,-1)\) , positive roots
 \(\{\alpha_1,\alpha_2,\alpha_1+\alpha_2=(1,0,-1)\}\) , half-sum \(\rho=(1,0,-1)\) with \(\|\rho\|^2=2\) in
the Killing normalization, and Weyl group \(S_3\) of order 6. The tangent space of the full flag
manifold \(K_6=SU(3)/T^2\) splits as \(T(K_6)=\mathfrak m_1\oplus\mathfrak m_2\oplus\mathfrak m_3\) ,
three real 2-planes, one per positive root, with \((-B)\) -orthonormal basis
 \(\{X_{ij}=E_{ij}-E_{ji},\,Y_{ij}=i(E_{ij}+E_{ji})\}/\sqrt{12}\) built from the Killing form
 \(B(X,Y)=6\,\mathrm{Tr}(XY)\) . The invariant metric is \(g_{K_6}(\vec u)=\sum_i u_i\langle
\cdot,\cdot\rangle_{\mathfrak m_i}\) , with squashing moduli in the Weyl-rigid chamber \(\vec u\in
[1/2,3/2]^3\) .

 Step 1 — reproduce the Einstein locus (zero-free-parameter check). The general-chamber Ricci
eigenvalues on scales \((x_1,x_2,x_3)\) are
$$
\mathrm{Ric}_1=\frac{x_1^2-x_2^2+6x_2x_3-x_3^2}{12x_1x_2x_3},\quad
\mathrm{Ric}_2=\frac{-x_1^2+6x_1x_3+x_2^2-x_3^2}{12x_1x_2x_3},\quad
\mathrm{Ric}_3=\frac{-x_1^2+6x_1x_2-x_2^2+x_3^2}{12x_1x_2x_3}.
$$
Setting \(\mathrm{Ric}_1=\mathrm{Ric}_2=\mathrm{Ric}_3\) and solving over the chamber returns
 exactly four solutions: the symmetric point \((1,1,1)\) and the three permutations of the
Kähler–Einstein point \((1,1,2)\) . This reproduces the classical classification of invariant Einstein
metrics on \(SU(3)/T^2\) (four total) — a check with zero adjustable parameters: either the algebra
returns exactly these four points, or the metric convention used downstream is wrong. This is
recorded as an independent validation of the engine, not a fitted feature.

 Step 2 — evaluate at the frozen witness \(\vec u=(1,1,1)\) . Substituting \(x_1=x_2=x_3=1\) :
 \(\mathrm{Ric}_i=5/12\) for all three eigenvalues (Killing-form normalization), and
 \(\mathrm{Scal}=\sum_k\dim(\mathfrak m_k)\,\mathrm{Ric}_k = 2\cdot3\cdot(5/12)=5/2\) .

 Step 3 — assemble the full Riemann tensor and its norm. Using the Nomizu curvature formula for
naturally reductive homogeneous spaces, contraction gives \(|\mathrm{Ric}|^2=25/24\) and
 \(|\mathrm{Riem}|^2=23/12\) , hence the scale-free ratios
$$
\frac{|\mathrm{Riem}|^2}{\mathrm{Scal}^2}=\frac{23/12}{25/4}=\boxed{\frac{23}{75}=0.3066666666666667},
\qquad
\frac{|\mathrm{Ric}|^2}{\mathrm{Scal}^2}=\frac{25/24}{25/4}=\frac16,
\qquad
\frac{\mathrm{Scal}}{\mathrm{Ric}_i}=6=\dim K_6.
$$

 Three independent routes, required to agree exactly. \(23/75\) is proved , not merely
computed, because it is the common output of three structurally unrelated methods: (i) direct
 \(SU(3)\) structure-constant contraction via the naturally-reductive curvature formula (pure
Lie-algebraic route); (ii) a curvature- free spectral heat-trace reconstruction from Peter–Weyl
eigenvalue sums, which never writes down a Riemann-tensor component; (iii) full Levi-Civita
connection and curvature assembly over the general 3-parameter metric \(\vec u\) , specialized to
 \((1,1,1)\) . All three return \(23/75\) exactly, with the acceptance criterion being exact rational
equality — not agreement "to a few percent." An error confined to any one method's internal
machinery (an algebra slip in route (i), a spectral truncation error in route (ii), an index-gymnastics
mistake in route (iii)) would show up as a disagreement between routes; it does not.

 Target-blindness self-check. Two candidate values had circulated before this ratio was
finalized: an earlier buggy-engine output of \(31/147\approx0.2109\) , and an unrelated firewall
placeholder \(\approx0.0667\) (numerically close to \(1/15\) ). The value actually returned, \(23/75=
0.30\overline{6}\) , matches neither . A reproducer can perform this exact check directly:
compute \(23/75\) , \(31/147\) , and \(1/15\) and confirm they are three distinct rationals with no simple
rescaling relating them. Landing on neither wrong prior candidate, while three unrelated
computational routes converge on the same third value, is direct, checkable evidence against
target-loading — not an assurance asked to be taken on faith.

 Cross-normalization check. The same geometry is separately recorded in a second absolute
normalization ("trip-unit," \(\mathrm{Ric}=1/2\) , \(\mathrm{Scal}=15\) , giving \(|\mathrm{Riem}|^2=69\) ,
 \(|\mathrm{Ric}|^2=75/2\) ). The dimensionless ratio must be normalization-invariant:
 \(69/15^2=69/225=23/75\) exactly — the identical rational, confirming both absolute normalizations
describe the same underlying geometry and that the physically meaningful content lives in the
scale-free ratios. The companion ratios \(\mathrm{Weyl}^2/\mathrm{Scal}^2=6/25\) and
 \(\mathrm{Scal}/\mathrm{Ric}_i=6=\dim K_6\) pass the identical cross-normalization test.

 E.3 The first-Bianchi machine-zero test — the gate's internal, parameter-free unit test

 Procedure. The algebraic first Bianchi identity \(R_{a[bcd]}=0\) is a theorem-level constraint
that any correctly assembled Riemann tensor on any manifold must satisfy identically, independent
of which curvature-computation route produced it. Running the assembled \(K_6\) curvature tensor at
the frozen witness through this identity is a pass/fail test with no free parameters and no
tolerance to tune: either the residual sits at floating-point round-off, or the pipeline has an
algebraic bug.

 Result. The corrected pipeline returns a first-Bianchi residual of
$$
\approx 2.5\times10^{-16},
$$
machine zero in double precision — a clean pass. Negative control (the diagnostic power of this
test): the earlier, buggy build of the identical pipeline returned a first-Bianchi residual of
 exactly \(1/7\) in the engine's \(\mathrm{Ric}=1/2\) normalization (equivalently \(1/6\) in raw
Killing-form units) — not numerical noise but a clean, exact, wrong rational, diagnostic of a
genuine algebraic error rather than round-off drift. This is the specific test that caught the
 \(\sim31\%\) curvature error that had been producing \(|\mathrm{Riem}|^2/R^2=31/147\approx0.2109\) and
the Einstein constant \(\kappa=7/12\) in place of the corrected \(23/75\) and \(\kappa=5/12\) .

 Reproduction instruction. Any physicist rebuilding the curvature pipeline from the root-system
data of section E.2 should run this identical Bianchi identity as the first gate, before trusting
any downstream heat-kernel number. A residual anywhere near \(1/7\) or \(1/6\) (rather than at the
 \(10^{-15}\) – \(10^{-16}\) floor) signals the same class of bug, and every downstream number computed
under it must be discarded and recomputed. Passing Bianchi certifies the input curvature tensor; it
does not by itself validate any \(a_6\) value built from that tensor — that is a separate,
further check (section E.6).

 E.4 Sphere cross-checks: validating the heat-kernel machinery where an independent answer exists

 \(K_6\) itself has no independently published \(a_6\) value in the literature to check against — no
external reference computes the graviton heat-kernel coefficient on the full \(SU(3)/T^2\) flag
manifold. The discipline applied instead is the standard one: validate the machinery on manifolds
where the exact answer is classical and independently known, before trusting it on the actual
target.

 Procedure. For round spheres \(S^n\) , the Seeley–DeWitt coefficients \(a_{2k}/a_0\) are known in
closed form from the classical Laplacian spectrum. The pipeline computes \(a_6/a_0\) for \(S^2\) , \(S^4\) ,
 \(S^6\) by two independent routes: (i) direct evaluation of the Gilkey local heat-kernel formula in
terms of the sphere's constant-curvature invariants, and (ii) the exact spectral sum over known
eigenvalues and degeneracies. The two routes are required to agree.

 Results (exact rationals, the two routes matching to \(\sim4\times10^{-14}\) ): 
$$
a_6(S^2)=\frac{4}{315},\qquad a_6(S^4)=\frac{74}{63},\qquad a_6(S^6)=\frac{1139}{63},\qquad
a_6^{\rm conf}(S^6)=\frac{5}{63}.
$$
As a further calibration, the lower \(S^6\) coefficients ( \(a_0=1\) , \(a_2=5\) , \(a_4=12\) in round-unit
normalization) are independently certified against the same known spectrum and used to calibrate
the \(a_4\) formula before it is trusted on \(K_6\) — the \(a_4=12\) match on \(S^6\) is the specific
calibration checkpoint recorded. Normalization-robust check: the scale-free combination
 \(a_4/a_2^2=66/125\) is Levi-Civita-immune (built from scalar invariants alone), so it must be
identical whichever of the two absolute normalizations of section E.2 the inputs are quoted in — a
reproducer can verify this directly.

 What this validates, and what it explicitly does not. These checks certify that the Gilkey
heat-kernel formula, the product/convolution rule \(a_{2k}(M_1\times M_2)=\sum_{i+j=k}a_{2i}(M_1)\,
a_{2j}(M_2)\) , and the coding of the \(\sim46\) -term cubic-curvature Gilkey basis at weight 6 are all
implemented correctly on manifolds with an independently known answer. They do not, by
themselves, certify the \(K_6\) graviton \(a_6\) value , because \(K_6\) is homogeneous but not locally
symmetric (unlike a round sphere, where \(\nabla\mathrm{Riem}\equiv0\) ), so it carries a nonzero
derivative-curvature sector that the sphere checks never exercise. This limitation is stated
explicitly rather than left to imply more validation coverage than actually occurred.

 E.5 The derivative-curvature sector: a possible blind spot, closed by an independent cross-check

 \(K_6=SU(3)/T^2\) is homogeneous but not locally symmetric, so \(\nabla\mathrm{Riem}\neq0\) , and the
weight-6 Gilkey basis therefore requires the derivative invariant \((\nabla\mathrm{Riem})^2\) (all
other first-derivative curvature invariants of this mass dimension vanish identically by symmetry
on this space). A naive "treat it like a sphere" shortcut would silently set this sector to zero;
the corpus instead carries it through explicitly and cross-checks it two structurally different
ways.

 Direct computation. Using the Nomizu reductive-homogeneous-space formalism, the second Bianchi
identity is verified with zero violations on the computed \(\nabla\mathrm{Riem}\) tensor, and its norm
is
$$
|\nabla\mathrm{Riem}|^2=\frac14\ \ (\text{Killing-form normalization}),\qquad
|\nabla\mathrm{Riem}|^2=54\ \ (\text{trip-unit normalization}),
$$
both values recorded in their respective absolute unit systems and derived from the same underlying
tensor. The companion invariants \(|\nabla\mathrm{Ric}|^2=0\) and \(|\nabla\mathrm{Scal}|^2=0\) vanish
exactly, as required since \(\mathrm{Ric}\) and \(\mathrm{Scal}\) are covariantly constant multiples of
the metric/a constant at the Einstein point even though the full Riemann tensor is not covariantly
constant; the companion graviton-endomorphism derivative object is \(|\nabla^{LC}E_{\rm grav}|^2=162\) 
(trip-unit).

 Independent cross-check via a different contraction pattern. A box identity supplies a second
route to the same invariant, built from a second-derivative self-contraction rather than a
first-derivative norm:
$$
R_{abcd}\,\Box R^{abcd} = -54 = -|\nabla\mathrm{Riem}|^2\quad(\text{trip-unit}).
$$
Because this identity is derived from a structurally different contraction, agreement in both sign
and magnitude with the direct norm computation is a genuine, non-trivial consistency check, not a
restatement of the same calculation. This sector is recorded as audited CHAIN-SAFE : it is
carried end-to-end through the \(a_6\) chain rather than silently dropped. Had the pipeline assumed
local symmetry by analogy with the sphere checks of section E.4, this entire sector would read as
zero, and the discrepancy would surface immediately as a failure of the \(R_{abcd}\Box R^{abcd}\) 
cross-check above — that failure mode is explicitly what this check is designed to catch, and it
does not occur.

 E.6 The route-disagreement audit on the graded \(a_6\) : reported as OPEN, without softening

 This is the single most important piece of negative evidence in the gate, reported here exactly as
it stands in the closure-of-record ledger, with no rounding or averaging.

 Procedure. Two structurally independent routes are run to compute the ghost-corrected graviton
 \(a_6\) heat-kernel coefficient:
- Route A (direct Gilkey/Lichnerowicz construction on the transverse-traceless graviton bundle
 \(\mathrm{Sym}^2_0(T)\) , dimension 20): consumes the certified Lichnerowicz endomorphism spectrum
 \(E_L\in\{1/6\ (\times6),\ 5/12\ (\times6),\ 7/6\ (\times6),\ 17/12\ (\times2)\}\) , with
 \(\mathrm{tr}\,E_L=40/3\) , \(\mathrm{tr}\,E_L^2=241/18\) , plus \(\Omega=\mathrm{Riem}\) , and requires
 the Gelfand–Tsetlin off-diagonal hopping-term matrix elements connecting the five
 Weyl-inequivalent \(T^2\) weight classes — an exact, well-posed computation via the standard
 \(SU(3)\) lowering-operator formula, but not yet enumerated .
- Route B (ghost + vector reconstruction): consumes the certified vector-bundle data
 ( \(E=\mathrm{Ric}\) , \(\mathrm{tr}(\Omega_{ab}\Omega^{ab})=-|\mathrm{Riem}|^2\) ) and the scalar
 backbone ratio \(a_6/a_2^3=7936/39375\) (banked and reproduced across 3+ independent engines) to
 assemble a ghost contribution.

 Result — the two routes disagree by an exact rational, six orders of magnitude outside
tolerance. Route A's graviton candidate returns \(-43/504\) ; a separate anchor-consistency check
for the same object independently expects \(-16/315\) — these two are already mutually inconsistent
before Route B is even brought in. Route B, moreover, turns out to compute the Bochner ghost
contribution \(149/1008\) (which implicitly sets \(E=0\) ), not the physical Faddeev–Popov ghost value
 \(-251/504\) (which correctly uses \(E=-\mathrm{Ric}\) ). The physical-ghost mismatch is exactly
$$
\left|-\frac{251}{504}-\frac{149}{1008}\right| = \frac{31}{48} \approx 0.6458,
$$
against a pre-registered tolerance of \(10^{-6}\) — a discrepancy roughly six orders of magnitude 
larger than the acceptance band. This is reported as FAIL / OPEN, not rounded, hidden, or averaged
away. The precisely located source of the debt is the graviton \(\mathrm{Sym}^2(T)\) Levi-Civita
off-diagonal leg (the known exact gap is \(2/21\) on the graviton sector, \(1/24\) on the vector sector)
— i.e., exactly the Gelfand–Tsetlin hopping-term computation flagged as not-yet-enumerated in Route
A. The trace \(\mathrm{tr}[a_6(L_{\rm grav})]\) is therefore OPEN / computation-debt , and a
previously circulated coefficient \(C\sim-6.39\) is explicitly not reproduced by either route and
must not be quoted as a live number.

 Why a disagreement is itself evidence, not merely an absence of evidence. A pipeline that always
agrees with itself, or that quietly patches a mismatch to hit a pre-existing target, would be worse
evidence than this. Route A and Route B are structurally independent — different bundle inputs
(graviton \(\mathrm{Sym}^2_0\) vs. vector/ghost reconstruction), different intermediate objects
(Lichnerowicz spectrum vs. scalar-backbone-plus-ghost assembly) — and their disagreement by a
 specific, exactly identified rational ( \(31/48\) ), traced to a specific, exactly identified 
missing computation (the GT off-diagonal matrix elements), demonstrates the two routes were run
independently rather than being disguised copies of each other. The debt is a named, closeable,
non-mysterious computation, not an unlocated black box.

 E.7 The exact-rational audit cascade (the parallel, later completion pass): what it banks and its explicit ceiling

 A separate, later audit/completion pass reproduces an internally self-consistent chain of exact
rationals via completion cycles, each value checked on at least two independent routes (the crux
geometric inputs on three, including from-scratch rebuilds). The chain:
$$
\begin{aligned}
\text{Graviton }\mathrm{Sym}^2(T_6)\text{ LC } a_6/a_0 &= -\frac{6373}{630}
= -\frac{3481}{360}\ (\text{algebraic}) + \left(-\frac{25}{56}\right)\ (\text{derivative sector}),\[2pt]
\text{Vector ghost }(E=-\mathrm{Ric})\ a_6/a_0 &= -\frac{251}{504}
= -\frac{713}{1260}\ (\text{algebraic}) + \frac{19}{280}\ (\text{derivative sector}),\[2pt]
\text{Physical defect } a_6 &= -\frac{7226}{35} = 21!\cdot!\left(-\frac{6373}{630}\right)
-12!\cdot!\left(-\frac{251}{504}\right),\[2pt]
\text{13D bulk graded } a_6\ (\text{keystone, at }K_2{=}5) &= -\frac{953329}{1260},\[2pt]
\mathbb{Z}_2\text{ smooth equivariant defect } \tfrac12 c_3^\gamma &= -\frac{337361}{840},\[2pt]
\text{AUD-0059 assembly} &= \frac12!\left(-\frac{953329}{1260}\right)+\left(-\frac{337361}{840}\right)
= -\frac{491353}{630}\quad(\text{EXACT, rational arithmetic}),\[2pt]
K_6\text{ scalar } a_6/a_0 &= \frac{992}{315} = \frac{8017}{2520}\ (\text{algebraic})
+\left(-\frac9{280}\right)\ (\text{derivative sector}).
\end{aligned}
$$

 Reproduction check for the assembly identity. Every additive step above is exact rational
arithmetic and can be checked by hand:
$$
\tfrac12\cdot\left(-\frac{953329}{1260}\right) = -\frac{953329}{2520},\qquad
-\frac{337361}{840} = -\frac{1012083}{2520},
$$
so
$$
-\frac{953329}{2520} - \frac{1012083}{2520} = -\frac{1965412}{2520} = -\frac{491353}{630}
$$
after reducing by \(\gcd(1965412,2520)=4\) . A reproducer should carry out exactly this reduction and
confirm the stated result. This kind of check has no tolerance band at all: rational arithmetic
either closes exactly, or it does not.

 The ceiling, stated without inflation. This cascade's status is explicitly AUDIT-CLOSED,
physics-OPEN — the audit certifies that the arithmetic is internally self-consistent and
multi-route-verified given its stated inputs , not that the physics question (a UV completion, or
even a single agreed value for the graviton+ghost graded \(a_6\) of section E.6) is resolved. It
carries a headline_green=false flag for exactly this reason. Several distinct objects appear in
this cascade and must never be conflated: the bulk \(a_6\) ( \(-953329/1260\) , or its AUD-assembled
form \(-491353/630\) ) is not the order-6 boundary \(a_6\) (which does not exist — section E.9), is
not the \(\mathbb{Z}_2\) -defect equivariant object ( \(-337361/840\) ), and is not the
 graded-Casimir supertrace of section E.6. A reproducer's checklist should include an explicit
verification that no downstream document has silently substituted one of these four for another —
this specific substitution has been a recurring failure mode this dossier must guard against.

 Reconciling the two layers. Section E.6 ("OPEN / FAIL_VALUE_MISMATCH") and this section
("AUDIT-CLOSED, physics-OPEN") are not in contradiction — they ask different questions of related
but distinct objects, and neither promotes the gate . Section E.6 adjudicates whether the
graviton+ghost graded total, computed via the two routes designed to cross-validate each other,
actually agrees; it does not. This section certifies that a larger, separately assembled
exact-rational chain (bulk, defect, and scalar sectors, related to each other via completion
cycles) is internally consistent and reproducible on its own terms. Both keep the gate un-closed;
the dossier carries both without picking one to hide the other.

 E.8 The color factor \(124/315\) : a caveat-carrying result, explicitly not yet a clean invariant

 The scalar \(K_6\) ratio \(b_3/b_0=124/315\) (the \(t^0\) /Seeley–DeWitt bracket ratio computed from the
actual \(K_6\) Peter–Weyl heat trace) was at one point labeled "DERIVED, dual-validated." That
status is deliberately downgraded here , as an act of evidentiary discipline, to
DERIVED-PENDING-INDEPENDENT-TARGET-BLIND-REPRODUCTION: the same metric-selected engine (selected at
 \(\mathrm{Scal}_{K_6}=7.5\) ) that is meant to validate this number is also the engine that produced
it, so the check is not yet independent of the thing it checks. Reproduction instruction: a
structurally different engine — one that does not share the metric-selection step — must reproduce
 \(124/315\) target-blind before it is promoted to a clean, unconditionally-DERIVED invariant. Until
that independent reproduction exists, every citation of \(124/315\) anywhere in this dossier carries
this caveat explicitly; dropping the caveat would silently upgrade an unverified number.

 E.9 Negative control: the \(\mathbb{Z}_2\) orbifold "boundary \(a_6\) " does not exist, and should not be found

 A specific negative result is preserved here as a live constraint. The order-6 mixed
Neumann \(\oplus\) Dirichlet boundary heat-kernel coefficient, which a naive reading of the
 \(S^1_Y/\mathbb{Z}_2\) orbifold might suggest is needed, does not exist in the published
mathematical literature — the standard boundary heat-kernel tower stops at \(a_5\) . The correct
treatment recognizes that the \(\mathbb{Z}_2\) action here (reflection \(\theta\mapsto-\theta\) , fixed
points \(\theta=0,\pi\) ) is a global isometric reflection on a closed manifold , not a
manifold-with-boundary problem: the twisted trace on \(S^1_R/\mathbb{Z}_2\) evaluates to exactly \(1\) 
(derivation: two fixed points \(\times\ 1/|1-dg|=1/|1-(-1)|=1/2\) each, summing to \(1\) ), \(t\) -independent,
with no boundary tower at all. The object legitimately emitted in the correct equivariant treatment
is the smooth equivariant defect \(\tfrac12c_3^\gamma=-337361/840\) (Donnelly's equivariant
heat-kernel formalism) — structurally different from a boundary coefficient. The checkable
consequence: any claim of a computed TOTAL (bulk \(a_6\) + boundary defect \(a_6\) ) for this gate is a
wrong-object artifact and must be refused; no such total has been, or should be, emitted. A
reproducer attempting to "complete" the calculation by searching the literature for an order-6
Neumann \(\oplus\) Dirichlet boundary coefficient will not find one — that failure to find it is the
correct, expected outcome, not a sign of reproducer error.

 E.10 Negative control: the refuted bulk magnitude, and the structural reason it fails

 An earlier claimed bulk value, \(-2.818\times10^{94}\,\mathrm{GeV}^6\) , for the graded \(a_6\) 
functional was tested and refuted at decision grade — a completed, reached verdict, retained
here as a frozen negative control that must never be revived. The mechanical reason it fails
(the reproducible part of the refutation): the heat-kernel expansion \(K(t)\sim(4\pi t)^{-d/2}
\sum_k a_{2k}t^k\) produces, in odd total spacetime dimension \(D=13\) , no finite local \(t^0\) slot for
a bulk magnitude at this mass dimension — the relevant object sits at the half-integer \(\zeta\) -function
pole \(s=7/2\) , which vanishes identically under dimensional regularization on an odd-dimensional
closed manifold, leaving no log or anomaly slot for a finite coefficient to occupy. A reproducer can
verify this structurally by enumerating the pole structure of the zeta function associated with
 \(L_{\rm grav}\) at half-integer argument and confirming the pole at \(s=7/2\) is absent for odd \(D\) —
a standard fact about heat-kernel zeta functions on odd-dimensional closed manifolds, not special
pleading invented for this gate. The number was additionally R2-contaminated (computed with a
known sign-flipped Rop engine bug) and scheme-anchored (backed into a value via an unjustified
scheme choice).

 A later, separate re-run of a nominally analogous dimensionful bulk quantity, using the
Bianchi-corrected pipeline, returns \(-2.995681680\times10^{94}\,\mathrm{GeV}^6\) — recorded
explicitly as a consistency-coefficient only , never gap-closing and still route-inconsistent, so
that its superficial numerical proximity to the old refuted value is never mistaken for a
rehabilitation of the refuted claim. In both cases the correct owed object is the finite trace
 \(\mathrm{tr}[a_6(L_{\rm grav})]\) , not a GeV \(^6\) magnitude — reiterating the point of section E.6.

 E.11 Frozen invariants: a checklist a reviewer should run first

 The following are preserved as permanent boundary markers, precisely because getting any one of
them wrong is diagnostic of silently reintroducing an already-identified error:

 \(|\mathrm{Riem}|^2(K_6)=23/12\) , ratio \(23/75\) — never \(31/147\) (the retired \(\sim31\%\) 
 curvature bug), never \(60\) (the unrelated \(S^6\) invariant).

 \(\mathrm{Scal}/\mathrm{Ric}_i=6=\dim K_6\) in both normalizations; \(|\mathrm{Ric}|^2/\mathrm{Scal}^2
 =1/6\) ; \(\mathrm{Weyl}^2/\mathrm{Scal}^2=6/25\) .

 Einstein constant \(\kappa=5/12\) — never \(7/12\) (retired buggy build value).

 First-Bianchi residual \(\approx2.5\times10^{-16}\) — never exactly \(1/7\) or \(1/6\) .

 \(\chi(K_6)=6=|S_3|\) , \(\chi(S^2)=2\) , \(\chi(S^1_Y/\mathbb{Z}_2)=1\) — topological, exact integers,
 zero tolerance.

 The bulk fiber triple: graviton dim \(91\) , ghost dim \(13\) , ghost-corrected weight \(65\) — never 
 the distinct graded pair \(67/11\) (Block-A graded weight \(67-2\cdot11=45\) ), which belongs only to
 the \(\mathbb{Z}_2\) -defect grading and must never be substituted into the bulk ledger.

 \(256a^2(a^2-1)^2\) (a retracted derivative-coupling coefficient) is dead; only
 \((\nabla\mathrm{Riem})^2\) survives as the nonzero derivative invariant on \(K_6\) .

 The graviton \(a_6\) is OWED at the named Gelfand–Tsetlin off-diagonal stratum — a named 
 computation debt, not a hidden gap; the scalar backbone ratio \(a_6/a_2^3=7936/39375\) , by
 contrast, is banked across 3+ independent engines and is not itself in dispute.

 \(\mathrm{tr}[a_6(L_{\rm grav})]\) not computed; \(C\sim-6.39\) not reproduced; the positivity
 functional \(P\) is UNSELECTED (three inequivalent readings exist) and no sign of \(P(a_6)\) is
 asserted.

 E.12 Reproducing the granularity-axiom grounding: a bibliographic, not computational, check

 Because the "+1" in the fixed grade REDUCED-TO-AXIOM / ANCHORED +1 rests entirely on the cost-floor
axiom pair \(\{\Delta_0>0,\ \text{Lorentz-scalar proper-time floor}\}\) , a reproducer should be able to
verify that its physical grounding is established physics, not a bespoke invention specific to this
program. This check is bibliographic and conceptual rather than computational, and is independent
of anything internal to this corpus:

 Margolus–Levitin bound , \(\tau\geq\pi\hbar/(2E)\) : the minimum time for a quantum system of
 energy \(E\) to evolve to an orthogonal state — a standard quantum-mechanics theorem.

 Landauer bound , \(\Delta E\geq k_BT\ln2\) : the minimum thermodynamic cost of erasing one bit —
 standard statistical mechanics.

 Bekenstein bound , \(S\leq2\pi k_BRE/(\hbar c)\) : the maximum entropy in a region of radius \(R\) 
 and energy \(E\) — a standard result linking quantum information to gravity/thermodynamics.

 None of these three is derived or re-derived inside this corpus; they are cited as external,
independently established anchors. A reproducer's check here is to confirm (i) that
AXIOM-COSTFLOOR — an irreducible action/cost quantum, applied to proper time rather than to a
spatial length — is a logically consistent reading of these three bounds taken together, and (ii)
that it does not smuggle in a preferred-frame spatial lattice, which would contradict Lorentz
invariance. The companion claim that the floored quantity transforms as a Lorentz scalar (theorem
T-LI) is the specific, checkable statement that keeps this axiom from being a hidden preferred-frame
assumption in disguise: a reproducer verifying T-LI should confirm the floor is stated as a bound on
proper-time intervals along a worldline (frame-independent), never on a coordinate distance in any
particular frame. A further checkable structural fact, cited but not re-derived here, is the ratio
 \(R_0/\ell_{\rm Planck}\sim194\) together with the derivation that \(R_0=(2\pi M_U)^{-1}\) is fixed
purely by gauge-coupling unification ( \(\alpha_1(M_U)=\alpha_2(M_U)=\alpha_3(M_U)\) , residual
 \(9.6\times10^{-11}\) ) with zero gravitational input — the reason every finite-grain attempt to
dissolve the strong-coupling wall relocates onto the shared UQF-9 sub-question P0 rather than
closing here.

 E.13 Full end-to-end reproduction checklist

 Collecting the above into one ordered procedure that an independent physicist can follow start to
finish: (1) build the \(A_2\) root system and the three \(SU(3)/T^2\) tangent root-planes; (2) solve for
the Einstein locus over the Weyl-rigid chamber \(\vec u\in[1/2,3/2]^3\) and confirm exactly four
solutions; (3) evaluate Ricci/Scalar/Riemann-norm at the symmetric witness \(\vec u=(1,1,1)\) by all
three independent routes of section E.2 and confirm exact rational agreement on
 \(|\mathrm{Riem}|^2/\mathrm{Scal}^2=23/75\) ; (4) run the first-Bianchi identity as a pass/fail gate
and confirm the residual sits at the \(10^{-15}\) – \(10^{-16}\) floor, never near \(1/7\) or \(1/6\) ; (5)
validate the Gilkey heat-kernel coding on round spheres \(S^2,S^4,S^6\) against the closed-form
spectral answer, confirming agreement to \(\sim4\times10^{-14}\) ; (6) carry the nonzero
derivative-curvature sector ( \(|\nabla\mathrm{Riem}|^2=1/4\) Killing-form / \(54\) trip-unit) through
the box cross-check \(R_{abcd}\Box R^{abcd}=-|\nabla\mathrm{Riem}|^2\) ; (7) assemble the BRST
ghost-corrected fiber weight \(91-2\cdot13=65\) and confirm it equals the independently derived
little-group count \(\dim\mathrm{Sym}^2_0(SO(11))=65\) ; (8) attempt the graviton+ghost graded \(a_6\) 
via Route A and Route B and confirm — honestly — that they currently disagree by \(31/48\) , tracing
the mismatch to the un-enumerated Gelfand–Tsetlin off-diagonal matrix elements; (9) check the
exact-rational audit-cascade assembly identity \(\tfrac12(-953329/1260)+(-337361/840)=-491353/630\) 
by hand; (10) confirm the order-6 boundary coefficient is absent from the published literature and
that the smooth equivariant defect \(-337361/840\) is the correct, distinct object in its place; (11)
confirm the refuted bulk magnitude \(-2.818\times10^{94}\,\mathrm{GeV}^6\) fails structurally because
 \(D=13\) is odd (no finite local \(t^0\) slot at the relevant half-integer zeta pole \(s=7/2\) ); (12)
confirm the granularity axiom's grounding in Margolus–Levitin/Landauer/Bekenstein and its
Lorentz-scalar (proper-time, not length) character, together with the gauge-only origin of \(R_0\) .

 A reproducer who completes steps (1)–(7) and (9)–(12) will reconstruct every banked, DERIVED or
DERIVED-GIVEN-E number in this dossier exactly. A reproducer who honestly attempts step (8) will
land on the same OPEN / FAIL_VALUE_MISMATCH this dossier reports — and that outcome is itself the
correct, faithful result of the reproduction, not a sign that something went wrong. It confirms that
the stated grade — REDUCED-TO-AXIOM / ANCHORED +1 on the conditional axiom pair \(\{\Delta_0>0,\) 
Lorentz-scalar proper-time floor \(\}\) , with the constructive UV completion of the interacting
graviton itself standing as an open global wall shared with every other approach to quantum gravity
— is the honest terminal state of the evidence exactly as it currently stands, neither more nor
less.

 Open gaps & the specialist closure path

 Fixed grade for this gate, stated once and never mutated by anything below: REDUCED-TO-AXIOM /
ANCHORED +1. The published row is anchored on the conditional axiom pair {Δ₀ > 0 (a positive
cost/action floor), Lorentz-scalar proper-time floor}; that is the one named, irreducible "+1" the
grade records, and it is TERMINAL + RESIDUALS-SHOWN. Nothing in this section reopens, downgrades, or
upgrades that pill. What follows is the honest specialist work-plan behind it: six residuals
(H1–H6), each pinned to (a) the precise open object, (b) why it is hard and the specific traps
already caught, (c) exactly what closes it — target-blind, with both a success criterion and what a
refuting result looks like, (d) the machinery to start from, described in full rather than
file-cited, and (e) the leverage — what else in the ledger moves if this residual falls. The six are
ordered by tractability, not by their power to move the roll-up: exactly one of them (H1) touches the
5C grade at all, and it does so only by construction — never by a per-gate plug.

 The frame that must not slip while reading what follows: the operator itself is not in question. On
the frozen 13-dimensional arena \(\mathcal{M}_4\times K_6\times S^2\times S^1_Y/\mathbb{Z}_2\) 
( \(K_6=SU(3)/T^2\) , the full \(A_2\) flag manifold; \(D=4+6+2+1=13\) ), de-Donder gauge-fixing the metric
fluctuation \(h_{MN}\) and adding the Faddeev–Popov ghost sector produces a fully specified
Lichnerowicz-type operator \(L_{\rm grav}=-(\nabla^2+E)\) on \(\mathrm{Sym}^2(T)\oplus(\text{FP ghost})\) ,
whose 4D massless spin-2 zero mode is the graviton with a Kaluza–Klein tower above it (leg 5A,
DERIVED-GIVEN-E). The ghost-corrected fiber weight \(91-2\cdot13=65=\dim\mathrm{Sym}^2_0(SO(11))\) is
forced by BRST nilpotency, not chosen. None of that is open. What is open is everything downstream of
asking what the interacting , all-orders quantum theory built on that operator does once the loop
expansion in \(G_NE^2\) stops converging above the cutoff — and that is where every item below lives.

 H1 — the roll-up itself: constructive UV completion (shared with UQF-9, inherited by UQF-14)

 (a) The precise open object. The 5C roll-up is UQF-9: a genuine, non-perturbative, constructive
UV completion of the fully interacting graviton on this specific frozen background — equivalently, a
demonstration that above the cutoff (read off the geometry at \(M_*\approx7.467\times10^{16}\) GeV via
 \(M_*^{11}=M_{\rm Pl}^2/\mathrm{Vol}(X_{\rm active})\) , or on the alternate threshold-closure reading
 \(M_U\sim1.0\times10^{16}\) GeV where \(\alpha_1=\alpha_2=\alpha_3\) closes to residual \(9.6\times10^{-11}\) )
the perturbative series in \(G_NE^2\) is replaced by a well-posed, UV-consistent quantum theory, whether
through resummation to a fixed point, embedding in a manifestly finite construction, or an equivalent
non-perturbative definition. This is emphatically not "compute more heat-kernel coefficients." The
operator is already fully specified; the open object is what happens to the interacting completion of
that operator's quantum field theory once loop corrections are resummed to all orders above the
cutoff, on this exact geometry, in this exact rulebook (de-Donder gauge, FP ghosts, \(\overline{\rm MS}\) /
heat-kernel scheme).

 (b) Why it is hard, and the specific traps. This is hard for the reason it is hard for every
research program in fundamental physics: perturbative quantization of Einstein gravity is
non-renormalizable, and each loop order's counterterm structure introduces a new independent
higher-derivative operator, so the naive expansion never closes on a finite parameter set. The
Seeley–DeWitt/heat-kernel ladder makes this structurally visible on this geometry: the sequence
 \(a_6<a_8<a_{10}<\cdots\) is unbounded, and by the nonseparability screen no finite subset of it can
determine whether the resummed theory converges — this is exactly why the scope-firewall
certificate ("one heat-kernel coefficient is not a UV completion") is terminal as a scope wall rather
than a hole to be patched. Four specific traps are already on record and must not be repeated:
- The signature mis-close. Treating "we computed \(a_6\) exactly" — or even a hypothetical future
 "we computed \(a_6,a_8,a_{10}\) exactly" — as evidence of closure. It is not: the ladder is infinite,
 and completion is a statement about the limit/resummation of the whole tower, not about any finite
 number of its terms. This is the gate's single most important named failure mode.
- False openness via a "solvable" framing. Marking H1 bounded:true with
 what_would_close_it = "solve the UV completion of quantum gravity" looks like an honest bound but
 is a category error — it restates the problem rather than bounding it. H1 is a wall , not a horse
 that runs faster with more compute or a cleverer gauge choice.
- Borrowing a partial result and calling it a completion. For instance, treating a candidate
 asymptotic-safety fixed point found in a truncated theory space elsewhere in the literature as if
 truncation-independence had already been shown for it. Truncation-dependence is exactly the open
 technical question in that program generally; importing an untested truncation here would smuggle
 the same unresolved issue into this geometry under a new name — a textbook instance of
 target-anchoring dressed as borrowed rigor.
- Conflating a positivity check with a completion. Even a fully correct, computed d=13
 graviton-minus-ghost \(a_6\) vector that passes a well-defined positivity functional (H4, below) is at
 most a linearized, conditional certificate. It says nothing about the resummed, fully interacting
 theory and must never be reported as though it did.

 (c) What closes it, target-blind, with success and refutation criteria. The construction to build
— once, and shared across UQF-9/UQF-14/5C — is a functional renormalization-group (FRG) /
asymptotic-safety-style flow evaluated on this specific frozen background, or an equivalent
non-perturbative definition (a lattice regularization respecting the isometries of
 \(K_6\times S^2\times S^1_Y/\mathbb{Z}_2\) , or a resurgence/Borel-summation treatment of the loop
series), run target-blind — i.e. without steering the truncation, ansatz, or regulator scheme toward
a preferred answer. Exactly three outcomes are legitimate, and all three are genuine termini rather
than partial credit:
1. Constructive close. A truncation-independent non-Gaussian UV fixed point is exhibited for the
 graviton-plus-matter beta functions on this geometry, stable under systematically enlarging the
 operator basis (the acid test separating a real fixed point from a truncation artifact). Success
 criterion: fixed-point couplings converge as truncation order increases, cross-checked against at
 least two independent regulator choices (e.g. a sharp cutoff versus an optimized/Litim-type
 regulator) agreeing within a pre-registered tolerance. This would promote H1 from OPEN (global
 wall) to CLOSED, and it would be the single largest event this framework — or arguably any
 framework — could report, since no research program has answered this question for any geometry.
2. Closed-negative — a valid terminus. A rigorous proof, not a truncation artifact and not a
 non-convergent numerical search misreported as a negative, that no non-Gaussian fixed point exists
 on this geometry for the graviton beta functions under a stated, defensible class of regulators.
 This is legitimate, terminal CLOSED-NEGATIVE content: it would not "solve quantum gravity," but it
 would settle, for this specific frozen shape, that the asymptotic-safety route is unavailable —
 real, falsifiable, publishable science.
3. Neither runnable — emit a WALL RECORD. If neither outcome is executable with currently
 available machinery (the realistic expectation, since this is precisely where every other
 research program is also stuck), the correct output is neither silence nor a fabricated partial
 answer: it is a WALL RECORD that names the specific construction attempted, classifies it
 precisely as OPEN, records the hidden bridge (the specific mathematical object — e.g. the resummed
 beta-functional itself — whose non-existence-of-known-technique is the actual obstruction), and
 marks it explicitly SHARED across UQF-9, UQF-14, and 5C. This is the outcome currently on record
 and the one the brief flags as most likely given the state of the art.

 A refuting result looks concretely like this: a demonstrated obstruction — a proof that the graviton
beta functions on this specific \(K_6\times S^2\times S^1_Y/\mathbb{Z}_2\) background develop a genuine
(non-artifactual) pole or instability at finite RG scale surviving every tested regulator, or a no-go
theorem tied specifically to the \(D=13\to4\) dimensional-reduction structure forbidding a UV fixed
point regardless of truncation. Either is outcome (2) above — a CLOSED-NEGATIVE, not a dossier
failure, and not evidence against the rest of the framework.

 (d) Machinery to start from. The natural entry point is the Wetterich-equation functional RG
flow adapted to this background: an effective average action \(\Gamma_k[g,h]\) built on the metric
split already fixed here (graviton fluctuation \(h_{MN}\) in de-Donder gauge, with the same
Faddeev–Popov ghost sector already forced by BRST), with a regulator \(R_k\) respecting the isometries
of \(K_6=SU(3)/T^2\) , \(S^2\) , and the \(\mathbb{Z}_2\) orbifold parity \(\theta\mapsto-\theta\) on \(S^1_Y\) .
The flow equation
$$
\partial_k\Gamma_k=\tfrac12\,\mathrm{STr}!\left[(\Gamma_k^{(2)}+R_k)^{-1}\partial_kR_k\right]
$$
would be evaluated using the same ghost-corrected supertrace structure already fixed at
 \(91-2\cdot13=65\) , but extended off the single-heat-kernel-coefficient truncation to a genuinely
running ( \(k\) -dependent) effective action. The already-tabulated KK spectrum supplies the mode sum any
such flow must be evaluated over: the \(K_6\) Casimir tower \(C_2(p,q)=(p^2+q^2+pq+3p+3q)/3\) with
dimension \(\dim(p,q)=(p+1)(q+1)(p+q+2)/2\) (giving, e.g., the adjoint \(\mathbf 8\) at \(C_2=3\) , the
 \(\mathbf{27}\) at \(C_2=8\) ), the \(S^2\) monopole tower \(\ell(\ell+1)/R_2^2\) with \(\ell\geq|N|/2\) , and the
 \(S^1_Y/\mathbb{Z}_2\) momentum lattice \(p_\theta=(n+\alpha)/R_Y\) . This is a large, multi-year research
program in its own right — not a computation that closes with more careful bookkeeping of objects
already in hand.

 (e) Leverage. This is the single highest-leverage item in the entire ledger for this gate: it is
the only item whose closure changes the 5C roll-up at all. Closing H1, in either direction, resolves
UQF-9 and UQF-14 simultaneously (H3, below, inherits H1's disposition automatically) and would
upgrade every "conditional on UQF-9" certificate elsewhere in the ledger — including whatever H4
eventually produces — from conditional to unconditional. No other item on this list has this
property.

 H2 / P0 — does gravity own an intrinsic minimal length, or does it inherit the color scale \(R_0\) ?

 (a) The precise open object. The compactification radius \(R_0\equiv(2\pi M_U)^{-1}
=1.591549430918954\times10^{-17}\ \mathrm{GeV}^{-1}\) is constructed purely from gauge-coupling
unification data — the two-loop Standard-Model threshold closure \(\alpha_1(M_U)=\alpha_2(M_U)
=\alpha_3(M_U)\) , residual \(9.6\times10^{-11}\) — and carries zero gravitational input by
construction. The banked Layer-2 screen T-DEEP records \(R_0/\ell_{\rm Planck}\sim194\) : \(R_0\) sits
parametrically above the Planck length but is not derived from any graviton-sector quantity. P0 asks
whether the graviton sector independently generates its own intrinsic minimal length/scale — the kind
of object a genuine UV completion of gravity specifically would be expected to produce — or whether
the only finite-grain scale available to gravity on this geometry is the inherited, purely
gauge-sourced \(R_0\) .

 (b) Why it is hard, and the traps. The central trap is circularity. Any attempt to "derive" a
graviton minimal length using \(R_0\) -derived inputs — for instance, reading a graviton cutoff off the
KK tower spacing, which is itself set by \(R_6=R_0\) at the chamber center — risks smuggling the gauge
scale in and then declaring it "gravity's own," which is target-anchoring dressed as a derivation, not
a derivation. The question must be posed so "inherits \(R_0\) " and "owns an independent scale" are
genuinely distinguishable outcomes rather than two descriptions of the same number. A second, related
trap: since \(R_6=R_2=R_0\) at the frozen chamber center \(\vec u=(1,1,1)\) (all internal radii coincide
before RG running), a careless argument could conflate "the graviton KK tower is spaced at \(1/R_6\) " —
a triviality true of every bulk field, gravitational or not — with "gravity has generated a scale of
its own." It has not; every bulk field on this geometry shares that radius.

 (c) What closes it, target-blind, with success and refutation criteria. The determination must
turn on finding, or rigorously ruling out, a graviton- specific invariant — a quantity that appears
only in the gravitational sector's quantum corrections (e.g. a coefficient in the resummed graviton
self-energy, or a genuinely gravitational anomaly scale) that is numerically or structurally
independent of \(R_0\) . Success criterion for "owns its own scale": exhibit a dimensionful or
dimensionless graviton-sector quantity whose value cannot be re-expressed as a function of \(R_0\) and
Standard-Model data alone — this would open a genuinely new finite-grain route into H1. Success
criterion for "inherits \(R_0\) ": a proof, not merely an absence of counterexample, that every
graviton-sector scale reduces order by order in the loop expansion to functions of \(R_0\) and the
already-anchored \(\{M_{\rm Pl},\alpha_i,y_t,|V_{us}|\}\) — this would confirm (not merely be consistent
with) T-DEEP's banked conclusion that Scale offers no finite-grain shortcut. A refuting result for
"inherits" is exactly the graviton-specific invariant described above; a refuting result for "owns its
own scale" is a demonstration that every candidate invariant proposed collapses, on inspection, into a
function of \(R_0\) — the null result that has held so far.

 (d) Machinery to start from. Start from the certified Lichnerowicz spectrum already fixed on
 \(\mathrm{Sym}^2_0\) at the Killing-form center: \(\mathrm{tr}\,E_L=40/3\) , \(\mathrm{tr}\,E_L^2=241/18\) ,
with eigenvalues \(\{1/6\,(\times6),\,5/12\,(\times6),\,7/6\,(\times6),\,17/12\,(\times2)\}\) . Compute
the one-loop graviton self-energy (or effective action) using this spectrum and classify whether any
UV-sensitive coefficient in it depends on the internal geometry in a way that cannot be absorbed into
a redefinition of \(R_0\) itself — i.e., whether the combination in which \(R_0\) appears is always the
same combination already fixing the gauge unification scale, or whether an independent combination
appears. This is a bounded, well-posed calculation, and the cleanest of the six items to execute
because it classifies scaling structure rather than requiring the full non-perturbative construction.

 (e) Leverage. P0 is the cleanest bounded handle on the H1 wall. If it resolves to "inherits
 \(R_0\) " — the outcome the existing T-DEEP screen already favors — it does not close H1, but it closes
off an entire class of would-be shortcuts: anyone proposing a finite-grain resolution of the graviton
UV problem via "the natural length scale of the compactification" would be shown, rigorously, to be
re-deriving \(R_0\) under a new name. If it resolves to "owns its own scale," it opens a genuinely new
avenue into H1 that does not currently exist in the ledger — high leverage, though the brief is
explicit that this is the less likely of the two outcomes given the banked T-DEEP result.

 H4 — the d=13 graviton-minus-ghost \(a_6\) vector and the positivity functional \(P\) 

 (a) The precise open object. Two distinct objects, in sequence. First, the trace
 \(\mathrm{tr}[a_6(L_{\rm grav})]\) in \(d=13\) — the physical, ghost-corrected graviton heat-kernel
coefficient — is not computed ; a previously claimed coefficient sign \(C\sim-6.39\) is explicitly
 not reproduced and must not be asserted. Second, and logically prior to that computation even
being meaningful as a certificate, a positivity functional \(P\) that would turn a computed \(a_6\) value
into a pass/fail statement about the linearized theory is unselected : three inequivalent candidate
readings of what "positivity of \(a_6\) " should mean at this order currently exist, with no argument yet
fixing which is physically correct. Until \(P\) is named and defended, the statement " \(P(a_6)\geq0\) " is
not a well-defined predicate, let alone an evaluated one.

 (b) Why it is hard, and the traps. Two compounding obstacles. First, a documented engine bug: the
sign of the curvature operator Rop is flipped in the computational pipeline (the "R2 bug"), and
this must be fixed as a drop-in correction before any d=13 number can be trusted — using the current
buggy engine and reporting a number would repeat exactly the class of error the first-Bianchi
machine-zero test already caught once (the curvature bug that flipped \(31/147\to23/75\) and
 \(\kappa=7/12\to5/12\) , at the ~31% level). Second, a cross-script contradiction already on record:
one independently-written implementation of the Gilkey \(a_6\) formula passes its internal
self-consistency check while a second, separately-coded cross-check implementation fails on the same
underlying machinery, emitting \(-8/405\) for the
 \(a_6(S^2)\) calibration case where the certified value is \(4/315\) . The two cannot both be trusted
simultaneously, and this contradiction must be resolved before any d=13 output from either is treated
as meaningful. Specific traps to avoid: (i) trusting a d=13 number produced before the R2 fix; (ii)
trusting whichever of the two contradictory scripts happens to agree with a prior expectation — this
is target-loading, exactly the sin the target-blind discipline exists to prevent; (iii) treating a
computed \(a_6\) , even a fully correct one, as itself closing anything beyond a conditional, linearized
certificate; (iv) fabricating a sign or magnitude for \(\mathrm{tr}[a_6]\) or for \(C\) under any pressure
to fill the gap — the discipline is that an honest "not computed" beats an invented value, without
exception.

 (c) What closes it, target-blind, with success and refutation criteria. Three sequential steps.
 Step 1 — select \(P\) . Establish, from first principles (e.g. reflection positivity of the
Euclidean path integral, or unitarity of the truncated propagator), which of the three candidate
readings of the positivity predicate is physically correct, argued independently of what value \(a_6\) 
turns out to have — a value-free adjudication rule. Success: a named, defended choice of \(P\) that a
referee could apply without knowing the answer in advance. Step 2 — fix the engine and resolve the
cross-script contradiction. Patch the sign-flipped Rop operator, then re-run both computation
paths on the calibration cases where the answer is already known exactly — the certified sphere
ledger \(a_6(S^2)=4/315\) , \(a_6(S^4)=74/63\) , \(a_6(S^6)=1139/63\) , \(a_6^{\rm conf}(S^6)=5/63\) — until both
agree with the certified values to the pre-registered tolerance. Success: both routes reproduce all
four sphere calibration values exactly; failure after the R2 fix indicates a second, currently
unidentified bug and must be reported as such, not patched around silently. Step 3 — compute the
d=13 graviton-minus-ghost \(a_6\) vector target-blind on the corrected, cross-validated machine lane,
and evaluate \(P(a_6)\) . Two outcomes, both legitimate: \(P(a_6)\geq0\) lifts legs 5A/5B to
certificate-grade, but strictly conditional on UQF-9 — it still never reaches 5C, by the
scope-firewall established elsewhere; \(P(a_6)<0\) refutes the linearized gate at decision grade —
a genuine, useful negative result, not a failed calculation.

 (d) Machinery to start from. The Gilkey heat-kernel formalism for the \(a_6\) (mass-dimension-6)
Seeley–DeWitt coefficient (Gilkey Thm 3.3.1 / Thm 4.8.16; Avramidi Ch. 4; Vassilevich eq. 4.29),
expressed in the standard curvature-invariant basis of roughly 46 independent terms built from \(R\) ,
 \(\mathrm{Ric}\) , \(\mathrm{Riem}\) , \(E\) , \(\Omega\) , and their covariant derivatives. The certified inputs
already available at the Killing-form center to feed this basis: the nine weight-6 invariants
 \(\mathrm{Scal}^3=125/8\) , \(\mathrm{Scal}\,|\mathrm{Ric}|^2=125/48\) ,
 \(\mathrm{Scal}\,|\mathrm{Riem}|^2=115/24\) , \(|\mathrm{Ric}|^3=125/288\) ,
 \(\mathrm{Ric}^{ab}\mathrm{Ric}^{cd}R_{acbd}=125/288\) , \(\mathrm{Ric}^{ab}R_a{}^{cde}R_{bcde}=115/144\) ,
the cubic-Riemann chain \(K_1=R_{ab}{}^{cd}R_{cd}{}^{ef}R_{ef}{}^{ab}=-113/72\) , the cubic-Riemann
ladder \(K_2=R_{abcd}R_{aecf}R_{ebfd}=-5/72\) , and the derivative invariant
 \(|\nabla\mathrm{Riem}|^2=1/4\) (Killing-form normalization; equal to \(54\) in the trip-unit
normalization, with the box cross-check \(R_{abcd}\Box R^{abcd}=-54\) ) — together with the certified
 \(E_L\) spectrum on \(\mathrm{Sym}^2_0\) and the ghost sector's \(E=\mathrm{Ric}=(5/12)\,\mathrm{Id}\) ,
 \(\Omega=\mathrm{Riem}\) , \(\mathrm{tr}(\Omega_{ab}\Omega^{ab})=-|\mathrm{Riem}|^2=-23/12\) . Note that
 \(|\nabla\mathrm{Riem}|^2\neq0\) is itself physically consequential: it is what it means for \(K_6\) to be
homogeneous but not locally symmetric, and it is precisely why the graviton \(a_6\) leg carries the
Gelfand–Tsetlin ladder term rather than vanishing identically as it would on a symmetric space. The
missing piece specifically is that Gelfand–Tsetlin off-diagonal hopping term : the first-order
(hopping) piece of the Lichnerowicz operator on \(\mathrm{Sym}^2_0\) mixes the five Weyl-inequivalent
 \(T^2\) weight classes through \(SU(3)\) representation matrix elements between adjacent GT patterns.
These matrix elements are exactly computable in closed form via the standard GT lowering-operator
formula (a square root of a product of pattern-entry differences) — this is a well-posed, finite
linear-algebra computation, not a research-level unknown. It is simply not yet enumerated, and
enumerating it is the concrete, mechanical task that unblocks Route A of the two-route \(a_6\) ledger
(the known located debt: LC gap \(2/21\) on the graviton leg, \(1/24\) on the vector leg).

 (e) Leverage. Closing H4 does not touch H1/5C — the scope firewall forbids that regardless of
outcome — but it is the highest-leverage item for tightening legs 5A/5B: it would convert "the
operator is supplied" into "the operator is supplied and its first nontrivial interacting-sector
coefficient is verified consistent with (or is shown to refute) linearized unitarity, conditional on
UQF-9." It would also resolve the standing two-route disagreement — Route A giving \(-43/504\) against
Route B's Bochner-ghost value \(149/1008\) (which uses \(E=0\) , not the physical \(E=-\mathrm{Ric}\) ),
producing a physical-ghost mismatch of \(|-251/504-149/1008|=31/48\approx0.646\) , six orders outside the
pre-registered \(10^{-6}\) tolerance — that currently keeps even the linearized graviton+ghost value at
OPEN/FAIL_VALUE_MISMATCH on the closure-of-record ledger.

 H5 — independent reproduction of the color factor \(124/315\) and the shared heat-kernel scheme object

 (a) The precise open object. Two related items. First, the scalar \(K_6\) ratio \(b_3/b_0=124/315\) 
(the Seeley–DeWitt \(a_{d/2}\) -bracket ratio computed from the actual \(K_6\) Peter–Weyl heat trace) was
previously labeled "DERIVED dual-validated" but has been correctly downgraded to
DERIVED-PENDING-INDEPENDENT-TARGET-BLIND-REPRODUCTION, because the engine that produced it is the
same metric-selected engine (selected at \(\mathrm{Scal}_{K_6}=7.5\) ) whose curvature output the result
is meant to validate — a circularity that must be broken by an independent route, not asserted away.
Second, the shared one-loop heat-kernel scheme object — the \(\overline{\rm MS}\) /heat-kernel
convention choice underlying this gate and, in the same family, the SG-7 \(\delta\) object and the
constant \(c_{\rm loop}\) elsewhere in the ledger — has not been settled target-blind and remains open
as either a to-be-derived object or a to-be-named axiom.

 (b) Why it is hard, and the traps. For \(124/315\) , the trap is subtle: a second run of the same 
engine, even one reproducing \(124/315\) to high precision, does not count as independent
reproduction — a captured terminal log or a re-execution of an identical pipeline is a named failure
mode ("a captured terminal log ≠ an independent reproduction"). True independence requires a
structurally different computational route: a different basis choice, a different regularization
(e.g. zeta-function regularization rather than proper-time heat-kernel truncation), or a hand
computation via the Peter–Weyl decomposition directly, rather than through the numerical engine tuned
(metric-selected) to produce the curvature data in the first place. For the scheme-object item, the
trap is reverse engineering: choosing a scheme specifically because it reproduces a wanted magnitude
elsewhere in the ledger is explicitly forbidden — the " \(\kappa^3/\pi\) kill-test" failure mode — and
any scheme that only works because it was tuned to a target must be refused and marked as
 relocated , not closed.

 (c) What closes it, target-blind, with success and refutation criteria. For \(124/315\) : reproduce
the value on a structurally independent engine. Success is exact rational agreement to the same
precision from a route sharing no code path — and ideally no numerical-optimization step — with the
original metric-selected engine; this upgrades the object from DERIVED-PENDING to a clean DERIVED
result. Failure — a different value, or a value that only agrees after adjusting a free parameter —
means the original \(124/315\) was an artifact of the specific metric selection and must itself be
relabeled as such: a legitimate, useful negative outcome, not a setback. For the scheme object:
either derive the shared one-loop heat-kernel scheme choice from an independent physical requirement
target-blind (promoting it to DERIVED-GIVEN-E), or explicitly name and verify a new
AXIOM-HEATKERNEL-SCHEME-OBJECT stated value-free — i.e. without reference to any numerical target it
is meant to reproduce. The latter path leaves the object formally OPEN but converts it from an
implicit, unexamined convention into an explicit, auditable axiom
(AXIOM-CLOSED-pending-verification).

 (d) Machinery to start from. For \(124/315\) : the Peter–Weyl decomposition
 \(L^2(K_6,E_\mu)=\bigoplus_{(p,q)}V_{(p,q)}\otimes\mathrm{Hom}_{T^2}(V_{(p,q)},E_\mu)\) together with the
exact Casimir formula \(C_2(p,q)=(p^2+q^2+pq+3p+3q)/3\) and dimension formula
 \(\dim(p,q)=(p+1)(q+1)(p+q+2)/2\) , tabulated here for the low representations —
 \((0,0)\to1,\,C_2=0\) ; \((1,0)/(0,1)\to3/\bar3,\,C_2=4/3\) ; \((1,1)\to8,\,C_2=3\) (the adjoint, lowest
nonzero scalar harmonic, 16 modes at \(C_2=3\) ); \((2,0)/(0,2)\to6/\bar6,\,C_2=10/3\) ;
 \((2,1)/(1,2)\to15/\overline{15},\,C_2=16/3\) ; \((3,0)/(0,3)\to10/\overline{10},\,C_2=6\) ;
 \((2,2)\to27,\,C_2=8\) ; \((3,3)\to64,\,C_2=15\) — a direct hand or independently coded summation of the
Seeley–DeWitt coefficients over this spectrum, truncated and Richardson-extrapolated or
zeta-regularized, is the structurally independent route. For the scheme object: work from the
definition of \(\overline{\rm MS}\) subtraction applied to the proper-time heat-kernel integral
 \(K(t)\sim(4\pi t)^{-d/2}\sum_k a_{2k}t^k\) and identify which regularization-independent physical
observable (e.g. a scattering-amplitude threshold, or an anomaly coefficient) the scheme choice is
actually required to reproduce, independent of any number this ledger wants it to produce.

 (e) Leverage. Same family as the SG-7 \(\delta\) object and \(c_{\rm loop}\) : closing the scheme
object once closes it everywhere it is shared — broader leverage per unit effort than most items on
this list, even though it does not touch 5C directly. Closing \(124/315\) independently removes the
single largest circularity caveat currently attached to the \(K_6\) scalar heat-kernel ledger,
strengthening (without promoting) confidence in every scalar-sector calculation built on top of it,
including inputs that feed H4.

 H6 — the \(\mathbb{Z}_2\) orbifold-defect heat-kernel coefficient (BLOCKED)

 (a) The precise open object. A hypothetical order-6 mixed Neumann \(\oplus\) Dirichlet boundary
heat-kernel coefficient for the \(S^1_Y/\mathbb{Z}_2\) orbifold that would combine with the bulk \(a_6\) 
into a "TOTAL = bulk + defect" object. This coefficient, at order 6, does not exist in the published
mathematical literature — the standard boundary heat-kernel tower for manifolds-with-boundary stops
at \(a_5\) , one half-integer order short of the one needed.

 (b) Why it is hard, and the traps. This is not an ordinary computation debt — it is a literature
gap, and possibly a wrong-question gap. Nobody has published the order-6 mixed-boundary-condition
coefficient because the manifold-with-boundary framing may be the wrong object entirely: the
reflection \(\theta\mapsto-\theta\) on \(S^1_Y\) (with isolated fixed points \(\theta=0,\pi\) ) is a global
isometric reflection on a closed manifold — the parent circle — not a genuine boundary in the
manifold-with-boundary sense. The correct framing is equivariant/orbifold (Donnelly-type), where the
two fixed points contribute a smooth equivariant defect rather than a boundary tower term. The trap,
already identified and refused, is fabricating a "TOTAL" bulk-plus-defect number by analogy with
ordinary boundary heat-kernel theory when the object being modeled is not a boundary problem at all —
the twisted trace on \(S^1_R/\mathbb{Z}_2\) is exactly \(1\) , \(t\) -independent, with no boundary tower
whatsoever, which is itself evidence that the "missing \(a_6\) -boundary-coefficient" framing is a
wrong-object artifact rather than a genuine hole.

 (c) What closes it, target-blind, with success and refutation criteria. Two legitimate paths.
 Path 1 — a genuine new mathematical construction. If a mathematically rigorous order-6
mixed-boundary-condition heat-kernel coefficient can be constructed for this specific orbifold setup —
likely a publishable result in its own right, since the general limitation (boundary tower stops at
 \(a_5\) ) is a known gap in the literature, not one specific to this geometry — then TOTAL = bulk + this
defect becomes well-defined. Success criterion: the construction reduces correctly to known results in
appropriate limits (pure Dirichlet, pure Neumann) and satisfies the same consistency checks (Bianchi-
type identities, or agreement with an independent index-theorem computation) used elsewhere in this
ledger. Path 2 — prove the Donnelly equivariant defect already supplies it. Show rigorously that
the smooth equivariant defect already computed here, \(\tfrac12c_3^\gamma=-337361/840\) (the AUD-0059
assembly layer), is the correct and complete object — i.e. that the equivariant/orbifold framing is
not an alternative to the boundary framing but the physically correct one, making the "order-6
boundary coefficient" question moot by dissolution rather than construction. Success criterion: a
proof, via the equivariant index theorem or a direct comparison in a solvable toy case, that the
equivariant defect and any well-defined boundary analogue must coincide, or that the boundary framing
is simply inapplicable to a global reflection on a closed manifold. A refuting-type outcome is a
demonstration that the equivariant defect and a rigorously constructed boundary coefficient (Path 1)
disagree — which would show the two framings are not interchangeable and reopen the question of which
is physically correct.

 (d) Machinery to start from. The Donnelly equivariant heat-kernel trace formalism already in use
here: for an isometry \(g\) with isolated fixed points, the equivariant trace picks up a contribution
 \(\sum_{\rm fixed\ pts}1/|1-dg|\) at each fixed point — here \(2\times\tfrac12=1\) for the two fixed
points of \(\theta\mapsto-\theta\) (each contributing \(1/|1-(-1)|=1/2\) ) — and the orbifold traces
 \(K^\pm=\tfrac12K_{\rm circle}\pm\tfrac12\) (with per-fixed-point \(a_0\) defects \(+1/4\) for even/+
parity and \(-1/4\) for odd/− parity) already tabulated give the leading structure. Extending this to
order 6, rather than the leading \(a_0\) -level defect already recorded, means working out the higher
equivariant heat-kernel expansion directly, using the fixed-point local model (tangent action
 \(dg=-1\) at each of \(\theta=0,\pi\) ) and the general equivariant Seeley–DeWitt expansion — Donnelly's
original papers on equivariant heat kernels for isometries with isolated fixed points, extended here
to mass-dimension-6 order.

 (e) Leverage. Low direct leverage on 5C — this is explicitly a blocked, literature-level item, not
a near-term computation — but resolving the framing question (Path 2) would carry moderate leverage
elsewhere in the ledger: it would settle, once and for all, that no "TOTAL = bulk + boundary" object
should ever be sought for this specific \(\mathbb{Z}_2\) construction, closing off a recurring source of
confusion — and a specific temptation to fabricate a TOTAL — in every future gate touching this
orbifold sector.

 H3 — above-cutoff graviton unitarity (UQF-14): inherited, not independently open

 (a) The precise open object. Whether the graviton remains unitary in scattering processes at and
above the cutoff \(M_*\) (or \(M_U\) on the threshold-closure reading). (b) Why it is not independently
tractable. This question is strictly downstream of H1: unitarity above the cutoff is exactly what a
non-perturbative UV completion would need to guarantee, or what its absence would violate, so there is
no route to answering it that does not pass through H1's construction. (c) What closes it: 
nothing, independently — H3 inherits whatever disposition H1 reaches (constructive close,
closed-negative, or wall record) automatically, without a separate argument. (d)/(e): no
independent machinery or leverage entry applies; H3 is listed here only to make explicit that it is
not a seventh open item requiring its own closure path, and that treating it as independently
tractable would repeat the exact false-openness error flagged for H1.

 What does NOT close this gate, stated once more for the record

 Two specific non-paths are worth naming explicitly because they are the gate's most tempting
mis-closes. First, computing \(\mathrm{tr}[a_6]\) to full satisfaction under H4 — even with \(P\) selected,
the engine bug fixed, the cross-script contradiction resolved, and \(P(a_6)\geq0\) obtained — would still
leave 5C exactly where it is now: OPEN (global wall), because the scope-firewall certificate is a
structural fact about an unbounded operator ladder, not a statement that happens to be true until \(a_6\) 
is known. Second, no amount of exact-rational bookkeeping in Layer B's audit cascade (the certified
chain running from the graviton \(\mathrm{Sym}^2(T_6)\) Levi-Civita ratio \(-6373/630\) through the
13D bulk graded keystone \(-953329/1260\) to the orbifold assembly \(-491353/630\) ) changes this either —
that cascade's ceiling is explicitly AUDIT-CLOSED, never physics-CLOSED, precisely because it computes
further heat-kernel data rather than constructing the resummed interacting theory. The only entry in
this list that can move the 5C grade is H1, and H1 moves it only through a genuine non-perturbative
construction, a rigorous non-existence proof, or an honest, explicitly shared wall record.

 The honest bottom line on leverage

 Ordered by what actually moves the roll-up: H1 alone can change the 5C/UQF-9/UQF-14 disposition,
and only through a genuine non-perturbative construction, a rigorous non-existence proof, or an honest
wall record — never through a per-gate shortcut. H2/P0 is the cleanest bounded diagnostic on
whether any finite-grain shortcut into H1 exists at all; current evidence, via the T-DEEP screen,
favors "no" ( \(R_0\) is a pure color/gauge object with zero gravitational input, and the graviton sector
has produced no invariant yet shown independent of it). H4 and H5 sharpen the linearized ,
conditional-on-UQF-9 certificate for legs 5A/5B — real, valuable, decision-grade work — but by the
scope-firewall established elsewhere in this dossier, no amount of H4/H5 progress can, even in the
best case, cross into 5C itself. H6 is presently blocked at the literature level and carries only
moderate framing-clarification leverage. H3 carries no independent leverage; it is a pure
inheritance of H1. The dossier's obligation is to keep this hierarchy explicit: work that feels
productive (H4, H5) must never be mistaken, by the dossier or by a future reader, for work that closes
the gate (H1) — that confusion is precisely the gate's signature mis-close, named and refused
throughout this section.

 Honest ceiling, scope & the endpoint

 Fixed grade for this gate (stated once, never mutated below): REDUCED-TO-AXIOM / ANCHORED +1. 
Everything in this section is written to make the ceiling of that grade explicit — not to hedge it
downward, and not to inflate it upward. The task here is bookkeeping of the sharpest kind: separate,
with zero ambiguity, what has actually been shown from what merely sounds adjacent to it, name every
anchor that was spent to reach the grade, and then state the endpoint in the fixed closing form. A
reader who reads only this section should come away able to reconstruct exactly what UQF-5C does and
does not license, without needing anything else in the dossier.

 II.1 What is explicitly NOT claimed

 Five distinct bright-lines are drawn here, each guarding against a specific, previously-observed
failure mode. None of the five is a matter of taste; each corresponds to a concrete way this gate has
been, or could be, mis-read.

 (1) Dissolved ≠ solved. The Granularity axiom (AXIOM-COSTFLOOR, §I.3 above) is a dissolution : it
removes exactly one divergence class — the \(a\to0\) runaway tail of the unbounded heat-kernel tower
 \(\{a_8,a_{10},\dots\}\) — by supplying an irreducible cost/action floor \(\Delta_0>0\) that regularizes
the small-parameter limit. Dissolving that one pathology is real, physical, and grounded in established
bounds (Margolus–Levitin \(\tau\geq\pi\hbar/2E\) ; Landauer \(\Delta E\geq k_BT\ln2\) ; Bekenstein
 \(S\leq2\pi k_BRE/\hbar c\) ). But a dissolved divergence class is not the same object as a solved 
strong-coupling theory. Ten walls remain completely untouched by this axiom, and — the sharpest point
of all — the finite \(a_6\) coefficient itself is one of the untouched ten. \(a_6\) was never divergent;
it is a finite, well-defined heat-kernel coefficient whose value is a computation debt (the graviton
Route A leg is OWED at the Gelfand–Tsetlin off-diagonal hopping-term stratum on \(\mathrm{Sym}^2_0\) ),
not a runaway that Granularity could ever regularize. Adopting AXIOM-COSTFLOOR does not move the
graviton \(a_6\) computation one inch closer to being finished, and it does not exhibit, or even
gesture at, a non-Gaussian fixed point. No fixed point is shown. No value of \(a_6\) is supplied by the
axiom. No constructive UV completion is produced. The correct one-line summary, carried verbatim from
the guardrails: dissolving one divergence class is not solving the theory. 

 (2) Selection ≠ derivation. Every geometric object this gate leans on is selected , not forced .
The frozen background \(\mathfrak{B}_{\rm active}=\mathcal{M}_4\times K_6\times S^2\times S^1_Y/\mathbb{Z}_2\) 
with \(K_6=SU(3)/T^2\) survives a battery of admissibility constraints (Weyl-rigid chamber
 \(\vec u\in[1/2,3/2]^3\) , threshold closure at \(M_U\) , the finestness of the \(\mathbb{Z}_6\) quotient) —
but surviving those constraints is not the same claim as being the unique shape compatible with a
consistent quantum theory of gravity. No uniqueness theorem is asserted or available. Likewise the
de-Donder gauge-fix and the \(\overline{\mathrm{MS}}\) /heat-kernel scheme are the standard, physically
sensible rulebook choices for this operator, and the BRST ghost sector they force is a mathematical
consequence of that choice — but the choice of scheme itself is a selection among rulebook
conventions that give equivalent physics, not a derivation that this is the only rulebook a UV
completion could use. Where the geometry forces something (the sign and multiplicity of the ghost
subtraction, the ghost-corrected weight \(91-2\cdot13=65\) , the ratio \(|\mathrm{Riem}|^2/\mathrm{Scal}^2
=23/75\) ), that forcing is flagged explicitly as DERIVED-GIVEN-E or DERIVED; where the geometry is
merely a surviving candidate among admissible options, it is flagged SELECTED. The word "ANCHORED" in
this gate's grade must never be read as "DERIVED": anchoring the row on \(\{\Delta_0>0,\ \text{Lorentz-
scalar proper-time floor}\}\) fixes one axiom pair as sufficient for the stated partial result; it does
not derive that axiom pair from something more primitive, and it does not derive the completion.

 (3) Given- \(E\) ≠ derivation of \(E\) . The Lichnerowicz endomorphism \(E\) that enters
 \(L_{\rm grav}=-(\nabla^2+E)\) — and, more generally, the observed Standard Model matter content that
fixes which bundle endomorphism sits inside every Seeley–DeWitt coefficient — is a given input to
this gate, not something the gate derives. The brief is explicit that given- \(E\) "is the single largest
charged input" to the entire heat-kernel ladder: it is given, not derived. Every DERIVED-GIVEN-E result
in this dossier (the operator \(L_{\rm grav}\) itself; the fiber weight \(65=91-2\cdot13\) ; the Einstein
constant \(\kappa=5/12\) ) is DERIVED conditional on \(E\) being what it is observed to be — swap the
matter content and the operator changes accordingly. The gate never claims to derive the matter
spectrum from the graviton sector, and it never claims that \(E\) itself follows from some deeper
principle internal to UQF-5C. This is a standing, explicitly-flagged input, not a hidden assumption.

 (4) A refuted number stays refuted; a certificate is not a hedge. The bulk magnitude
 \(-2.818\times10^{94}\,\mathrm{GeV}^6\) that once circulated as a candidate value for the graded \(a_6\) 
functional is REFUTED at decision grade , not merely "superseded" or "uncertain." It was
R2-contaminated, scheme-anchored, and — the decisive, structural reason — ill-posed at odd \(D=13\) :
the relevant local heat-kernel slot sits at the half-integer zeta pole \(s=7/2\) , which vanishes
identically in dimensional regularization and carries no log or anomaly term to hold a finite GeV \(^6\) 
answer. There is, quite simply, no such number to compute at that dimension and that order; the
correctly-posed object is the finite trace \(\mathrm{tr}[a_6(L_{\rm grav})]\) , not a dimensionful bulk
magnitude. This refutation is a reached verdict — a piece of completed, decision-grade science — and
it is listed here as a non-claim in the specific sense that the dossier must never present it as still
live or reopen it as a pending number. Symmetrically, the scope-firewall statement ("one heat-kernel
coefficient can never be a non-perturbative UV completion") is a proved structural certificate , not
a hedge invented to excuse an unfinished calculation: the unbounded ladder \(a_6<a_8<a_{10}<\cdots\) is a
mathematical fact about the Seeley–DeWitt expansion, and no finite truncation of an unbounded ladder
can, by construction, decide convergence of the full interacting series. Confusing a certificate for a
hedge — or a hedge for a certificate — is exactly the error this non-claim forecloses.

 (5) Computing \(a_6\) , even exactly, would not close this gate. This is the gate's signature
mis-close and is refused explicitly and by name: " \(a_6\) is computed, therefore 5C is closed" is false
regardless of which value of \(a_6\) eventually gets computed, and regardless of how many independent
routes confirm it. \(a_6\) is input to the UV-completion question; it is never identical with the
completion. Even in the best realistic case — the Gelfand–Tsetlin hopping-term wall (H4) is cleared,
the graviton Route A and Route B values are reconciled to within the pre-registered \(10^{-6}\) 
tolerance, the positivity functional \(P\) is selected from among its three currently inequivalent
readings, and \(P(a_6)\geq0\) comes back — the resulting certificate would lift legs 5A/5B to
certificate grade conditional on UQF-9 , and would still never reach 5C. The wall this gate names is
categorically different in kind from a finite coefficient, however exactly known: it is the existence
(or provable non-existence) of a non-Gaussian, truncation-independent fixed point for the fully
interacting graviton above the cutoff. No amount of heat-kernel bookkeeping, on its own, supplies or
refutes that fixed point. This is why the dossier's own maximum-leverage assessment is stated
plainly: closing every bounded work package available to this gate (H4, H5, H2/P0) would sharpen the
 linearized certificate to an unconditional decision-grade result, and even then the sharpened result
would remain conditional on UQF-9, with the 5C wall standing exactly where it stands today. The only
thing that closes 5C is closing UQF-9.

 II.2 The anchors paid

 The grade REDUCED-TO-AXIOM / ANCHORED +1 was purchased with a specific, enumerable set of anchors.
Listing them plainly is the honest accounting this section exists to do; no anchor here is hidden
inside a derivation, and no anchor is claimed as more than it is.

 The conditional axiom pair \(\{\Delta_0>0,\ \text{Lorentz-scalar proper-time floor}\}\) — the "+1." 
 This is the one new , gate-specific axiom charged to UQF-5C: an irreducible quantum of cost/action
 (not of length), proved to act on a Lorentz-scalar quantity (proper time) via theorem T-LI, so that
 no preferred frame is introduced. It is grounded in, but not identical to, three pieces of
 independently established physics — Margolus–Levitin, Landauer, and Bekenstein — which are
 themselves MEASURED-ANCHOR / established results, not new posits. The axiom is explicitly
 AXIOM-OPEN / not atomic : it is a named, relocatable floor, not a claimed fundamental fact about
 nature, and its only certified consequence is the dissolution of the single divergence class
 described in §II.1(1). This is the entirety of the "+1" — one named axiom, isolated and auditable,
 rather than smuggled into a derivation.

 The observed matter content, given- \(E\) . As detailed in §II.1(3), the bundle endomorphism \(E\) 
 entering \(L_{\rm grav}=-(\nabla^2+E)\) is a given input, charged against every DERIVED-GIVEN-E result
 in this gate (the operator itself, the fiber weight \(65\) , the Einstein constant \(\kappa=5/12\) ). It is
 not counted as part of the "+1" because it is not a new axiom specific to this gate — it is the same
 observed Standard Model content every gate in the ledger takes as given — but it is named here so the
 full cost of the DERIVED-GIVEN-E results is visible.

 The frozen background selection. The specific shape \(\mathcal{M}_4\times K_6\times S^2\times
 S^1_Y/\mathbb{Z}_2\) , \(K_6=SU(3)/T^2\) , survives the admissibility screens but is SELECTED rather than
 proven uniquely forced. This is not charged as a new axiom for this gate specifically (it is the
 shared background anchor for the entire ledger), but it is named here because every curvature number
 quoted in this section — \(|\mathrm{Riem}|^2/\mathrm{Scal}^2=23/75\) , \(\kappa=5/12\) , the ghost-corrected
 weight \(65\) — is a property of this selected shape, not a shape-independent theorem.

 What is explicitly NOT among the anchors paid. The gate does not consume, and its grade does not
 rest on, any of the four calibration anchors \(\{M_{\rm Pl},\alpha_i(M_Z),y_t,|V_{us}|\}\) or the
 neutrino-count anchor \(N_\nu\) used elsewhere in the ledger; it also does not rest on the derived
 UV-floor numbers, both of which are read off the geometry as consequences of \(M_{\rm Pl}\) and are
 explicitly flagged as "not a closure and not a new anchor" in every place they appear. Two consistent
 readings of that floor exist in the corpus and are reported here without silently picking one:
 the full-precision Planck-normalization computation gives \(M_*^{11}=M_{\rm Pl}^2/\mathrm{Vol}(X_{\rm
 active})=4.023836152402511\times10^{185}\,\mathrm{GeV}^{11}\) , i.e.
 \(M_*=7.467050992135091\times10^{16}\) GeV; a separate threshold-closure reading elsewhere in the ledger
 quotes \(M_*\approx6.01\times10^{16}\) GeV for the same conceptual floor. Both are DERIVED geometry
 read-offs of where the perturbative series in \(G_NE^2\) stops converging, not new anchors and not a
 closure; the unification scale itself, \(M_U\sim1.0\times10^{16}\) GeV (two-loop closure residual
 \(9.6\times10^{-11}\) ), is the more tightly pinned companion number and is not in tension between
 sources. There is no \(N\) -sigma pull for this gate: every quantity charged above is an exact geometric
 rational or an established physical bound, not a fit to a measured central value with an error bar.
 Stating that plainly forecloses any temptation to dress this gate's ceiling up with statistical
 language it does not use.

 II.3 Locating the ceiling precisely: what is banked versus what stands open

 Before the closing statement, it is worth being maximally explicit about where the line sits, because
this is the single most commonly blurred boundary in the whole gate.

 Banked, terminal, and not reopened by anything in this section: the operator-supply result (5A,
DERIVED-GIVEN-E); the BRST-forced ghost-corrected fiber weight \(91-2\cdot13=65=\dim
\mathrm{Sym}^2_0(SO(11))\) ; the curvature ratio \(|\mathrm{Riem}|^2/\mathrm{Scal}^2=23/75\) proved by
three independent target-blind routes and cross-checked against the first-Bianchi machine-zero test
( \(\approx2.5\times10^{-16}\) ); the sphere heat-kernel cross-checks \(a_6(S^2)=4/315\) , \(a_6(S^4)=74/63\) ,
 \(a_6(S^6)=1139/63\) , \(a_6^{\rm conf}(S^6)=5/63\) , agreeing between independent routes to
 \(\sim4\times10^{-14}\) ; the scope-firewall certificate that one coefficient can never be a UV
completion; the decision-grade refutation of \(-2.818\times10^{94}\,\mathrm{GeV}^6\) ; and the T-DEEP
structural result that the compactification radius \(R_0\) is a pure color/gauge object
( \(R_0/\ell_{\rm Planck}\sim194\) ) with zero gravitational input, so that no finite-grain shortcut to 5C
exists without relocating onto UQF-9's P0 sub-target.

 Standing open, and not touched by the axiom or by any result above: the graded graviton+ghost
 \(a_6\) value itself, where two independent routes disagree by \(31/48\approx0.646\) against a
pre-registered \(10^{-6}\) tolerance; the trace \(\mathrm{tr}[a_6(L_{\rm grav})]\) , wholly uncomputed
(the coefficient \(C\sim-6.39\) is not reproduced and is not asserted here); the positivity functional
 \(P\) , which is currently UNSELECTED among three inequivalent readings, so that " \(P(a_6)\geq0\) " is not
yet even a well-posed predicate to evaluate; the order-6 \(\mathbb{Z}_2\) orbifold-defect boundary
coefficient, which is BLOCKED because the published literature's boundary heat-kernel tower stops at
 \(a_5\) and no order-6 mixed Neumann \(\oplus\) Dirichlet term has been constructed; and — standing above
all of these as the object that actually decides the gate — the non-perturbative, truncation-
independent fixed point (or rigorous non-existence proof) for the fully interacting graviton, which is
UQF-9 and is not attempted here.

 The smallest remaining object, named plainly. If a single next object is asked for, it is not a
number this gate can produce internally: it is the shared UQF-9 construction — a target-blind
functional-renormalization-group / asymptotic-safety-type non-Gaussian fixed point for the interacting
graviton, stable under truncation, built once and inherited by every gate (5C, UQF-14) that currently
carries this wall as OPEN. Three, and only three, outcomes legitimately close it: a target-blind
construction exhibiting a truncation-independent fixed point (constructive close); a rigorous proof
that no such fixed point exists on this geometry (CLOSED-NEGATIVE, a fully valid terminus); or, failing
both, an explicit WALL RECORD naming the attempted construction, its blocker, and its shared status
across UQF-9/UQF-14/5C. Nothing smaller than that construction closes this gate; in particular, no
further heat-kernel bookkeeping on \(a_6\) , however exact, is that object (§II.1(5)).

 II.4 The closing endpoint statement

 Nothing left to extract from this gate's internal resources beyond what is stated above. The row is
anchored, not derived, and the anchoring is stated here in full:

 Nothing left. Anchored on: Shape: the frozen \(\mathcal{M}_4\times K_6\times S^2\times S^1_Y/\mathbb{Z}_2\) 
arena with \(K_6=SU(3)/T^2\) , all three layers (× Stage / ⊕ Rulebook / ⊗ Actors) pinned, SELECTED by
admissibility constraints and supplying the de-Donder-gauge-fixed, BRST-ghost-corrected graviton
operator \(L_{\rm grav}=-(\nabla^2+E)\) with fiber weight \(91-2\cdot13=65\) and curvature input
 \(|\mathrm{Riem}|^2/\mathrm{Scal}^2=23/75\) ; Granularity: the named axiom pair
 \(\{\Delta_0>0\ (\text{irreducible cost/action floor}),\ \text{Lorentz-scalar proper-time floor}\}\) 
(AXIOM-COSTFLOOR + theorem T-LI), grounded in Margolus–Levitin/Landauer/Bekenstein, the single "+1"
that dissolves exactly one UV-divergence class; Scale: the UV floor read off the geometry,
 \(M_*=7.467050992135091\times10^{16}\) GeV by the full-precision Planck-normalization computation
(a separate threshold-closure reading elsewhere in the ledger quotes \(M_*\approx6.01\times10^{16}\) GeV
for the same conceptual floor — both DERIVED read-offs, neither a closure), companioned by the more
tightly pinned unification scale \(M_U\sim1.0\times10^{16}\) GeV and internal radius
 \(R_0=1.591549430918954\times10^{-17}\,\mathrm{GeV}^{-1}\) , which together name precisely where the
perturbative expansion in \(G_NE^2\) breaks down without themselves supplying what replaces it;
Observables: none of the four calibration anchors \(\{M_{\rm Pl},\alpha_i,y_t,|V_{us}|\}\) or \(N_\nu\) are
consumed here — the gate rests entirely on exact geometric curvature rationals and established physical
bounds, with no statistical pull to quote; Dissolution: the Granularity axiom dissolves exactly one
divergence class — the \(a\to0\) runaway of the higher heat-kernel coefficient tower
 \(\{a_8,a_{10},\dots\}\) — while leaving the finite \(a_6\) obligation and the non-perturbative fixed-point
question completely untouched, so that dissolution here is real but strictly partial, and the
constructive strong-coupling UV completion of the interacting graviton stands as an open global wall —
carried under UQF-9, inherited by UQF-14, shared by every research program in quantum gravity including
this one — that only a target-blind non-perturbative construction, or a rigorous proof that none
exists, can ever close.

 Closure ledger — UQF-5C — UV completion (shared)

 Status (fixed): REDUCED-TO-AXIOM · ANCHORED +1

 The technical closure LEDGER (separate document)

 Gate: UQF-5C — UV completion (shared) / interacting–strong-coupling graviton. 
 Fixed grade (do not change): REDUCED-TO-AXIOM / ANCHORED +1. This ledger is the auditor's
record: identity, anchor register, root stack, numbered derivation chain with exact values, a
credit-ladder grade on every leg, the anti-claims, and the endpoint line. The per-gate closure-
of-record additionally carries an OPEN (global wall) roll-up disposition for the shared
constructive UV-completion object (§H1/UQF-9); both statements are recorded here without either
one silently absorbing the other.

 L0. Layer-0 wall identity

 Field 
 Value 

 Gate 
 UQF-5C 

 Object under adjudication 
 The interacting , strong-coupling, non-perturbative UV completion of the graviton on the frozen 13D geometry 

 Wall class 
 Global, shared — identical open problem confronts every quantization program (strings, asymptotic safety, loop quantum gravity, causal sets, this geometry). Not a private defect. 

 Regime where the wall lives 
 Above the cutoff, where the perturbative expansion in \(G_N E^2\) stops converging (loop counterterms from \(R^3,R^4,\dots\) curvature-invariant operators proliferate without bound) 

 Canonical frontier count 
 Exactly two structural frontiers exist on the board (Gap-13, UQF-4). UQF-5C is not a third frontier — it is the graviton leg of the UQF-9 shared wall, inherited downstream by UQF-14 (above-cutoff unitarity). 

 Roll-up carrier 
 H1 = UQF-9 (constructive UV completion / strong-coupling closure). Only closing UQF-9 closes the 5C roll-up. 

 Closure-of-record disposition (conservative ledger reading) 
 OPEN (global wall) for the roll-up; the graded graviton+ghost \(a_6\) value leg is OPEN/FAIL_VALUE_MISMATCH (§D8 below) 

 Public-board pill (this dossier's fixed grade) 
 ANCHORED +1 / REDUCED-TO-AXIOM , read as TERMINAL + RESIDUALS-SHOWN on the conditional axiom pair {Δ₀>0, Lorentz-scalar proper-time floor} 

 Reconciliation 
 Both descriptions are the same physics: a genuine banked structural result (operator supply + BRST-forced fiber weight + a granularity axiom that dissolves one divergence class) sitting next to an unclosed strong-coupling wall. The dossier states both; neither is hidden under the other. 

 L1. Layer-1 endpoint anchor

 The published row is ANCHORED on the conditional axiom pair

 \[
\{\ \Delta_0 > 0\ \text{(a positive cost/action floor)},\quad \text{Lorentz-scalar proper-time floor}\ \}.
\]

 This is the one named, irreducible axiom (AXIOM-COSTFLOOR) that supplies the "+1" in
REDUCED-TO-AXIOM / ANCHORED +1. It is read as TERMINAL + RESIDUALS-SHOWN : the axiom pair
terminates this leg of the ledger honestly, while the residual (the strong-coupling wall itself)
is shown, not absorbed. Binding distinctions carried into every downstream row:

 SELECTED ≠ FORCED. The frozen background \(K_6\times S^2\times S^1_Y/\mathbb{Z}_2\) is
 selected by constraints, not proven the unique geometry that could carry this axiom.

 ANCHORED ≠ DERIVED. Anchoring on {Δ₀>0, Lorentz-scalar floor} is not a derivation of the
 completion; it is a floor axiom under which the residual divergence structure is characterized.

 AXIOM-CLOSED ≠ atomic. The cost-floor axiom is a named, relocatable floor (a specific
 quantum of action/cost, not a smallest length), not a proof that no finer floor could exist.

 L2. Layer-2 root stack

 Tier A — Shape / Scale / Granularity (full precision, all three ×⊕⊗ levels)

 A1. Shape root. 

 × Stage (metric geometry): \(\mathcal{M}_4\times K_6\times S^2\times S^1_Y/\mathbb{Z}_2\) ,
 \(D=4+6+2+1=13\) , \(K_6=SU(3)/T^2\) the full \(A_2\) flag manifold (real dimension 6, primitive
 factor). Frozen branch identity is audit-tracked (hash-level provenance exists but is not 
 quoted here per the inline-only rule; it certifies which object was tested, not the physics).

 ⊕ Rulebook: de-Donder (harmonic) gauge-fix on the metric fluctuation \(h_{MN}\) ;
 Faddeev–Popov ghost sector; \(\overline{\rm MS}\) /heat-kernel renormalization scheme; \(\mathbb{Z}_2\) 
 orbifold parity on \(S^1_Y/\mathbb{Z}_2\) ( \(\theta\mapsto-\theta\) , isolated fixed points
 \(\theta=0,\pi\) ); cubic-curvature (mass-dimension-6) Gilkey operator basis (~46 terms) as the
 readout basis for \(a_6\) .

 ⊗ Actors: \(L_{\rm grav}=-(\nabla^2+E)\) acting on the graviton \(\mathrm{Sym}^2(T)\) bundle
 plus the FP ghost bundle; the graded/ghost-corrected fiber supertrace; the holonomy
 decomposition of these bundles under \(K_6\times S^2\times S^1_Y\) .

 Status: Frozen background is SELECTED by constraints, not proven uniquely forced. This
 qualifier is carried at full strength into every geometric input below.

 A2. Scale root. 

 The strong-coupling wall is the regime at and above the UV cutoff where the perturbative
 series in \(G_NE^2\) ceases to converge.

 Geometry read-off of the UV floor (a derived scale, not a closure, not a new measured
 anchor ):
$$
M_* \approx 6.01\times10^{16}\ \text{GeV (order-of-magnitude reading used at this gate)},\qquad
M_U \sim 1.0\times10^{16}\ \text{GeV},\qquad
R_0 = 1.591549430918954\times10^{-17}\ \text{GeV}^{-1}.
$$
 (Cross-reference to the geometry pack's Planck-normalized value:
 \(M_*=7.467050992135091\times10^{16}\) GeV from \(M_*^{11}=M_{\rm Pl}^2/\mathrm{Vol}(X_{\rm active})\) ,
 \(\mathrm{Vol}(X_{\rm active})=3.704417261398702\times10^{-148}\ \mathrm{GeV}^{-9}\) — the two
 figures are the same geometric read-off quoted at different rounding/normalization stages in
 the corpus; both are DERIVED, neither is a closure.)

 This root is where the gate's openness structurally lives : no amount of refining \(M_*\) or
 \(R_0\) turns the scale root into a completion. Scale supplies where the wall sits, not a lever
 across it.

 A3. Granularity root (the "+1" engine). 

 AXIOM-COSTFLOOR: an irreducible quantum of cost/action , explicitly not a smallest
 length, applied Lorentz-invariantly.

 What it dissolves: exactly one UV-divergence class — the \(a\to0\) runaway tower
 \(\{a_8,a_{10},\dots\}\) (the unbounded higher heat-kernel coefficients that a naive continuum
 limit would let run to arbitrarily short distance).

 What it leaves untouched: 10 walls remain , including the finite coefficient \(a_6\) 
 itself. DISSOLVED ≠ SOLVED.

 Grading: AXIOM-OPEN / not atomic — a named floor that can be relocated (to a different
 value, or in principle to a different physical mechanism) but not eliminated by this argument.

 Lorentz-scalar property (theorem T-LI): the floored quantity is proved to be a Lorentz
 scalar, so no preferred frame is introduced. This supplies the {Lorentz-scalar proper-time
 floor} half of the anchor pair.

 Physical grounding of the cost floor (MEASURED-ANCHOR / established physics, not a new
 invention of this framework): 
$$
\text{Margolus–Levitin: } \tau \geq \frac{\pi\hbar}{2E},\qquad
\text{Landauer: } \Delta E \geq k_BT\ln 2,\qquad
\text{Bekenstein: } S \leq \frac{2\pi k_B R E}{\hbar c}.
$$
 These three bounds are independently established results in physics (quantum speed limits,
 thermodynamics of computation, holographic entropy bound); AXIOM-COSTFLOOR packages them as a
 Lorentz-invariant cost/action floor rather than deriving them from the 13D geometry.

 Tier A summary table. 

 Root 
 Level pinned 
 Terminal status 
 Grade 

 Shape 
 ×⊕⊗ full (K₆×S²×S¹_Y/ℤ₂, de-Donder+FP, \(L_{\rm grav}\) ) 
 SELECTED, supplies operator 
 DERIVED-GIVEN-E (structure only) 

 Scale 
 UV floor \(M_*, M_U, R_0\) 
 Locates the wall; no lever across it 
 DERIVED (geometry read-off), not a closure 

 Granularity 
 AXIOM-COSTFLOOR + T-LI 
 Dissolves ONE class of 10+; rest OPEN 
 REDUCED-TO-AXIOM (the +1) 

 Tier B — screens

 Physical Equivalence / Invariance screen: gauge redundancy under diffeomorphisms/graviton
 gauge transformations forces the de-Donder gauge-fix and the Faddeev–Popov ghost
 subtraction; the physical coefficient must be the frame-independent, ghost-corrected object
 (graviton \(-2\cdot\) ghost, §D2 below). This screen is satisfied: the object this ledger tracks is
 already the gauge-invariant one.

 Nonseparability screen: a single heat-kernel coefficient ( \(a_6\) ) cannot, by construction,
 compose into a statement about strong-coupling (non-perturbative, all-orders) consistency. This
 screen is the formal reason the scope-firewall certificate (§D11 below) is airtight — it is not
 an incidental caution but a structural fact about what one finite coefficient can and cannot
 carry.

 Record Interface screen: every curvature ledger entry, every heat-kernel run, and the
 refuted magnitude (§D12) are independently reproducible/reviewable computations, satisfying the
 interface screen (no black-box numbers).

 Layer-2 T-DEEP screen (banked): \(R_0/\ell_{\rm Planck}\sim 194\) . \(R_0\) is certified to be a
 pure color/gauge object with zero gravitational input (it is set by the KK/unification
 scale \(M_U\) via \(R_0=(2\pi M_U)^{-1}\) , not by any Planck-scale/gravitational construction).
 Consequence (DERIVED, banked): every finite-grain dissolution attempted at 5C relocates onto
 UQF-9 through the isolated sub-target P0 — "does gravity own an intrinsic shortest length,
 or does it merely inherit the color scale \(R_0\) ?" 5C therefore has no finite-grain shortcut. 
 This screen result must not be presented as dissolving the wall — it locates where any future
 dissolution would have to occur (P0, carried at UQF-9), and rules out a large class of would-be
 quick fixes at 5C itself.

 D. The numbered derivation-chain ledger (every step, exact value, credit-ladder grade)

 Each row: the object, its full-precision value with layer pin, its derivation route, and its
credit-ladder grade. Grades used: DERIVED-GIVEN-E, DERIVED (target-blind), DISSOLVED-GIVEN-root,
MEASURED-ANCHOR, CERTIFIED-IRREDUCIBLE, REDUCED-TO-AXIOM, CLOSED-NEGATIVE, OPEN, REFUTED.

 D1. Operator supply (leg 5A). 
De-Donder (harmonic) gauge-fixed metric fluctuation \(h_{MN}\) on the frozen 13D background yields
$$
L_{\rm grav} = -(\nabla^2 + E),
$$
a Lichnerowicz-type Laplace operator, whose 4D massless spin-2 mode is the graviton, with a
Kaluza–Klein tower above it indexed by \(K_6\times S^2\times S^1_Y\) harmonics.
 Layers: × Stage = \(\mathrm{Sym}^2(T)\) bundle over the frozen 13D base; ⊕ Rulebook = de-Donder
gauge + \(\overline{\rm MS}\) heat-kernel scheme; ⊗ Actors = connection \(\nabla=\) Levi-Civita,
endomorphism \(E\) = the Lichnerowicz Weitzenböck term
 \((E_Lh)_{ab}=\mathrm{Ric}_{ac}h^c{}_b+\mathrm{Ric}_{bc}h^c{}_a-2R_{acbd}h^{cd}\) .
 Grade: DERIVED-GIVEN-E (terminal as structure). Allowed claim: the geometry supplies the
operator. Forbidden claim: that this "certifies the quantum theory" or "derives GR / derives E."

 D2. Ghost-corrected fiber weight (BRST-forced, not chosen). 
$$
\text{graviton bulk fiber dim (D=13)} = \dim\mathrm{Sym}^2(\mathbb{R}^{13}) = \frac{13\cdot14}{2}=91,
$$
$$
\text{ghost fiber dim} = 13,\qquad \text{ghosts enter with multiplicity } -2,
$$
$$
\boxed{\text{graviton}(91) - 2\cdot\text{ghost}(13) = 65.}
$$
Cross-check: the physical massless graviton degrees of freedom at \(D=13\) are
 \(\dim\mathrm{Sym}^2_0(SO(11)) = \frac{11\cdot12}{2}-1 = 66-1=65\) — the little-group count
independently reproduces \(65\) . Also \(65=91-26\) .
 Do NOT use "67/11." These are the graded constants of a different object,
 \(\gamma=\mathrm{Sym}^2(\mathrm{diag}(\mathbf{1}_{12},-1))\) , giving
 \(\mathrm{tr}\,\gamma_{\rm grav}=67\) , \(\mathrm{tr}\,\gamma_{\rm ghost}=11\) , Block-A graded weight
 \(67-2\cdot11=45\) — used only in the \(\mathbb{Z}_2\) -defect grading (§D9), never the bulk
 \(91/13/65\) triple. These two objects (bulk supertrace vs. \(\mathbb{Z}_2\) -graded block trace) are
kept structurally distinct throughout this ledger.
 Grade: DERIVED-GIVEN-E , sign and multiplicity forced by BRST nilpotency (not a free
modeling choice).

 D3. Curvature input win — the proved ratio. 
At the frozen Weyl-chamber center \(\vec u=(1,1,1)\) , in the Killing-form normal metric
 \(g=(-B)|_{\mathfrak m}\) (dimensionless normalization):
$$
\mathrm{Ric}_i=\frac{5}{12},\quad \mathrm{Scal}=\frac{5}{2},\quad
|\mathrm{Ric}|^2=\frac{25}{24},\quad |\mathrm{Riem}|^2=\frac{23}{12},
$$
$$
\boxed{\ |\mathrm{Riem}|^2/\mathrm{Scal}^2 = 23/75 = 0.3066666666666667\ },\qquad
|\mathrm{Ric}|^2/\mathrm{Scal}^2=\frac16,\qquad \mathrm{Scal}/\mathrm{Ric}_i=6=\dim K_6.
$$
Equivalently in the R₆-physical normalization, \(\mathrm{Ric}_i=1/(2R_6^2)\) ,
 \(\mathrm{Scal}=3/R_6^2\) ; the ratios above are metric-scale-invariant and agree in both.
 Proved by three independent target-blind routes: (1) \(SU(3)\) structure constants +
naturally-reductive curvature formula; (2) curvature-free spectral heat-trace; (3) full
Levi-Civita Riemann tensor over the 3-parameter squashing metric \((x_1,x_2,x_3)\) . Load-bearing
trip-unit cross-check: \(\mathrm{Scal}=15/2\) (one convention) or \(15\) (another),
 \(\|\mathrm{Ric}\|^2=75/8\) , \(\|\mathrm{Riem}\|^2=69/4\) , \(\mathrm{Weyl}^2=27/2\) , Einstein
 \(R/6=5/4\) ; \(\mathrm{Weyl}^2\) is FORCED by the \(d=6\) curvature decomposition (Riem = Weyl + Ricci
part in 6 dimensions).
 No-target-loading self-check: \(23/75=0.30667\) matches neither prior (wrong) target — not
the buggy engine value \(0.2109\) , nor the firewall value \(0.0667\) — so it cannot have been
back-solved to a desired answer.
 Grade: DERIVED (target-blind, 3 independent routes). Necessary, not sufficient — does NOT
by itself close the gate. 

 D4. Einstein constant. 
$$
\kappa = \frac{5}{12}\quad(\text{the Ricci eigenvalue on the corrected } K_6;\ \text{buggy } 7/12\ \text{retired}).
$$
 Grade: DERIVED-GIVEN-E. 

 D5. First-Bianchi target-blind falsification test (theorem-criterion). 
Corrected curvature tensor: first-Bianchi residual \(\approx 2.5\times10^{-16}\) (machine zero,
i.e. numerically exact). The prior buggy build gave residual exactly \(1/7\) in the engine's
 \(\mathrm{Ric}=1/2\) normalization ( \(1/6\) in raw \(-B\) units) — a finite, wrong, nonzero number. This
test is what caught the \(\sim31\%\) curvature bug ( \(31/147\to23/75\) ; \(\kappa=7/12\to5/12\) ).
 Grade: DERIVED (theorem-criterion, machine zero). Passing Bianchi validates the curvature
tensor's internal consistency; it does not validate \(a_6\) itself.

 D6. Derivative-sector curvature (K₆ homogeneous, not locally symmetric). 
Because \(\nabla\mathrm{Riem}\neq0\) on \(K_6\) , only \((\nabla\mathrm{Riem})^2\) survives as an
independent weight-6 invariant:
$$
|\nabla\mathrm{Riem}|^2 = 54\ (\text{trip-unit; three independent routes agree}) \;=\; \frac14\ (\text{Killing-form normalization}),
$$
$$
|\nabla\mathrm{Ric}|^2=0,\qquad |\nabla\mathrm{Scal}|^2=0,\qquad |\nabla^{LC}E_{\rm grav}|^2=162\ (\text{trip-unit}),
$$
box cross-check: \(R_{abcd}\Box R^{abcd} = -54 = -\|\nabla\mathrm{Riem}\|^2\) .
Weight-6 cubic invariants at the Killing-normalized Einstein center (exact rationals):
$$
K_1=R_{ab}{}^{cd}R_{cd}{}^{ef}R_{ef}{}^{ab}=-\frac{113}{72},\qquad
K_2=R_{abcd}R_{aecf}R_{ebfd}=-\frac{5}{72},
$$
$$
\mathrm{Scal}^3=\frac{125}{8},\quad \mathrm{Scal}\cdot|\mathrm{Ric}|^2=\frac{125}{48},\quad
\mathrm{Scal}\cdot|\mathrm{Riem}|^2=\frac{115}{24},\quad \mathrm{Ric}^3=\frac{125}{288},\quad
\mathrm{Ric}\cdot|\mathrm{Riem}|^2=\frac{115}{144}.
$$
This sector is carried end-to-end in the banked \(a_6\) chain (audited CHAIN-SAFE) — it is not a
symmetric-space-blind omission; \(K_6\) 's failure to be locally symmetric is tracked explicitly,
not swept under a round-sphere approximation.
 Grade: DERIVED (exact rationals, multi-route).

 D7. Sphere cross-checks (machinery validation, not a K₆ result). 
$$
a_6(S^2)=\frac{4}{315},\qquad a_6(S^4)=\frac{74}{63},\qquad a_6(S^6)=\frac{1139}{63},\qquad
a_6^{\rm conf}(S^6)=\frac{5}{63}.
$$
Two independent routes agree to \(\sim4\times10^{-14}\) . Scale-free, normalization-robust
consistency ratio (Levi-Civita–immune, since scalars are LC-immune):
$$
a_4/a_2^2 = \frac{66}{125}.
$$
Topology used throughout: \(\chi(K_6)=6=|S_3|\) (order of the Weyl group of \(A_2\) ),
 \(\chi(S^2)=2\) , \(\chi(S^1_Y/\mathbb{Z}_2)=1\) .
 Grade: DERIVED (exact rationals; validates the \(a_6\) heat-kernel machinery on cases where it
CAN be checked in closed form — it does not by itself supply the \(K_6\) graviton value).

 D8. The graded \(a_6\) object — Layer A (closure-of-record, conservative reading): OPEN /
FAIL_VALUE_MISMATCH. 
The physical, ghost-corrected graviton+ghost \(a_6\) value is adjudicated OPEN : two independent
routes disagree. Route A (graviton \(K_6\) -bundle candidate) gives \(-43/504\) , which is
anchor-inconsistent against \(-16/315\) . Route B earns only the Bochner ghost value \(149/1008\) 
(computed with \(E=0\) ) — not the physical Faddeev–Popov ghost value ( \(E=-\mathrm{Ric}\) , giving
 \(-251/504\) ). The physical-ghost mismatch is
$$
\left|-\frac{251}{504} - \frac{149}{1008}\right| = \frac{31}{48}\approx0.6458,
$$
six orders of magnitude outside the pre-registered \(10^{-6}\) tolerance. The exact located debt:
the graviton \(\mathrm{Sym}^2(T)\) Levi-Civita off-diagonal leg (the known Levi-Civita gap:
 \(2/21\) graviton, \(1/24\) vector), which requires the SU(3) Gelfand–Tsetlin off-diagonal matrix
elements on Peter–Weyl harmonic sections — an exact, in-principle-computable, but not yet
enumerated object (the 5-Weyl-inequivalent-class hopping stratum on \(\mathrm{Sym}^2_0\) ).
The trace \(\mathrm{tr}[a_6(L_{\rm grav})]\) (the \(d=13\) vector object) is OPEN / computation
debt ; the reported sign \(C\sim-6.39\) is NOT reproduced and must not be fabricated.
 Magnitude leg (R1): DISSOLVED-as-ill-posed — at odd \(D=13\) there is no finite local \(t^0\) 
heat-kernel slot (it sits at the half-integer zeta-pole \(s=7/2\) , which is zero in dimensional
regularization with no log/anomaly slot), so no finite GeV \(^6\) value exists to anchor at all .
The correctly-posed owed object is the finite trace , not a dimensionful magnitude.
 Grade: OPEN (FAIL_VALUE_MISMATCH for the graded value; DISSOLVED-as-ill-posed for the would-be
dimensionful magnitude — two different dispositions for two different sub-objects, both
non-closing).

 D9. The \(a_6\) object — Layer B (2026 audit/completion cascade): AUDIT-CERTIFIED chain, ceiling
AUDIT-CLOSED / physics-OPEN. 
This layer banks an exact-rational \(a_6\) computation chain via completion cycles, each value
multi-route verified, but explicitly flagged headline_green=false (ceiling AUDIT-CLOSED, never
physics-CLOSED):

 Object 
 Exact value 
 Composition / note 

 Graviton \(\mathrm{Sym}^2(T_6)\) LC \(a_6/a_0\) 
 \(-6373/630\) 
 \(=-3481/360\) (algebraic) \(+(-25/56)\) (derivative sector); multi-route 

 Physical defect \(a_6\) 
 \(-7226/35\) 
 \(=21\cdot(-6373/630) - 12\cdot(-251/504)\) 

 Vector ghost ( \(E=-\mathrm{Ric}\) ) \(a_6/a_0\) 
 \(-251/504\) 
 \(=-713/1260\) (algebraic) \(+19/280\) (derivative) 

 \(\mathbb{Z}_2\) smooth equivariant defect \(\tfrac12c_3^\gamma\) 
 \(-337361/840\) 
 Donnelly equivariant construction; NOT an order-6 boundary object 

 13D bulk graded \(a_6\) (KEYSTONE) 
 \(-953329/1260\) 
 at frozen \(K_2=5\) ; multi-route 

 AUD-0059 (orbifold assembly) 
 \(-491353/630\) 
 \(=\tfrac12(-953329/1260)+(-337361/840)\) EXACTLY 

 \(K_6\) scalar \(a_6/a_0\) 
 \(992/315\) 
 \(=8017/2520\) (algebraic) \(+(-9/280)\) (derivative) 

 Bianchi-exact re-run dimensionful bulk 
 \(-2.995681680\times10^{94}\ \mathrm{GeV}^6\) 
 consistency coefficient only; never gap-closing; still route-inconsistent 

 Object-identity firewall (binding): bulk \(a_6\) ( \(-953329/1260\) / AUD-0059 \(-491353/630\) )
 \(\neq\) order-6 boundary \(a_6\) \(\neq\) \(\mathbb{Z}_2\) -defect equivariant ( \(-337361/840\) ) \(\neq\) 
graded-Casimir supertrace. These four objects are never conflated.
 Reconciliation instruction (binding): Layer A (OPEN/FAIL_VALUE_MISMATCH) is the
closure-of-record disposition; Layer B is audit-certified computational progress whose ceiling
is explicitly AUDIT-CLOSED, never physics-CLOSED. Neither promotes the gate.
 Grade: DISSOLVED-GIVEN-root is NOT claimed here — this is computational bookkeeping toward a
future closure attempt, graded OPEN at the physics level, with sub-components individually
CERTIFIED as internally-consistent exact-rational computations (AUDIT-CERTIFIED, a narrower
credit than DERIVED).

 D10. The color-factor candidate \(124/315\) — PENDING, not banked as clean. 
The scalar \(K_6\) ratio \(b_3/b_0=124/315\) (the \(t^0\) /Seeley–DeWitt \(a_{d/2}\) -bracket ratio,
computed from the actual \(K_6\) Peter–Weyl heat trace) was earlier labeled "DERIVED
dual-validated." It is now DOWNGRADED to
 DERIVED-PENDING-INDEPENDENT-TARGET-BLIND-REPRODUCTION , because the same
R2-carrying, metric-selected engine (selected at \(\mathrm{Scal}_{K_6}=7.5\) ) produced it. Binding
caveat: always report \(124/315\) with the metric-selected caveat; it is not the "clean
route-independent invariant" until an independent route reproduces it (work package H5). Only
the purely dimensionless ratios immune to Levi-Civita rescaling (e.g. \(a_4/a_2^2=66/125\) , D7
above) are currently clean in this sense.
 Grade: DERIVED-PENDING (not yet an independent-route win).

 D11. Scope-firewall certificate — the structural terminal. 
"One heat-kernel coefficient \(a_6\) is NOT a UV completion" is a certificate , proved
structurally, not empirically: the operator ladder \(a_6 < a_8 < a_{10} < \dots\) is unbounded, so
no single finite coefficient can ever settle strong-coupling (non-perturbative, all-orders)
consistency. This is exactly why the Tier-A Nonseparability screen (above) holds.
 Grade: CERTIFIED-IRREDUCIBLE (terminal as a scope wall — this is a proof of a structural
impossibility, not an unfinished computation).

 D12. Refuted magnitude. 
The dimensionful bulk figure \(-2.818\times10^{94}\ \mathrm{GeV}^6\) was REFUTED at decision
grade — it is R2-contaminated, scheme-anchored, and ill-posed at odd \(D=13\) (see D8: no finite
local \(t^0\) slot exists at odd dimension, so no such GeV \(^6\) number can legitimately exist as a
gap-closing quantity). This is a reached verdict — a piece of completed science, not an open hole.
 Grade: REFUTED / CLOSED-NEGATIVE (dead; never revived; superseded by the Bianchi-exact
re-run figure in D9, which itself remains route-inconsistent and non-gap-closing).

 D13. Granularity dissolution (repeated from Tier A3, entered here as a ledger line). 
AXIOM-COSTFLOOR dissolves exactly the \(\{a_8,a_{10},\dots\}\) runaway divergence class.
 Grade: DISSOLVED-GIVEN-root (given the Granularity root/AXIOM-COSTFLOOR) for that one class
only; the finite \(a_6\) coefficient and 9 other named walls are untouched (OPEN).

 D14. Endpoint anchor (repeated from L1, entered here as the terminal ledger line). 
The gate's published grade rests on {Δ₀>0, Lorentz-scalar proper-time floor}.
 Grade: REDUCED-TO-AXIOM (ANCHORED +1) — the fixed grade of this gate.

 E. Anchor register — consumed / reproduced / tested-against

 This gate does not consume the flavor/calibration anchor set
 \(\{M_{\rm Pl},\alpha_i(M_Z),y_t,|V_{us}|,N_\nu\}\) used elsewhere on the board. It rests on
geometric curvature invariants plus the two conditional axioms. There is no N-sigma pull to
quote anywhere in this gate: every quantity below is either an exact geometric rational or an
established physical bound, never a fit to a measured central value with an error bar.

 Object 
 Value 
 Role 
 Kind 

 \(\|\mathrm{Riem}\|^2/R^2\) 
 \(23/75\) 
 Curvature input feeding \(a_6\) 's Riem³ sector 
 DERIVED , target-blind, 3 routes (necessary, not sufficient) 

 Einstein constant \(\kappa\) 
 \(5/12\) 
 Fixes the Ricci-trace normalization of \(L_{\rm grav}\) 
 DERIVED-GIVEN-E 

 First-Bianchi residual 
 \(2.5\times10^{-16}\) 
 Target-blind correctness theorem-criterion 
 DERIVED (theorem-criterion, machine zero) 

 Color factor \(124/315\) 
 \(124/315\) 
 Candidate scalar heat-kernel ratio 
 DERIVED-PENDING (metric-selected, not independently reproduced) 

 Sphere spectra 
 \(4/315,\,74/63,\,1139/63,\,5/63\) 
 Machinery validation 
 DERIVED (exact rationals) 

 Cost-floor bounds (Margolus–Levitin, Landauer, Bekenstein) 
 see Tier A3 formulas 
 Physical grounding of AXIOM-COSTFLOOR 
 MEASURED-ANCHOR / established physics 

 \(M_*\) 
 \(\approx6.01\times10^{16}\) GeV (order-of-magnitude reading); \(7.467050992135091\times10^{16}\) GeV (Planck-normalized full-precision figure) 
 Locates the UV floor 
 DERIVED (geometry read-off); NOT a closure, NOT a new anchor 

 given- \(E\) 
 the observed matter content, which fixes which Seeley–DeWitt terms are populated 
 Single largest charged input to the whole heat-kernel ladder 
 GIVEN , not derived 

 Refuted magnitude \(-2.818\times10^{94}\ \mathrm{GeV}^6\) 
 — 
 Historical wrong value 
 REFUTED (decision-grade); dead 

 Consumed: the curvature-invariant ratios (D3, D4, D6), the sphere calibration set (D7), the
BRST-forced fiber counting (D2), and the cost-floor bounds (Tier A3) are all consumed as inputs
to this gate's structure.
 Reproduced: the sphere spectra and the first-Bianchi machine-zero are reproduced across
independent routes/engines (a positive validation of the shared machinery).
 Tested-against (not yet passed): the graded graviton+ghost \(a_6\) value is tested against a
 \(10^{-6}\) pre-registered tolerance and fails by \(31/48\) (D8) — this is the live, honestly
open residual.

 F. Anti-claims and negative controls (binding, carried verbatim in spirit)

 NOT claimed: "UQF-5C is closed / we have a UV completion of quantum gravity." A
 constructive, non-perturbative UV completion of the interacting graviton is a global open
 problem shared by every approach (strings, asymptotic safety, loops, this geometry) —
 unsolved by all of them alike. The geometry supplies the operator (D1) but no constructive
 lever across the wall.

 NOT claimed: " \(a_6\) is computed, therefore 5C is closed." This is the gate's signature
 mis-close. \(a_6\) is one missing object feeding toward a UV completion; it is never
 identical to a UV completion (D11 formalizes why).

 NOT revived: the refuted magnitude \(-2.818\times10^{94}\ \mathrm{GeV}^6\) (D12); the
 retracted \(256a^2(a^2-1)^2\) Berger derivative coefficient (only \((\nabla\mathrm{Riem})^2\) 
 survives on \(K_6\) , D6); the buggy build values \(\kappa=7/12\) , \(\|\mathrm{Riem}\|^2/R^2=31/147\) ,
 first-Bianchi \(=1/6\) — all superseded, all dead.

 NOT claimed: "we out-UV-complete every possible theory" — a universal negative over all
 conceivable mathematics. DISSOLVED as a shared ceiling on all knowledge (nobody can prove
 this for any theory), never claimed as a proven advantage of this geometry.

 NOT claimed: a third structural frontier exists. Canonically there are exactly two
 (Gap-13, UQF-4); UQF-5C is the graviton leg of the shared UQF-9 wall, carried under UQF-9/UQF-14.

 Negative control — dissolved unicorns (never claimed proven, always framed as shared
 ceilings): 
 - "No future theory could UV-complete gravity better than ours" — unprovable for anyone;
 honest ceiling is that 5C is a global wall shared by all programs.
 - " \(a_6\) alone settles the strong-coupling theory" — forbidden by the unbounded
 \(a_6<a_8<a_{10}<\dots\) ladder (D11).
 - "THE absolutely unique UV completion of the interacting graviton" — demanding forced
 uniqueness over the open-ended space of quantization schemes is a unicorn; the bounded,
 legitimate claim is narrower: this geometry supplies the operator (D1), and the missing
 object is precisely named (UQF-9).

 Fabrication guards (things this ledger explicitly does not assert a value for): 
 \(\mathrm{tr}[a_6]\) is not computed ; \(C\sim-6.39\) is not reproduced ; the positivity
 functional P is UNSELECTED (three inequivalent readings exist; no sign of \(P(a_6)\) is
 asserted); a TOTAL bulk+defect \(a_6\) is not emitted ; the order-6 mixed
 Neumann \(\oplus\) Dirichlet boundary coefficient is BLOCKED (does not exist in the published
 heat-kernel literature; the boundary tower stops at \(a_5\) ) — the gap01/uqf5c ledgers regard the
 "boundary-coefficient" framing itself as a wrong-object artifact, since a global isometric
 reflection on a closed manifold is not a manifold-with-boundary problem (the twisted trace on
 \(S^1_R/\mathbb{Z}_2\) is exactly \(1\) , \(t\) -independent, with no boundary tower); the smooth
 equivariant defect \(\tfrac12c_3^\gamma=-337361/840\) (D9) is the object that IS emitted, and it
 must not be summed into a fabricated TOTAL.

 Standing methodological guards: a reverse-engineered scheme that is true-by-construction
 RELOCATES rather than closes (the \(\kappa^3/\pi\) kill-test); a captured terminal log is never
 treated as an independent reproduction; DISSOLVED \(\neq\) SOLVED; frozen-branch identity
 provenance is an audit anchor only and never validates the physics; SELECTED \(\neq\) FORCED;
 ANCHORED \(\neq\) DERIVED; AXIOM-CLOSED \(\neq\) atomic; given- \(E\) \(\neq\) a derivation of \(E\) .

 G. Open-hole ledger (H1–H6) — tractability order, not 5C-leverage order

 Only H1 = closing UQF-9 closes the 5C roll-up. The rest sharpen 5A/5B but never cross the
firewall into a 5C close.

 ID 
 Object 
 What would close it 
 Current grade 

 H4 
 \(d=13\) \(a_6\) vector + positivity functional \(P\) 
 Select \(P\) among 3 inequivalent readings; fix the R2 engine bug; compute the graviton-minus-ghost \(a_6\) vector target-blind; evaluate \(P(a_6)\) 
 OPEN — \(P\) UNSELECTED; blocked further by a cross-script Gilkey contradiction (one script emits \(-8/405\) for \(S^2\) against the certified \(4/315\) ) that must be resolved before any \(d=13\) number is trusted 

 H5 
 Independent \(124/315\) reproduction + heat-kernel scheme object 
 Reproduce \(124/315\) on a structurally independent engine; separately settle or axiomatize the shared one-loop scheme object 
 OPEN / DERIVED-PENDING 

 H2/P0 
 Does gravity own a minimal length, or inherit \(R_0\) ? 
 Target-blind determination; "inherits \(R_0\) " confirms no finite-grain shortcut (consistent with T-DEEP); "owns its own scale" opens a genuinely new route 
 OPEN (cleanest bounded handle on the wall) 

 H6 
 \(\mathbb{Z}_2\) orbifold-defect heat-kernel / order-6 boundary object 
 A specialist construction (new literature result) or a proof the Donnelly equivariant form supplies it 
 BLOCKED — object does not exist in the published literature 

 H1 
 UQF-9: constructive UV completion / strong-coupling closure 
 (1) target-blind non-Gaussian fixed point construction (constructive close), (2) proof no such fixed point exists on this geometry ( CLOSED-NEGATIVE , a valid terminus), or (3) neither runnable → emit a WALL RECORD 
 OPEN (global wall) — the only leg that would close the 5C roll-up 

 H3 
 Above-cutoff graviton unitarity (UQF-14) 
 Inherits H1's disposition; no independent close available 
 OPEN (downstream of H1) 

 False-openness trap (named explicitly so it is never repeated): marking H1 as
 bounded:true with what_would_close_it="solve QG UV completion" is a category error — H1 is a
 wall , not a bounded task with a work plan ("a horse"). The correct record is a WALL RECORD:
name the construction precisely, classify it OPEN/wall, record the hidden bridge (H1→UQF-9), and
mark it SHARED across UQF-5C/UQF-9/UQF-14.

 H. Credit-ladder summary (every leg, one line each)

 Leg 
 Grade 

 Operator supply \(L_{\rm grav}=-(\nabla^2+E)\) (D1) 
 DERIVED-GIVEN-E 

 Ghost-corrected fiber weight \(91-2\cdot13=65\) (D2) 
 DERIVED-GIVEN-E (BRST-forced) 

 \(\|\mathrm{Riem}\|^2/R^2=23/75\) (D3) 
 DERIVED (target-blind, 3 routes) 

 Einstein constant \(\kappa=5/12\) (D4) 
 DERIVED-GIVEN-E 

 First-Bianchi machine-zero (D5) 
 DERIVED (theorem-criterion) 

 Derivative-sector invariants (D6) 
 DERIVED (exact rationals) 

 Sphere cross-checks (D7) 
 DERIVED (exact rationals; machinery validation) 

 Graded graviton+ghost \(a_6\) value (D8, Layer A) 
 OPEN / FAIL_VALUE_MISMATCH 

 Dimensionful \(a_6\) magnitude at odd \(D=13\) (D8) 
 DISSOLVED-as-ill-posed (no slot exists) 

 Audit-cascade exact-rational chain (D9, Layer B) 
 AUDIT-CERTIFIED components; OPEN at physics level (ceiling AUDIT-CLOSED, physics-OPEN) 

 Color factor \(124/315\) (D10) 
 DERIVED-PENDING 

 Scope-firewall certificate: \(a_6\ne\) UV completion (D11) 
 CERTIFIED-IRREDUCIBLE 

 Refuted magnitude \(-2.818\times10^{94}\) GeV \(^6\) (D12) 
 REFUTED / CLOSED-NEGATIVE 

 Granularity dissolves \(\{a_8,a_{10},\dots\}\) (D13) 
 DISSOLVED-GIVEN-root (one class only) 

 Endpoint anchor {Δ₀>0, Lorentz-scalar floor} (D14) 
 REDUCED-TO-AXIOM (ANCHORED +1) — the fixed grade 

 H1 = UQF-9 roll-up 
 OPEN (global wall) — shared, not private 

 I. Endpoint line

 UQF-5C is ANCHORED +1 / REDUCED-TO-AXIOM on the conditional axiom pair
 \(\{\Delta_0>0,\ \text{Lorentz-scalar proper-time floor}\}\) , read as TERMINAL +
RESIDUALS-SHOWN. The frozen 13D geometry (× Stage: \(\mathcal{M}_4\times K_6\times S^2\times
S^1_Y/\mathbb{Z}_2\) , \(D=13\) ; ⊕ Rulebook: de-Donder + FP ghosts + \(\overline{\rm MS}\) heat-kernel
scheme; ⊗ Actors: \(L_{\rm grav}=-(\nabla^2+E)\) ) cleanly supplies the interacting graviton's
kinetic operator and its BRST-forced ghost-corrected fiber weight ( \(91-2\cdot13=65\) ), backed
by a proved target-blind curvature ratio ( \(23/75\) ) and a machine-zero correctness theorem
(first-Bianchi). The AXIOM-COSTFLOOR granularity root dissolves exactly one UV-divergence
class (the \(a_8,a_{10},\dots\) runaway tower) under a proved Lorentz-scalar floor, while leaving
 10 other named walls open, including the finite \(a_6\) coefficient itself — whose graded
graviton+ghost value is presently OPEN / FAIL_VALUE_MISMATCH at \(31/48\) outside the
pre-registered tolerance, with the would-be dimensionful magnitude DISSOLVED-as-ill-posed at
odd \(D=13\) . The scope-firewall certificate — one heat-kernel coefficient cannot be a UV
completion, given the unbounded \(a_6<a_8<a_{10}<\dots\) ladder — is CERTIFIED-IRREDUCIBLE . The
gate's roll-up (H1) is carried as a shared global wall with UQF-9 (inherited by UQF-14):
 OPEN , not private to this geometry, not closeable by any per-gate plug, and not to be
conflated with the anchored status of the operator-supply and axiom legs above. Both the ANCHORED
+1 pill and the OPEN (global wall) roll-up describe the same physics faithfully; neither is
softened, upgraded, or hidden behind the other. PROMOTIONS:0.