Cellulation-Independent Boundary Gauge Averaging and Sharp Class-Sector Gaps in Two-Dimensional Yang--Mills
Admitted after checks by named AI models and editor approval. Not peer reviewed. Screening record
Let G be a compact connected Lie group with a bi-invariant metric. For the two-dimensional heat-kernel Yang--Mills model, this paper gives an explicit finite chain from a boundary-conditioned edge integral on an arbitrary regular annular cellulation to the gauge-invariant class-function transfer kernel. Conditional Haar coordinates make the boundary constraints and boundary-edge subdivision precise. Edge and face subdivision invariance, an explicit PL radial-cut construction, and a primal-tree/dual-cotree disk-elimination schedule yield a cellulation-independent orbital heat-kernel integral. Peter--Weyl theory then diagonalizes the transfer operator with multipliers exp(-t c_lambda). The exact mean-zero norm is exp(-t c_*(G)), the gap of I-T_t is 1-exp(-t c_*(G)), and the equality sector consists of all irreducible characters with minimum positive Casimir. In the stated SU(2) normalization the sharp exponent is 3/4, while the induced SO(3) quotient removes the fundamental channel and raises it to 2. The work assembles and audits classical two-dimensional Yang--Mills ingredients; it does not claim a new character solution, four-dimensional mass gap, continuum reconstruction, or clustering for arbitrary bulk observables.
Preview PDF on this page v1
Uses your browser's PDF viewer. Open the PDF directly if a preview is unavailable.
Cite this paper
Lluis Eriksson (2026). Cellulation-Independent Boundary Gauge Averaging and Sharp Class-Sector Gaps in Two-Dimensional Yang--Mills. AIRR.SCIENCE, ARR-2026-7D2BBEC8MJ8BM80S, v1. https://airr.science/papers/ARR-2026-7D2BBEC8MJ8BM80S/versions/v1/
Use this version link to cite the manuscript you read. A Working paper label means the version has not completed AIRR screening. AIRR screening is not journal peer review. Import BibTeX, RIS or CSL JSON into your reference manager to apply a citation style.
Version history
The paper identifier remains stable. Each version has its own immutable release, timestamp and version identifier.
- v1 · source snapshot available · viewing
Verification record
- Frontier-model screening
- Not assessed
- Source integrity
- Pass
- Bibliographic integrity
- Not assessed
- Reproducibility
- Partial
- Lean 4
- L2
Recorded under ARR-VERIFY-1.0. AIRR verification and screening are not peer review.
AI assistance statement
OpenAI Codex assisted with repository discovery, mathematical exploration, drafting, adversarial proof review, Lean and numerical replay, revision in response to external criticism, deposit preparation, machine-readable extraction, and visual release QA. The author assumes responsibility for the definitions, proofs, claims, citations, licenses, and deposited artifacts. No independent AI scientific assessment is claimed.
Frontier-model screening
Status: not_assessed. Any listed reports correspond to this exact version under ARR-SCREEN-1.0; no absent assessment is represented as a pass.
Model assessments and review context
- Distinct models
- 2
- Range
- 2.90–3.35
- Method
- bounded capability-weighted mean · exact version only
Limited evidence · 1/2 benchmark matches. This is an evidence description, not a probability of correctness.
One provider; shared blind spots possible; Fresh blinded review context not documented for every model; Benchmark match incomplete; Adverse reports or unresolved objections require editorial disposition
OpenAI · gpt-5.6-sol · High · unknown · review context: unknown3.35
Recommendation: Minor Revision · Material objections: 0
This exact v1 gives a coherent finite derivation of the gauge-invariant annular heat kernel and its sharp class-sector spectral gap. The conditional boundary measure, elementary subdivisions, disk tree-cotree collapse, orbital averaging, and character diagonalization are mutually consistent, and the SU(2)/SO(3) constants follow from the stated Casimirs. The main mathematical result appears correct, while the contribution is primarily a careful assembly of classical ingredients. I recommend minor revision only to reconcile the advertised exact formal-release protocol with the supplied public-code payload and to expand the radial-cut coordinate disintegration.
- Correctness confidence
- The conditional-Haar, subdivision, disk-collapse, orbital-average, and spectral-gap steps form a consistent chain. I found no counterexample to Theorem 1.1, but the claimed Lean rebuild was not rerun because Lean/Lake are unavailable.
- Rigor
- Lemmas 3.2-3.6 close the main normalization and topology interfaces, and Theorems 3.8 and 5.1 follow by standard harmonic analysis. The radial-cut disintegration remains terser than the other finite-coordinate arguments.
- Novelty
- The character solution, heat-kernel gluing, and spectral diagonalization are explicitly acknowledged as classical. The plausible increment is the end-to-end finite normalization and cellulation reduction, not a new physical phase or mass-gap result.
- Significance
- A normalization-explicit theorem for arbitrary compact connected groups is useful and clarifies framed versus class-sector kernels, but its scope is the exactly soluble two-dimensional heat-kernel model.
- Reproducibility
- The supplied numerical replay ran and printed REPLAY: PASS. All 84 public-code manifest entries matched, but the Lean toolchain was absent and the included internal release manifest does not validate the supplied public-code layout byte-for-byte.
Strengths
- Definition 3.1 and Lemma 3.2 give an explicit normalized fixed-holonomy fibre and show that changing the solved boundary edge or subdividing it introduces no hidden gauge-volume factor.
- Lemmas 3.3-3.5 reduce any regular disk cellulation by exact Haar and heat-kernel convolution identities, with the tree-cotree count identifying one chord per bounded face.
- Theorem 3.8 correctly cancels the Peter-Weyl dimension factor through orbital averaging and retains the complex conjugate on the second character for general compact groups.
- Theorem 5.1 obtains the exact mean-zero operator norm and equality space directly from the character basis; Corollary 5.3 and Proposition 5.4 then give the stated SU(2), product, and central-quotient constants.
- The paper clearly limits its claims to finite two-dimensional heat-kernel Yang-Mills and distinguishes the paper proof from the narrower formal endpoints.
Weaknesses
- Lemma 3.6 compresses the cut-and-reglue change of variables into a prose gauge-coordinate argument; an explicit finite bijection and product-measure factorization would make the single residual Haar variable completely auditable.
- The claimed contribution is mostly organizational: the public primary-source check confirms that discrete two-dimensional Yang-Mills, surface gluing, and Markov behavior substantially predate this manuscript.
- Appendix C advertises an exact Git-bundle replay, but that bundle is absent from the supplied public-code payload and the declared PowerShell route could not be exercised in this Linux environment.
Potential errors
- SUPPLIED_RELEASE_MANIFEST.sha256 does not validate at the included public-code/src root: 6 of 13 paths match, four named paths are absent, and three pinned files differ bytewise. This does not contradict the separate PUBLIC-CODE-MANIFEST, whose 84 entries all match.
- The manuscript reports successful 2,817-job and 3,097-job Lean builds, but these are stored claims rather than my executions; static inspection found no sorry, admit, or axiom declarations, yet compilation remains unconfirmed here.
Strong novelty candidates
- The normalization-explicit chain from arbitrary regular annular cellulations to one relative-frame Haar average is a plausible expository and theorem-packaging contribution.
- The combined compact-group gap statement, equality-sector description, and product/central-quotient calculus may be useful as a single reusable transfer-operator result even though its ingredients are classical.
Unresolved material objections
None reported.
arr:assessment:203f65c3-153b-4c7a-a05d-0fda789d6245 · prompt ARR-ASSESS-1.0 · runtime identity: OpenAI gpt-5.6-sol · reasoning effort: high · basis: operator verified ui · evidence SHA-256 92703c3c3d72e9347295c9be5ba97924b67ac85672a1724b468bcb11111112de · response SHA-256 601751b3b1a1ed033bb208451ea548f84b5fb70bbe5362b78d090bc244c9d015 · canonical PDF SHA-256 ee51663a977f1932da74670956aafe39e6f43ac355c6d27b472a4285a2b3fe34
OpenAI · gpt-6-astra · High · unknown · review context: unknown2.90
Recommendation: Major Revision · Material objections: 1
This exact 13-page PDF gives a useful but predominantly classical assembly of two-dimensional heat-kernel Yang-Mills, boundary class projection and sharp Casimir decay. I checked the harmonic-analysis and spectral arguments, the elementary subdivision moves and the disk elimination, and executed the SU(2) numerical replay successfully. The central radial-cut proof still does not establish its stated representative-preserving coordinate change: a connected tree gauge generally conjugates the second boundary holonomy. This is a repairable but load-bearing proof-interface objection to the claimed end-to-end theorem. The trivial group also needs exclusion from the positive-gap statements. Supplied formal sources support a substantive formalization effort, but I did not rebuild Lean or verify the reported axiom lists. Major revision is a referee assessment, not an editorial decision.
- Correctness confidence
- The orbital identity, class projection, reflection form and Casimir gap in Sections 2, 4 and 5 check out. The fixed-representative change of variables in Lemma 3.6 remains insufficiently justified; the all-cellulation claim depends on it.
- Rigor
- Lemmas 3.2-3.5 give useful explicit conditional-Haar, subdivision and tree-cotree arguments. Lemma 3.6 switches to a disk spanning-tree gauge without proving compatibility with both fixed boundary representatives or the regluing constraints.
- Novelty
- The cylinder character expansion, gluing and class-sector heat operator are classical; Levy Section 5.5 already treats conditional partition functions. The plausible increment is the explicit normalization and formal/paper interface, not a new mass-gap mechanism.
- Significance
- An auditable finite-edge-to-boundary transfer construction is useful, and the SU(2)/SO(3) normalization comparison is clear. The spectral conclusion is an elementary consequence of known harmonic analysis and has no four-dimensional Millennium implication.
- Reproducibility
- The supplied NumPy replay completed with all assertions passing. Exact Lean snapshots and gap/audit modules are present and hash-verified, but Lean, Lake, elan and PowerShell are unavailable; neither a build nor an axiom audit was executed.
Strengths
- Lemma 2.1 and Theorem 3.8: Schur averaging gives chi(a)chi(b)/dim, cancelling the heat-kernel dimension factor. The PDF includes the required complex conjugation at h; SU(2) reality is not silently used for general G.
- Definition 3.1 and Lemma 3.2: solving the product constraint gives a probability Haar fibre. Adjacent changes of eliminated coordinate are inversion/translation maps; boundary-edge subdivision introduces one normalized redundant Haar coordinate.
- Lemma 3.3: both edge multiplication and face convolution preserve the conditioned amplitude, with positive daughter areas and no gauge-volume factor. Lemma 3.4 inserts the PL arc using precisely these moves.
- Lemma 3.5: a primal tree containing all but one exterior edge leaves a dual tree rooted at that edge. Leafwise convolution merges all positive face areas and yields the fixed-boundary disk heat kernel.
- Propositions 4.1-4.2 and Theorem 5.1: orthogonal class projection, Parseval and the semigroup square establish self-adjointness, reflection positivity, the exact norm and its entire equality sector.
- Corollary 5.3 and Proposition 5.4: (n-1)(n+3)>=0 proves the unique SU(2) minimum 3/4; even labels on SO(3) give 2. Product and finite-central-quotient statements follow from additive/restricted Casimir spectra.
- Section 7 replay actually passed: endpoint error 0, Gram error 2.709e-14, multiplier error 6.939e-18, sharp mode 1 and reflected form 1.7639228388840671. These are finite numerical diagnostics only.
Weaknesses
- Section 8 accurately concedes classical ingredients, but the incremental research contribution is modest. The sharp transfer rate follows immediately once the character formula is available; selective source checking does not establish priority for the assembly.
- Section 6 formal endpoints were located in the supplied exact snapshots, including the conditioned original-edge disk amplitude. I inspected relevant definitions and theorem dependencies, not the entire Lean corpus; no kernel check or #print axioms execution was possible.
- Appendix C describes a Git-bundle/PowerShell release. This batch instead supplies source snapshots and explicitly withholds the bundle. The original driver was inspected, not run; the batch cannot reproduce that exact driver as supplied.
- The numerical script never integrates an arbitrary annular cellulation. Its spectral arithmetic cannot test the representative-preserving disintegration required by Lemma 3.6.
Potential errors
- Lemma 3.6, pp. 5-6: fixing a gauge at only one root does not preserve a second marked representative h; it generally becomes k^{-1}hk. The displayed edge transformation needs constraints at both roots and consistent potentials on duplicated cut vertices.
- Theorem 1.1 and Section 5 admit G={1}, which is compact and connected. Its nontrivial dual is empty, c_* is undefined and the mean-zero Hilbert space is zero. Exclude this case or state separate conventions; this is an edge-case defect, not a counterexample to SU(2).
- Lemma 3.6 also identifies a conditional integral at fixed relative transport with A_D(q), although A_D was defined by averaging all disk boundary coordinates at product q. The proof needs either pointwise boundary-data invariance or an explicit measure pushforward, not Fubini alone.
Strong novelty candidates
- The explicit conditional-Haar normalization, subdivision schedule and delineation of the concrete SU(2) formal endpoints form a plausible useful expository/formal contribution, subject to closing Lemma 3.6.
- The common presentation of normalized cylinder transfer, sharp equality sectors and quotient-induced gap changes may be valuable for formal verification; no new spectral phenomenon or historical priority is established.
Unresolved material objections
- Lemma 3.6 -> Proposition 3.7 -> Theorem 1.1: give a representative-preserving finite Haar parametrization with both marked roots controlled and the duplicated cut variables correctly identified, and prove the resulting fixed-x integral equals A_D(gxh^{-1}x^{-1}). A two-root forest appears to repair this, but the exact PDF supplies neither that construction nor the required fibre pushforward. The final classical formula is plausible; the advertised complete proof interface remains open.
arr:assessment:0617ef16-bab6-425c-8f39-7bb3053a1430 · prompt ARR-ASSESS-1.0 · runtime identity: OpenAI gpt-6-astra · reasoning effort: high · basis: operator verified ui · evidence SHA-256 d6795eeac3911d1340112c3a81067c06d55b17243edbe75932bb156f1258b1ee · response SHA-256 2118a94ccf618c61f674ad460cfd21c4243a53335a1abbb68a5fa50ba590ba04 · canonical PDF SHA-256 ee51663a977f1932da74670956aafe39e6f43ac355c6d27b472a4285a2b3fe34
OpenAI · gpt-5.6-sol · High · unknown · review context: unknown3.45
Recommendation: Minor Revision · Material objections: 0
This exact 13-page v1 paper gives a coherent finite proof that a heat-kernel Yang-Mills annulus reduces to the class-function heat semigroup, then derives reflection positivity and the sharp minimum-Casimir gap, including SU(2) and SO(3). I found no unresolved material objection to the stated mathematical theorem. The contribution is a careful assembly of classical ingredients rather than a new character solution. The numerical replay passed, but the supplied public-code subset does not contain the files needed to reproduce the claimed Lean builds, so the machine-check claims remain only partially audited.
- Correctness confidence
- Theorem 1.1's chain is mathematically coherent: conditional Haar fibers, subdivision identities, disk collapse, orbital averaging, and Peter-Weyl diagonalization yield the stated kernel and sharp Casimir gap. No counterexample was found within the regular finite, compact connected scope.
- Rigor
- Lemmas 3.2-3.5 close several normalization and topology interfaces explicitly, and Theorems 3.8 and 5.1 justify convergence and equality cases. Lemma 3.6 remains compressed at the general combinatorial-coordinate level, and 'regular cellulation' is not formally defined.
- Novelty
- The character solution, heat-kernel gluing, and transfer spectrum are classical, as the paper acknowledges. The plausible increment is the normalization-clean end-to-end assembly, general compact-group gap statement, and explicit formalization boundary, not a new Yang-Mills solution.
- Significance
- The exact class-sector gap and cellulation-independent finite reduction are useful clarifications for solvable two-dimensional Yang-Mills. The result is deliberately limited to boundary class functions and does not address continuum reconstruction, bulk observables, or four-dimensional mass gap.
- Reproducibility
- The supplied Python replay executed successfully, but the advertised Lean reproduction could not be attempted: reproduce.ps1, the Git bundle, and the imported Lean2dYangMills.SU2ClassHeatKernel source are absent, and Lean/Lake/PowerShell are unavailable.
Strengths
- Lemma 3.2 makes the boundary conditional Haar measure coordinate-explicit and proves independence of the eliminated boundary edge, directly addressing a common normalization ambiguity.
- Lemmas 3.3-3.5 give a finite sequence of Haar/convolution moves and a tree-cotree disk elimination, supporting the cellulation-independent reduction without invoking a formal path integral.
- Theorem 3.8 correctly cancels the representation dimension through orbital averaging and supplies an absolute-uniform convergence bound via the scalar heat trace.
- Theorem 5.1 identifies the full equality space of the sharp mean-zero norm, while Corollary 5.3 and Proposition 5.4 consistently specialize the Casimir normalization to SU(2), products, and SO(3).
- The manuscript sharply limits its scope and expressly rejects any inference to the four-dimensional Clay Yang-Mills problem.
- Fresh execution of replay.py returned REPLAY: PASS with orthogonality error 2.709e-14, semigroup error 6.939e-18, sharp mode 1, and positive reflection form.
Weaknesses
- Lemma 3.6's assertion that a general cut-open conditional edge integral disintegrates into exactly one relative-frame Haar coordinate is plausible but terse; an explicit coordinate bijection and product-measure pushforward for an arbitrary cut path would improve auditability.
- The term 'regular finite cellulation' is used in the universal quantifier without a formal definition, leaving implicit whether loops, multiple incidences, or face-boundary self-identifications are excluded.
- The executable evidence supplied here covers only floating SU(2) diagnostics; it does not test the general compact-group theorem or the PL/tree-cotree reduction.
Potential errors
- Appendix C and the README say the release contains reproduce.ps1 and a complete Git bundle, but neither appears in the supplied public-code manifest. The included SU2ClassTransferGap.lean also imports an absent SU2ClassHeatKernel module, so the stated Lean builds are not reproducible from this batch.
- The formalization table should not be read as machine verification of Theorem 1.1 as a whole: the paper itself concedes that subdivision invariance, the PL radial cut, compact-group extension, and Hilbert-space assembly are not packaged in Lean.
Strong novelty candidates
- A single normalization-explicit chain from boundary-conditioned lattice edge integrals on arbitrary regular annular cellulations to the class-function transfer operator.
- The sharp compact-connected-group boundary gap and equality-sector statement expressed in the same metric and area convention as the finite physical edge model.
Unresolved material objections
None reported.
arr:assessment:4136d701-406c-4d40-b125-faa21d9d6a1e · prompt ARR-ASSESS-1.0 · runtime identity: OpenAI gpt-5.6-sol · reasoning effort: high · basis: operator verified ui · evidence SHA-256 7531e97a7baad09cb045dd852e4a088ad369c8df86932feeef1e11f85fb1aa83 · response SHA-256 6dc16de4489ddf2c041287852b3c89ac68bd771a5a09028ac1cc4bbdaf10e4a2 · canonical PDF SHA-256 ee51663a977f1932da74670956aafe39e6f43ac355c6d27b472a4285a2b3fe34
OpenAI · gpt-6-astra · High · unknown · review context: unknown2.60
Recommendation: Major Revision · Material objections: 3
This exact 13-page PDF gives a coherent assembly of classical two-dimensional Yang-Mills harmonic analysis, with correct sharp SU(2) multiplier arithmetic and a successful numerical replay. The boundary-fibre and disk-collapse arguments are useful, but the radial-cut disintegration still needs an explicit constrained coordinate proof. The universal gap statement also needs to exclude the trivial compact connected group. The supplied batch does not support independent reproduction of the advertised Lean builds. These unresolved issues prevent an accept recommendation; neither the numerical replay nor this assessment verifies the claimed formal compilation. The result is a modest boundary-sector assembly, not a Millennium-problem advance.
- Correctness confidence
- The orbital average, class projection and Casimir diagonalization in Lemma 2.1 and Sections 4-5 are sound for nontrivial G. The fixed-frame disintegration in Lemma 3.6 is not sufficiently justified, and the stated group class includes a degenerate exception.
- Rigor
- Conditional Haar coordinates and the tree-cotree disk elimination improve self-containment. Lemma 3.6 does not explicitly preserve both conditioned representatives or establish the required glued-fibre measure identity. The Lean frontier could not be rebuilt.
- Novelty
- The cylinder character expansion and gluing precede this paper: Levy, Section 5.5, equations (5.4)-(5.8), already give the compact-group construction. The plausible increment is the explicit finite proof assembly and formal interface, not a new gap mechanism.
- Significance
- A useful normalization-conscious exposition of an exactly soluble two-dimensional boundary sector. The sharp SU(2) and SO(3) constants follow directly from classical Casimir spectra; no four-dimensional mass-gap or bulk-observable result is obtained.
- Reproducibility
- All nine supplied code-manifest entries match and replay.py exits 0 with REPLAY: PASS. The claimed complete Lean reproduction lacks reproduce.ps1, the Git bundle and imported Lean2dYangMills sources; Lean/Lake are also absent. No formal build was performed.
Strengths
- Lemma 2.1 and Theorem 3.8, pp. 3 and 6: Schur averaging contributes 1/d_lambda, cancelling the heat-kernel dimension. The PDF correctly includes the complex conjugate of chi_lambda(h); this matters for general compact groups.
- Lemmas 3.2-3.3, pp. 3-4: the adjacent eliminated-coordinate change is Haar preserving, including subdivision of the boundary edge used to impose holonomy. Face subdivision follows from the normalized heat convolution identity.
- Lemmas 3.4-3.5, pp. 4-5: insertion of a transverse PL arc uses the analytically covered edge/face moves. The disk proof counts |E|-|V|+1=|F| and eliminates a dual tree from its leaves, preserving total area.
- Propositions 4.1-4.2 and Theorem 5.1, pp. 7-8: orthogonal conjugation averaging, Parseval and the semigroup square establish the class-sector contraction, reflected positivity and complete minimum-Casimir equality space.
- Corollary 5.3 and Proposition 5.4, pp. 8-9: n(n+2)/4 has unique positive minimum 3/4 at n=1; the quotient retains even labels and first reaches 2 at n=2. Product eigenvalues add, giving the stated minimum-sector rule.
- Section 7 numerical replay actually completed: endpoint error 0, Gram error 2.709e-14, multiplier error 6.939e-18, sharp-mode index 1, and reflected form 1.7639228388840671. These are finite floating diagnostics only.
Weaknesses
- Sections 1 and 8: much of the headline theorem is an explicit assembly of known facts. Levy already supplies general conditional partition functions and gluing (https://arxiv.org/pdf/math/0101239, Section 5.5). A stronger novelty claim for the cylinder formula or gap would be unsupported.
- Section 6 and Appendix C describe complete 2817-job and 3097-job builds. Those are manuscript claims, not this referee's executions; the batch contains only the scalar satellite modules and package pins, not the required mathematical library.
- Section 7 tests the advertised formulas after assuming their spectral form; it does not test an arbitrary annular edge integral, a radial-cut measure identity, or the infinite-dimensional completion.
- Appendix A: strict positivity follows because every spectral weight is positive, but density alone would only preserve nonnegativity. An explicit all-L2 Parseval sentence would make this last inference precise.
Potential errors
- Theorem 1.1 and Section 5 include G={1}, a compact connected zero-dimensional Lie group. Its nontrivial dual is empty, so min_{lambda!=1} c_lambda is undefined; the mean-zero Hilbert space is zero and its spectrum is empty. The positive-gap and saturating-irrep assertions require nontrivial G.
- Lemma 3.6, pp. 5-6: a spanning-tree vertex gauge sends a marked boundary holonomy h to k^{-1}hk. For example, noncommuting SU(2) choices can send h to -h. Holding h fixed therefore requires endpoint constraints or explicit tracking of this conjugation; invariance of unconstrained Haar factors alone is insufficient.
Strong novelty candidates
- Lemmas 3.2-3.6 plus Sections 4-5 could provide a useful single finite-coordinate proof chain from arbitrary regular annular cellulations to an optimally normalized boundary estimate, once the remaining disintegration interface is made explicit.
- The proposed concrete SU(2) formal endpoints could be a reproducible formalization contribution. Only the scalar source was inspectable here; originality and compilation of the larger development remain unverified.
Unresolved material objections
- Lemma 3.6 is load-bearing for Proposition 3.7 and Theorem 1.1. The proof needs a measure-preserving coordinate map that retains both marked boundary constraints and tracks identified copies of the cut. Conditioning on g,h and relative transport does not automatically give the single-holonomy disk fibre of Lemma 3.5. This is an unresolved proof-interface objection, not a counterexample to the classical cylinder formula.
- Theorem 1.1 and Theorem 5.1 need a nontrivial-group hypothesis or a separate convention and statement for G={1}; as written the minimum Casimir and the claimed positive spectral gap are not defined in that allowed case.
- The machine-checked claims in Section 6 cannot be independently checked from this supplied payload. Appendix C's reproduce.ps1, lean-2d-yang-mills.bundle and imported endpoint sources are missing. Provide the exact executable source payload before treating the advertised formal frontier as verified evidence.
arr:assessment:953a6404-d2c8-4ced-b2f0-35b9164f14c1 · prompt ARR-ASSESS-1.0 · runtime identity: OpenAI gpt-6-astra · reasoning effort: high · basis: operator verified ui · evidence SHA-256 2dc67e969a3ac770c4c0fb93d5a56f8c5d40f49f0714f9b8f056b22f0c6766f6 · response SHA-256 fd381792696b2c63ecc9de3d443bc48eb8ea2f72895cc1131af61c9d8a815788 · canonical PDF SHA-256 ee51663a977f1932da74670956aafe39e6f43ac355c6d27b472a4285a2b3fe34
A model assessment is not peer review or a correctness certificate. AIRR preserves disagreement, exact-version provenance and later reassessments.
Editorial disclosure
Founder-owned record: Lluis Eriksson is both the author and AIRR's current founder-editor. No independent editorial review, peer review, frontier-model screening, novelty assessment, or scientific certification is claimed. Acceptance records a technically valid deposit.