Pith. sign in

REVIEW 1 major objections 5 minor 21 references

Graded Frobenius Algebras

T0 review · 1 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

Pith's one-line read This paper constructs determinant-twisted graph cobordism PROPs and proves that (c,d)-graded Frobenius algebras are exactly the map data satisfying five explicit sign-bearing relations, with suspension shifting (c,d) to (c−1,d+1).

desk verdict A careful, mostly rigorous construction of a graded Frobenius PROP with explicit signs; the main weakness is an unproved normal-form assertion in the sufficiency proof of Theorem 5.1. read the letter →

arxiv 2412.13909 v2 pith:VVR6FFIB submitted 2024-12-18 math.AT math.QA

classification math.ATmath.QA MSC 18M0557R5655P50
keywords gradedFrobeniusalgebrasPROPs2DTQFTsgraphcobordismsdeterminanttwistingssuspensionofsignconventionsstringtopology
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper gives a workable definition of a graded Frobenius algebra when the multiplication and comultiplication have arbitrary integer degrees $c$ and $d$. It builds categories $pGCob_{c,d}$, $fGCob_{c,d}$, and $GCob_{c,d}$ whose morphisms are graph cobordisms twisted by determinant line bundles, so the degrees are carried by the geometry: the multiplication generator has degree $c$ and the comultiplication generator degree $d$. The main theorem characterizes algebras over these PROPs as the data $(A,\mu,\eta,\nu,\varepsilon)$ satisfying graded associativity, unitality, coassociativity, counitality, and the graded Frobenius relation with the explicit signs of Theorem 5.1, plus graded commutativity or graded symmetry in the symmetric cases. A central consequence is that the chosen signs are stable under suspension: suspending a $(c,d)$-graded algebra yields a $(c-1,d+1)$-graded algebra, so the convention does not depend on a choice of shift. This matters because the naive ungraded-sign definition collapses to zero when $c-d$ is odd, while the motivating examples—manifold cohomology, Hochschild homology, and string topology—are non-trivial in those ranges.

What carries the argument

The load-bearing object is a family of PROPs—symmetric monoidal categories whose objects are finite sets and whose morphisms are operations with multiple inputs and outputs—built from graph cobordisms. For a graph $G$ from $X$ to $Y$, the morphism space uses $\det_{c,d}(G)=\det(G,\partial_{\mathrm{in}})^{\otimes c}\otimes\det(G,\partial_{\mathrm{out}})^{\otimes d}$, where $\det$ is the determinant (top exterior power) of relative homology, concentrated in degree minus the Euler characteristic; this twist places the multiplication generator in degree $c$ and the comultiplication generator in degree $d$. The determinant isomorphisms—additivity over short exact sequences and the identification of the determinant of a chain complex with the determinant of its homology—allow the paper to compute how edge collapses and gluings act on orientations, which yields the explicit signs in Theorem 5.1. The suspension PROP $\Sigma=\mathrm{End}_{\Sigma 1}$ supplies the universal shifting property, and the isomorphisms $\Sigma\otimes GCob_{c,d}\cong GCob_{c-1,d+1}$ (and the analogues for $fGCob$ and $pGCob$) transfer suspension to algebras.

What would settle it

The simplest check is Example 5.8: suspend the explicit rank-two $(c,d)$-graded algebra $R_{c,d}$ and verify with the paper's formulas that it satisfies the $(c-1,d+1)$ relations; any extra sign in the suspended associativity or Frobenius relation would disprove suspension stability.

Watch

Extended reading notes

Core claim

The paper's central claim is Theorem 5.1: for integers $c,d$ and a monoidal category $C$ enriched over graded abelian groups, the data $(A,\mu,\eta,\nu,\varepsilon)$ with $A\in C$, $\mu\in C_c(A\otimes A,A)$, $\eta\in C_{-c}(1,A)$, $\nu\in C_d(A,A\otimes A)$, $\varepsilon\in C_{-d}(A,1)$ defines a monoidal functor $pGCob_{c,d}\to C$ uniquely up to isomorphism if and only if the graded associativity, unitality, coassociativity, counitality, and Frobenius relations hold, namely $\mu\circ(\mu\otimes\mathrm{id})=(-1)^c\mu\circ(\mathrm{id}\otimes\mu)$, $(-1)^c\mu\circ(\eta\otimes\mathrm{id})=(-1)^{c(c-1)/2}\mathrm{id}=\mu\circ(\mathrm{id}\otimes\eta)$, the dual relations with $d$, and $(\mu\otimes\mathrm{id})\circ(\mathrm{id}\otimes\nu)=(-1)^{cd}\nu\circ\mu=(\mathrm{id}\otimes\mu)\circ(\nu\otimes\mathrm{id})$. In a symmetric category, adding $\mu\circ\tau=(-1)^c\mu$ characterizes symmetric monoidal functors out of $GCob_{c,d}$, while adding $\varepsilon\circ\mu\circ\tau=(-1)^c\varepsilon\circ\mu$ characterizes functors out of $fGCob_{c,d}$. The paper further establishes that these signs are not an artifact: no choice of orientations removes them, and with the chosen orientations they are preserved under suspension, which shifts $(c,d)$ to $(c-1,d+1)$.

Load-bearing premise

The load-bearing premise is that the graph categories faithfully encode the geometric cobordism categories, so that no relations are lost or added when surfaces are replaced by graphs; if two graphs represented the same surface without being connected by graph morphisms, the twisted morphism spaces would not describe 2D TQFTs.

Editorial extensions

If this is right

  • For a closed oriented $d$-manifold, the cup product and the Thom intersection coproduct give $H^*(M)$ a $(0,d)$-graded commutative Frobenius algebra, and the Poincaré coproduct is the suspended Thom coproduct on $\Sigma^{-d}H^*(M)$, which explains the sign difference between the two conventions.
  • For a $(0,d)$-graded symmetric Frobenius algebra $A$, the normalized Hochschild homology $HH_*(A)$ carries a $(d,d)$-graded structure with the explicit operations of Example 6.3, matching the determinant-twisted open-closed TQFT description.
  • Suspension gives an equivalence between $(c,d)$-graded and $(c-1,d+1)$-graded algebras over the PROPs, so $c+d$ is a suspension invariant; every non-trivial algebra with $c+d\neq0$ contains the rank-two example $\Sigma^{-c}R\oplus\Sigma^d R$ as a subobject.
  • When $c$ and $d$ have opposite parity, the forest subdioperad of $GCob_{c,d}$ has the same operations as the full PROP up to $\mathbb{Z}/2$ cokernel, which is what allows the infinite-dimensional Rabinowitz loop homology example to be described by the dioperad.
  • The signs in Theorem 5.1 coincide, up to a uniform regrading of the multiplication and comultiplication, with the biunital coFrobenius convention, so the two frameworks describe the same finite-dimensional structures.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • One extension the author leaves implicit is that the determinant-twisting calculus should transfer to other graph-indexed PROPs, such as the open-closed TQFTs with zippers sketched in Remark 6.9; forcing the zipper relations in a two-vertex-type dioperad is a concrete next step.
  • Because suspension shifts $(c,d)$ to $(c-1,d+1)$ while preserving $c+d$, I would expect any complete invariant of graded Frobenius algebras to respect this suspension class; the paper does not pursue such invariants.
  • The triviality result for odd $c-d$ in the naive graded setting suggests a useful diagnostic for the literature: any published non-trivial graded Frobenius structure with operations of odd relative degree must be hiding either signs or a suspension, and Theorem 5.1 gives the dictionary for finding them.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

1 major / 5 minor

Summary. The paper constructs PROPs GCob_{c,d}, fGCob_{c,d}, and pGCob_{c,d} from graphs with determinant twistings, defining (c,d)-graded (commutative/symmetric) Frobenius algebras as algebras over these PROPs. Theorem 5.1 then gives an equivalent description in terms of an object A with a multiplication µ of degree c, a unit η of degree −c, a comultiplication ν of degree d, and a counit ε of degree −d, subject to graded associativity, unitality, coassociativity, counitality, and a graded Frobenius relation with signs depending on c and d; in the symmetric cases, graded commutativity or graded symmetry is added. The paper also proves stability under suspension, establishing isomorphisms Σ⊗GCob_{c,d} ≅ GCob_{c−1,d+1} and the corresponding behavior of the maps, and it discusses examples from cohomology of manifolds, Hochschild homology, and loop homology.

Significance. If the main theorem is correct, the paper provides a useful unifying framework for graded Frobenius algebras: the PROP definition gives a canonical sign convention for arbitrary degrees (c,d), and Theorem 5.1 is the first systematic map-and-relations description in this setting. The signs are derived from determinant twistings and orientation choices, not fitted to examples, which gives the computation of Koszul signs in Lemma 7.6 a high degree of reliability. The suspension-stability result (Proposition 5.11) and the examples from manifold cohomology, Hochschild homology, and loop homology provide nontrivial external benchmarks. The paper is clearly written and the technical core is carried out in considerable detail. However, the sufficiency direction of Theorem 5.1 contains a significant gap that needs to be addressed before the result can be considered established.

major comments (1)
  1. [Section 7.2] The 'if' direction of Theorem 5.1 rests on the assertion: 'An analogous argument for (fat) graphs shows that any two decompositions into the four graphs are related by a sequence of those six relations.' This is not demonstrated. The cited normal-form theorems of Kock [Koc04] and Lauda–Pfeiffer [LP08] are for closed and open surfaces, not for the graph categories GCob_{c,d}, fGCob_{c,d}, and pGCob_{c,d}. Proposition 2.13 establishes an equivalence of 1-categories, but it does not by itself provide a presentation of the morphism categories by the four elementary graphs and the six relations. If the graph categories admit relations among decompositions beyond those generated by Lemma 7.6, the sufficiency direction fails and Theorem 5.1 would overcount graded Frobenius algebras. The subsequent π1-invariance argument inherits the same weakness: the reduction of arbitrary zig-zags to the relations in Lemma 7.6 depends on the same unproved normal-form assertion. A proof, or a precise reference showing that the graph decomposition relations are generated by the six surface relations, is needed.
minor comments (5)
  1. [Definition 4.3] The third displayed definition repeats 'fGCob_{c,d}' instead of 'pGCob_{c,d}' for the planar case.
  2. [Lemma 7.6, Graded unitality] The line beginning 'comp2(ωin() ⊗ ωin()) = −(e2 ∧ comp2(ωin() ⊗ ωin()) = ...' appears garbled; it should read 'comp2(ωin() ⊗ ωin()) = −(e2 ∧ e1 ∧ e0 ∧ e′2 ∧ e′1 ∧ e′0)−1 ...'.
  3. [Proposition 6.4] The two displayed isomorphisms in the statement are both labeled (10); the second should be renumbered.
  4. [Proposition 2.13] The faithfulness argument for fCob invokes [ES15, Theorem A] and says that 'restricting to the open part and taking π0' gives the claim; since this is a key step, a short explanation of how the cited theorem implies π0-level faithfulness of fCob would help the reader.
  5. [Throughout] The distinction between the 2-categories pGCob and the associated 1-categories pGCob is visually subtle in the typeset text; a bolder typographic distinction (e.g., different fonts) would reduce the chance of confusion.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; signs are computed from determinant twistings and cross-checked against external examples.

full rationale

The paper's derivation chain is not circular. The graded Frobenius PROP is constructed from graph cobordisms and determinant twistings (Definitions 2.12, 4.3, and 4.4), and no parameter is fitted to the examples used as benchmarks in Section 6. The signs in Theorem 5.1 are computed, not assumed: Lemma 7.6 evaluates compositions of the chosen orientation generators using the explicit formulas in Lemmas 7.3 and 7.4, while Remark 5.3 records that the signs depend on choices in det_{c,d}(G). The examples from manifold cohomology, Hochschild homology, and loop homology are external checks; the paper does not fit the sign relations to them. The only load-bearing gap is in Section 7.2, where the 'if' direction of Theorem 5.1 imports the surface normal-form theorem of Kock and Lauda-Pfeiffer and asserts an 'analogous argument' for graphs. That is a completeness/proof gap, not circularity: the needed normal-form statement is an external presentation theorem, not an assumption equivalent to the conclusion, and it is not a self-citation. Citations to [ES15], [Koc04], [LP08], [WW16], and [CO22b] are independent published results rather than self-referential support. Accordingly, no circular step of any of the seven enumerated kinds is present.

Assumptions & free parameters 0 free parameters · 3 assumptions · 1 invented entities

The central claim rests on standard results in cobordism theory and determinant linear algebra; no free numerical parameters are fitted. The main theorem's proof uses a normal-form assertion imported from the surface case, which is the least self-contained assumption.

assumptions (3)
  • standard math The graph cobordism categories GCob, fGCob and pGCob are equivalent to the geometric cobordism categories Cob_2^closed, Cob_2^open and Cob_2^planar.
    Used in Section 2 (Proposition 2.13, Corollary 2.15) to justify that functors out of these graph categories capture the usual TQFTs. The proof uses generation results of Kock [Koc04] and Lauda-Pfeiffer [LP08] and, for fat graphs, Egas Santander's theorem [ES15, Theorem A].
  • standard math The determinant functor det_{c,d} is well-defined and the colimits defining GCob_{c,d}(X,Y) can be computed via the π1-action on det_{c,d}(G) (Proposition 4.5).
    Relies on Lemma 4.1 (det of chain complexes) and the fact that graph morphisms induce isomorphisms on det. This is used throughout Section 4 and in the proof of Theorem 5.1 to identify morphism spaces as direct sums over graph classes.
  • domain assumption Any two decompositions of a graph cobordism into elementary graphs (⋔, η, ν, ε, twist) are related by the six generating relations.
    Imported from the surface case ([Koc04], [LP08]) by asserting 'an analogous argument for (fat) graphs' in Section 7.2. This is load-bearing for the if-direction of Theorem 5.1 and is not proven in the paper.
invented entities (1)
  • Graded graph cobordism PROP GCob_{c,d} (and fat/planar variants fGCob_{c,d}, pGCob_{c,d}) independent evidence
    purpose: Defines (c,d)-graded (commutative/symmetric) Frobenius algebras as algebras over this PROP; morphisms are generated by graphs twisted by det(G,∂in)^{⊗c}⊗det(G,∂out)^{⊗d}.
    The definition is grounded in external examples: manifold cohomology (Example 6.1), Hochschild homology (Example 6.3), loop space homology (Example 6.5), and Rabinowitz loop homology (Example 6.8). These examples are not used to fit the sign conventions; they are checked against the PROP after the fact. The categories are constructed from existing notions (graphs, determinant lines, cobordism categories) rather than postulated out of thin air.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Graded Frobenius Algebras." pith.science (2026). https://pith.science/paper/VVR6FFIB

@misc{pith2026241213909,
  author       = {Pith},
  title        = {Pith review of: Graded Frobenius Algebras},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VVR6FFIB}},
  note         = {Machine review of arXiv:2412.13909}
}
read the original abstract

We construct a PROP which encodes 2D-TQFTs with a grading. This defines a graded Frobenius algebra as algebras over this PROP. We also give a description of graded Frobenius algebras in terms of maps and relations. This structure naturally arises as the cohomology of manifolds, loop homology and Hochschild homology of Frobenius algebras. In addition, we give a comprehensive description of the signs that arise in suspending algebras over PROPs.

Figures

Figures reproduced from arXiv: 2412.13909 by the authors.

Figure 1
Figure 1. An example of a closed cobordism from S 1 ⊔ S 1 ⊔ S 1 to S 1 ⊔ S 1 . We use the convention of drawing the incoming boundary to the left and the outgoing boundary to the right. Definition 2.2. The category of open 2D-cobordisms Cobopen 2 has objects Ob(Cobopen 2 ) = (a X [0, 1] [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 2
Figure 2. An example of an open cobordism from [0, 1] to ∅. The two surfaces are homeomorphic as they both have one boundary and Euler characteristic −1. We fix some orientation on [0, 1] × R. Definition 2.3. The category of planar 2D-cobordisms Cobplanar 2 has objects Ob(Cobplanar 2 ) = (a X [0, 1] [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. A planar cobordism between [0, 1] and ∅. 2.2. Graphs. We later give a description of those three categories using graphs. For this, we introduce what we mean by a graph. Definition 2.4. A graph G = (V, H, σ, s, Lin, Lout) consists of the following data: • a finite set V (G) := V called the vertices; • a finite set H(G) := H called the half-edges; • an involution σ : H → H with no fixed points, the orbits {h, σh} are… view at source ↗
Figures from the paper (11 more)
Figure 4
Figure 4. Figure 4: A morphism of graphs sends Lin to Lin and Lout to Lout and possibly collapses edges. Here the graph morphism collapses two edges. Definition 2.8. A morphism of fat graphs is a morphism of graphs f : G → G′ that can be written as a sequence of edge collapses of edges G …
Figure 5
Figure 5. Figure 5: Composition of a graph cobordism between {a, b, c} and {x} and a graph cobordism between {x} and {y}. We note that this defines a functor as morphisms in GCob(X, Y ) and GCob(Y, Z) can be glued together in a similar manner. The unit idX ∈ GCob(X, X) is given by the gra…
Figure 6
Figure 6. Figure 6: The closed cobordism Cob(G) corresponding to a graph G. Similarly we can construct for a fat graph cobordism between X and Y an open cobordism fCob(G) between ` X[0, 1] and ` Y [0, 1]. It is given by the following construction. Vertices v of arity 0 contribute a copy o…
Figure 7
Figure 7. Figure 7: The open cobordism fCob(G) corresponding to a fat graph G [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]
Figure 8
Figure 8. Figure 8: Embedding fCob(G) in grey as the bottom half of Cob(G) [PITH_FULL_IMAGE:figures/full_fig_p036_8.png]
Figure 9
Figure 9. Figure 9: The graph on the bottom can be lifted to a graph which can be decom￾posed into the four elementary graphs. This graph can for example be written as ( ⊗ ) ◦ ( ⊗ id ⊗ id) ◦ (id ⊗ ⊗ id) ◦ ( ⊗ id) ◦ ( ⊗ ) ◦ ◦ ◦ ( ⊗ id). (Note the order-reversing convention for the composit…
Figure 11
Figure 11. Figure 11: The graph with v the lone internal ver￾tex, v0 the lone incoming vertex and v1 and v2 the out￾going vertices. The vertex v1 corresponding to the first coordinate of the output and v2 to the second coordinate [PITH_FULL_IMAGE:figures/full_fig_p046_11.png]
Figure 13
Figure 13. Figure 13: The graph with v the lone internal ver￾tex and v0 the lone incoming vertex. Using the identification as in Proposition 7.1, we define generators ωin( ) := (e2 ∧ e1 ∧ e0) −1 ⊗ h0 ∧ σh0 ∧ h1 ∧ σh1 ∧ h2 ∧ σh2 ⊗ v ∧ v0 ∈ det( , ∂in), ωout( ) := 1 ∈ det( , ∂out) ∼= 1, ωin(…
Figure 14
Figure 14. Figure 14: The two composition of two multiplications give the same graph after collapsing two edges. As we glued v ′ 0 to v1, we can collapse e ′ 0 and e1 to obtain the graph (see top of [PITH_FULL_IMAGE:figures/full_fig_p050_14.png]
Figure 14
Figure 14. Figure 14: This differs from the naming scheme of the other composition by an even permutation. [PITH_FULL_IMAGE:figures/full_fig_p051_14.png]
Figure 15
Figure 15. Figure 15: The two composition of a unit and a multiplications give the graph with one vertex representing id after collapsing four edges. Now we can collapse e0 and e ′ 1 as in the top of [PITH_FULL_IMAGE:figures/full_fig_p052_15.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

21 extracted references · 16 canonical work pages

  1. [1]

    Two-dimensional topological quantum field theories and F robenius algebras

    Lowell Abrams. Two-dimensional topological quantum field theories and F robenius algebras. Journal of Knot theory and its ramifications , 5(05):569--587, 1996

  2. [2]

    A polarized view of string topology

    Ralph L Cohen and V \'e ronique Godin. A polarized view of string topology. London Mathematical Society Lecture Note Series , 308:127, 2004

  3. [3]

    La th\'eorie de l'homotopie des 2-cat\'egories

    Jonathan Chiche. La th\'eorie de l'homotopie des 2-cat\'egories . arXiv preprint arXiv:1411.6936 , 2014

  4. [4]

    Poincar\' e duality for loop spaces

    Kai Cieliebak, Nancy Hingston, and Alexandru Oancea. Poincar\' e duality for loop spaces. arXiv preprint arXiv:2008.13161 , 2020

  5. [5]

    Loop coproduct in M orse and F loer homology

    Kai Cieliebak, Nancy Hingston, and Alexandru Oancea. Loop coproduct in M orse and F loer homology. Journal of Fixed Point Theory and Applications , 25(2):59, 2023

  6. [6]

    Reduced symplectic homology and string topology

    Kai Cieliebak and Alexandru Oancea. Reduced symplectic homology and the secondary continuation map. arXiv preprint arXiv:2209.02972 , 2022

  7. [7]

    Topological F robenius A lgebras

    Kai Cieliebak and Alexandru Oancea. Topological F robenius A lgebras. arXiv preprint arXiv:2209.02979 , 2022

  8. [8]

    Rabinowitz Floer homology as a Tate vector space

    Kai Cieliebak and Alexandru Oancea. Rabinowitz Floer homology as a Tate vector space . arXiv preprint arXiv:2407.21741 , 2024

Show all 21 references
  1. [9]

    Comparing fat graph models of moduli space

    Daniela Egas Santander. Comparing fat graph models of moduli space. arXiv preprint arXiv:1508.03433 , 2015

  2. [10]

    Koszul duality for dioperads

    Wee Liang Gan. Koszul duality for dioperads. arXiv preprint math/0201074 , 2002

  3. [11]

    Loop products and closed geodesics

    Mark Goresky and Nancy Hingston. Loop products and closed geodesics. Duke Mathematical Journal , 150(1):117, 2009

  4. [12]

    Modular operads

    Ezra Getzler and Mikhail M Kapranov. Modular operads. Compositio Mathematica , 110(1):65--125, 1998

  5. [13]

    Higher string topology operations

    V \'e ronique Godin. Higher string topology operations. arXiv preprint arXiv:0711.4859 , 2007

  6. [14]

    Algebraic Topology

    Allen Hatcher. Algebraic Topology . Cambridge University Press, 2002

  7. [15]

    Product and coproduct in string topology

    Nancy Hingston and Nathalie Wahl. Product and coproduct in string topology . arXiv preprint arXiv:1709.06839 , 2017

  8. [16]

    Frobenius algebras and 2-d topological quantum field theories

    Joachim Kock. Frobenius algebras and 2-d topological quantum field theories . Number 59 in London Mathematical Society Student Texts. Cambridge University Press, 2004

  9. [17]

    Pre-Calabi-Yau algebras and topological quantum field theories

    Maxim Kontsevich, Alex Takeda, and Yiannis Vlassopoulos. Pre-Calabi-Yau algebras and topological quantum field theories . arXiv preprint arXiv:2112.14667 , 2021

  10. [18]

    Open--closed strings: T wo-dimensional extended TQFT s and Frobenius algebras

    Aaron D Lauda and Hendryk Pfeiffer. Open--closed strings: T wo-dimensional extended TQFT s and Frobenius algebras . Topology and its Applications , 155(7):623--666, 2008

  11. [19]

    Operads and PROPs

    Martin Markl. Operads and PROPs . Handbook of algebra , 5:87--140, 2008

  12. [20]

    Loop coproducts in string topology and triviality of higher genus TQFT operations

    Hirotaka Tamanoi. Loop coproducts in string topology and triviality of higher genus TQFT operations. Journal of Pure and Applied Algebra , 214(5):605--615, 2010

  13. [21]

    Hochschild homology of structured algebras

    Nathalie Wahl and Craig Westerland. Hochschild homology of structured algebras. Advances in Mathematics , 288:240--307, 2016

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.