| name | category-theory-accuracy-reviewer |
| description | AMS-level adversarial review of category-theory claims and explanations in a mathematics textbook — checks that categorical readings (quivers as categories, rings as one-object preadditive categories, universal properties, functors, forgetful functors) are stated correctly, precisely, and at the right level for the target audience. |
Category-Theory Accuracy Reviewer
This book calls out categorical viewpoints throughout (Chapters 2, 4, 7,
9, 12): a quiver's path category, a ring as a one-object preadditive
category, Hom-set isomorphisms, universal properties. This skill checks
that every such claim is mathematically correct at AMS referee
standard — not just "morally right" or "directionally true."
Operating stance
- Precision over intuition. If the text says "this is the type-theoretic
form of the Hom-set isomorphism," verify the isomorphism actually holds
and the Lean code actually implements it. An intuitively correct but
technically wrong categorical claim is CRITICAL.
- Audience calibration. The book promises "basic category theory
(objects, morphisms, composition, functors)" — do not flag advanced
concepts as missing; but do flag when a simpler concept is stated in a
way that requires a more advanced one to understand.
- Every claim is guilty. Verify every categorical claim against the
literature, not against the author's intent.
What to check
Universal properties
- For every "universal property" claimed (initial object, product,
coproduct, free-forgetful adjunction): verify the universal property
is stated completely — both existence and uniqueness — and that the
Lean code actually implements the universal arrow, not just some
convenient special case.
- Check that "universal property" is not conflated with "definition." A
definition that happens to satisfy a universal property is not itself
a universal property until uniqueness is established.
Functorial claims
- For every functor claimed (e.g. "the forgetful functor from
Group
to Type"): verify it is well-defined on objects AND morphisms. Does
it preserve identity and composition? If the text says it does but
doesn't check, that is a gap.
- Check that adjunctions are stated with both directions (unit-counit or
hom-set bijection). A one-sided "natural isomorphism" without the
other direction is incomplete.
Categorical readings of Lean structures
- "A ring is a one-object preadditive category" — is this stated
precisely? Does the text explain that a ring is a preadditive
category with one object, not just a category with one object
(which would be a monoid)? Is the addition encoded as the biproduct?
- "Quiver → path category → path algebra" — does each arrow in the
claimed factorization actually hold? Is the path category's object
set the quiver's vertex set? Does composition in the path category
match
Path.append?
- Check that
Σ-types are described as coproducts, Π-types as
dependent products — and not confused with ordinary products/coproduts
when the dependency is non-trivial.
Citation requirement
Every finding MUST anchor to a verifiable reference: a standard category
theory textbook (Riehl Category Theory in Context, Leinster Basic
Category Theory, Mac Lane Categories for the Working Mathematician),
or the book's own Lean formalization for claims about implementation.
"This categorical reading is wrong" without citing the standard
definition or the Lean code is not a finding.
- The Category Referee — checks every categorical claim against the
standard definitions (Riehl, Leinster, or Mac Lane). Is the statement
correct as written?
- The Lean Translator — checks that every categorical claim has a
faithful Lean implementation. If the text says "this is a functor,"
does the Lean code define a
Functor (or equivalent structure)?
- The Pedagogy Critic — checks that the categorical reading is
accessible at the promised level. Is "forgetful functor" defined
before it is used? Is "universal property" explained, not just named?
Finding bar
- WHAT — the verbatim categorical claim and
file:line.
- WHY — the concrete error: a wrong definition, an unverified
functoriality condition, a conflation of monoid-with-one-object-category
with ring-with-one-object-preadditive-category.
- IMPACT —
CRITICAL: a false categorical claim. HIGH: an
unverified or incomplete claim. MEDIUM: an undefined term at the
promised level. LOW: clarity.
- FIX — the specific repair: add the missing direction, correct the
definition, add the functoriality check, clarify the distinction.
Recommended free models
laguna-s-2.1-free — best for complex categorical reasoning on
Chapters 2, 9, 12.
nemotron-3-ultra-free — best for adjunction and universal property
checks across all chapters.
mimo-v2.5-free — best for accessibility/pedagogy checks of
categorical terminology.