# Expert review guide

The first priority is the main interface theorem, not the packaging or coined release name.

1. Check the row-gradient convention, the Hadamard rank-one reduction, and the determinant identity. Challenge the support-size-one and support-size-two exceptional cases as well as the generic case. Search rank-one compatibility and differential-inclusion literature for an equivalent classification.
2. Check the quantitative inequality's normal margin and orientation sign argument. The bound is intentionally local and asymmetric: A is exactly orthogonal; B can have shear. Do not read it as an unproved global near-orthogonal rigidity theorem.
3. Check that scale-gain balance and primal coordinate exactness are distinct checks. For holed or invalid input complexes, no undeclared simply connectedness or embedding certificate should be inferred.
4. Check that C is primitive and that chart origins are real. Inspect the integer right inverse and the fake-torsion counterexample. Check the coupled H*alpha+p range before accepting any rounded period.
5. Check all boundary, singular-extension, and extraction exclusions. No claim of a universal hex mesher should survive into a public abstract or submission letter.

## Priority and positioning

The expected relevant literature includes MINT, CubeCover, Odeco integrability, Beben's frame topology, Lipman's simplicial global inversion, and gain-graph balance. The literature search used primary papers, author pages, and official project documentation. It was targeted, not exhaustive. Absence of an exact phrase in a web search does not establish priority.

## Submission readiness

The archive is suitable for independent technical review. Before a journal submission, the human submitter should verify authorship policy, licensing, publication ethics, and the claimed novelty. No acceptance, review, DOI, or public deposit is represented as already obtained.