Type safety, test validity, naming conventions, dead code, documentation accuracy, and general maintainability of code that exists and resolves.
The correctness-of-existing-code layer: is the code that does exist and does resolve also maintainable, type-safe, and honestly tested?
In scope. Type safety (any-type leakage, unchecked casts, @ts-ignore without justification), test validity (assertion-free tests, over-mocked suites, coverage theater), naming conventions and consistency, dead code and unused exports, documentation accuracy (docs vs. behavior drift), cyclomatic complexity and duplication, maintainability hygiene.
Not in scope. Plural tooling doing the same job (that's dependency-coherence). References that don't resolve (that's reference-integrity). Stubs or scaffolding leaked into production (that's placeholder-hygiene).
Distinct because. Concerns correctness and maintainability of code that exists and resolves. The neighboring taxons carve off specific defect families (reference failure, plural tooling, scaffold leakage) that would otherwise overwhelm this taxon. A pattern about "test file has no expect() calls" is code-quality. A pattern about "import points at module that doesn't exist" is reference-integrity.
Conceptual sub-structure. Type safety, test validity, naming, dead code, documentation accuracy, maintainability. Formal sub-taxon paths likely warranted once catalog passes ~3,000 patterns in this taxon.