Lean, table-and-code driven content that assumes Claude's competence; avoids explaining basic testing concepts and every token earns its place via concrete counts, tool tables, and a YAML spec.

Provides an executable YAML spec, a concrete tool comparison with named CLIs/libraries (PICT, AllPairs, CATS), and copy-paste-ready parametrize snippets in the linked reference file.

A clear 7-step sequence with an explicit validation checkpoint in Step 5 (review the coverage report, accept constraint-driven gaps or raise strength), forming a real feedback loop.

SKILL.md is an overview with a single, clearly signaled one-level-deep reference (references/output-formats.md), which exists and appropriately offloads the per-runner format snippets.

Lists multiple concrete actions: generating parameterized test inputs, combining boundary-value/equivalence-class/pairwise cases, all-pairs reduction of the cross-product, and emitting in runner-native parametrize format.

Clearly answers both what (generate/reduce/emit parametrize cases) and when via an explicit 'Use when a function or endpoint takes 3+ inputs whose interactions matter and full Cartesian product would explode' trigger.

Uses natural trigger phrasing users would say: 'function or endpoint takes 3+ inputs whose interactions matter' and 'full Cartesian product would explode', alongside 'parameterized test inputs' and 'pairwise'.

Occupies a clear niche (combinatorial pairwise test generation for 3+ interacting inputs) with distinct triggers that are unlikely to collide with sibling skills like boundary-value-generator.