umap-learn

Quick Review

Excellent skill documentation for UMAP dimensionality reduction. The description clearly conveys the skill's capabilities (fast nonlinear manifold learning, visualization, clustering preprocessing, supervised variants), enabling straightforward invocation. Task knowledge is comprehensive with detailed parameter tuning guidance, complete workflows for multiple use cases (visualization, clustering, supervised learning, pipelines), and practical code examples. Structure is logical with clear sections progressing from basics to advanced features, though slightly lengthy for a single file. Novelty is strong—UMAP configuration is nuanced (parameter interactions, clustering vs visualization settings, metric selection) and would consume many tokens for a CLI agent to discover through trial. The preprocessing requirements, supervised/parametric variants, and clustering integration represent complex domain knowledge that meaningfully reduces cost. Minor improvement possible by moving some advanced sections to separate reference files, but overall this is a high-quality, production-ready skill.