umap-learn
UMAP dimensionality reduction. Fast nonlinear manifold learning for 2D/3D visualization, clustering preprocessing (HDBSCAN), supervised/parametric UMAP, for high-dimensional data.
80
66%
Does it follow best practices?
Impact
89%
2.11xAverage score across 6 eval scenarios
Passed
No known issues
Optimize this skill with Tessl
npx tessl skill review --optimize ./scientific-skills/umap-learn/SKILL.mdQuality
Discovery
82%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This is a strong, technically specific description that clearly identifies the skill's domain and capabilities with excellent trigger terms for its target audience. The main weakness is the absence of an explicit 'Use when...' clause, which would help Claude know exactly when to select this skill. The description is concise and avoids fluff while packing in relevant technical keywords.
Suggestions
Add an explicit 'Use when...' clause, e.g., 'Use when the user asks about UMAP, dimensionality reduction, embedding high-dimensional data, or visualizing clusters.'
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Lists multiple specific concrete actions: dimensionality reduction, 2D/3D visualization, clustering preprocessing with HDBSCAN, supervised/parametric UMAP. These are concrete, well-defined capabilities. | 3 / 3 |
Completeness | Clearly answers 'what does this do' with specific capabilities, but lacks an explicit 'Use when...' clause or equivalent trigger guidance. Per the rubric, a missing 'Use when...' clause caps completeness at 2. | 2 / 3 |
Trigger Term Quality | Includes strong natural keywords users would say: 'UMAP', 'dimensionality reduction', 'manifold learning', '2D/3D visualization', 'clustering', 'HDBSCAN', 'high-dimensional data', 'supervised', 'parametric'. Good coverage of terms a data scientist would use. | 3 / 3 |
Distinctiveness Conflict Risk | UMAP is a very specific algorithm/technique with distinct terminology (HDBSCAN, manifold learning, parametric UMAP). This is unlikely to conflict with other skills given its narrow, well-defined niche. | 3 / 3 |
Total | 11 / 12 Passed |
Implementation
50%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The skill is highly actionable with excellent, executable code examples covering a wide range of UMAP use cases. However, it is far too verbose for a skill file — it reads more like comprehensive library documentation than a concise instruction set for Claude, explaining concepts Claude already understands and including extensive parameter descriptions that could be in a reference file. The workflow clarity is decent but lacks validation checkpoints.
Suggestions
Reduce content by 50-60%: move the Parameter Tuning Guide, Advanced Features, and Common Issues sections into separate referenced files, keeping only a quick-start and the most critical parameter recommendations inline.
Remove explanatory text Claude already knows (e.g., 'UMAP is a dimensionality reduction technique', 'Unlike t-SNE, UMAP scales well', descriptions of what metrics are) and keep only the actionable guidance.
Add explicit validation checkpoints to workflows — e.g., after UMAP embedding, check embedding shape and inspect a sample scatter plot before proceeding to clustering.
Consolidate redundant parameter combination examples that appear in multiple sections into a single reference table or cheat sheet.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The content is excessively verbose at ~300+ lines. It explains concepts Claude already knows (what UMAP is, what dimensionality reduction means, how scikit-learn conventions work), includes lengthy parameter descriptions that read like documentation rather than actionable instructions, and has significant redundancy (e.g., the same parameter combinations repeated across multiple sections). | 1 / 3 |
Actionability | The skill provides fully executable, copy-paste ready code examples throughout — from basic usage to clustering workflows, ML pipelines, parametric UMAP, and inverse transforms. All code blocks are complete with imports and realistic parameter choices. | 3 / 3 |
Workflow Clarity | Workflows are clearly sequenced with numbered steps (e.g., the clustering workflow and typical workflow), but there are no explicit validation checkpoints or error-recovery feedback loops. For example, the clustering workflow lacks a step to verify embedding quality before proceeding to HDBSCAN, and there's no guidance on what to do if results look wrong mid-workflow beyond the troubleshooting section at the end. | 2 / 3 |
Progressive Disclosure | The content references a `references/api_reference.md` file for detailed API docs, which is good progressive disclosure. However, the main SKILL.md itself is monolithic — the parameter tuning guide, advanced features (Parametric UMAP, AlignedUMAP, inverse transforms), and troubleshooting could all be split into separate referenced files. No bundle files were provided to verify the reference exists. | 2 / 3 |
Total | 8 / 12 Passed |
Validation
90%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 10 / 11 Passed
Validation for skill structure
| Criteria | Description | Result |
|---|---|---|
metadata_version | 'metadata.version' is missing | Warning |
Total | 10 / 11 Passed | |
- Repository
- K-Dense-AI/claude-scientific-skills
- Commit
cbcae7b
- Reviewed
Table of Contents
Is this your skill?
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.