umap-learn — Claude Code skill: what it does, how to install
This skill guides UMAP-based dimensionality reduction, 2D/3D embeddings, and clustering preparation. Use it when an agent needs supervised or semi-supervised UMAP workflows.
npx skills add K-Dense-AI/scientific-agent-skills --skill umap-learn -a claude-codeWhat umap-learn does
The skill covers umap-learn for nonlinear dimensionality reduction, 2D/3D embeddings, and feature preparation before clustering. It guides parameter choices for n_neighbors, min_dist, n_components, and metric, with separate fit and transform steps through the scikit-learn estimator interface. Label-aware workflows include supervised and semi-supervised UMAP, DensMAP, AlignedUMAP, and Parametric UMAP.
Examples use StandardScaler and Matplotlib alongside UMAP. Python 3.9+ and umap-learn are required; HDBSCAN, Matplotlib, and TensorFlow with Keras 3 are optional for specific workflows. Network access is needed to install the package, while local fitting does not need it.
Choose the embedding size according to the task: 2–3 dimensions are used for visualization, while 5–10 can support clustering preprocessing. UMAP axes, island areas, and gaps between groups do not carry calibrated physical or probabilistic meaning. For predictive work, fit preprocessing and embeddings on training data only.
What it helps with
- build 2D and 3D embeddings from feature data
- prepare higher-dimensional embeddings for clustering
- tune
n_neighbors,min_dist, andmetric - guide embeddings with supervised labels
- compare local and broader neighborhood structure
Example prompts for the agent
- Create a 2D UMAP embedding for
data, choose suitable parameters, and plot the result. - Prepare these features for clustering with 10 UMAP components and explain the metric choice.
- Run supervised UMAP with
labelswhile keeping test data outside the fit.
How to install umap-learn
1Skills CLI (Vercel)
npx skills add K-Dense-AI/scientific-agent-skills --skill umap-learn -a claude-code2Manually: put the files in place
~/.claude/skills/umap-learn/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/umap-learn/
1Skills CLI (Vercel)
npx skills add K-Dense-AI/scientific-agent-skills --skill umap-learn -a codex2Manually: put the files in place
.agents/skills/umap-learn/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add K-Dense-AI/scientific-agent-skills --skill umap-learn -a cursor2Manually: put the files in place
.cursor/skills/umap-learn/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install K-Dense-AI/scientific-agent-skills umap-learn2Manually: put the files in place
.github/skills/umap-learn/SKILL.md1Skills CLI (Vercel)
npx skills add K-Dense-AI/scientific-agent-skills --skill umap-learn -a gemini-cli2Platform CLI
gemini skills install https://github.com/K-Dense-AI/scientific-agent-skills --path skills/umap-learnManually: put the files in place
.devin/skills/umap-learn/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add K-Dense-AI/scientific-agent-skills --skill umap-learn -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about umap-learn
How do I install umap-learn in Claude Code?
Run npx skills add K-Dense-AI/scientific-agent-skills --skill umap-learn -a claude-code, then check ~/.claude/skills/umap-learn/SKILL.md.
What does the umap-learn skill do?
It helps apply UMAP for nonlinear dimensionality reduction, 2D/3D visualization, clustering preparation, and label-aware embeddings. It also covers DensMAP, AlignedUMAP, and Parametric UMAP.
What does umap-learn require?
It requires Python 3.9+ and umap-learn. HDBSCAN, Matplotlib, and TensorFlow with Keras 3 may be needed for specific examples, and installation needs network access.
