Key Takeaways
- K-Dense Scientific Skills provides 170+ skills across 17 scientific domains for Claude Code and compatible agents
- Domains include RNA-seq, scRNA-seq, spatial transcriptomics, drug discovery, clinical research, and more
- One-command install: copy skill files to
.claude/skills/directory - Used by MedgeClaw (140 skills) as its default skill library
- Complementary to Orchestra Research's ML-focused skills — K-Dense covers domain science
What Are K-Dense Scientific Skills?
K-Dense Scientific Skills is a curated library of 170+ skills across 17 scientific domains, designed for Claude Code and any agent supporting the SKILL.md standard. Each skill encodes domain expertise — from proper RNA-seq normalization to clinical trial protocol design.
Core attributes:
- Skills count: 170+ across 17 domains
- Target agent: Claude Code (also compatible with Codex, Gemini)
- Format: Standard
SKILL.mdMarkdown files - Notable user: MedgeClaw (uses 140 K-Dense skills as default)
The 17 Domains
| Domain | Example Skills |
|---|---|
| RNA-seq | Differential expression, normalization, batch correction |
| scRNA-seq | Clustering, trajectory analysis, cell type annotation |
| Spatial Transcriptomics | Spot deconvolution, spatial pattern analysis |
| Genomics | Variant calling, WGS/WES analysis pipelines |
| Proteomics | Mass spec analysis, protein quantification |
| Drug Discovery | Virtual screening, ADMET prediction, docking |
| Clinical Research | Trial design, endpoint analysis, regulatory |
| Epidemiology | Outbreak modeling, surveillance data analysis |
| Biostatistics | Survival analysis, mixed models, meta-analysis |
| Imaging | Medical image analysis, histopathology |
| Ecology | Species distribution, biodiversity analysis |
| Chemistry | Molecular property prediction, reaction planning |
| Materials Science | Crystal structure, property prediction |
| Neuroscience | EEG/fMRI analysis, neural data processing |
| Agriculture | Crop genomics, phenotype analysis |
| Bioinformatics Core | Sequence alignment, phylogenetics, annotation |
| Data Management | FAIR data, metadata standards, reproducibility |
Quick Start
Install All Skills
git clone https://github.com/k-dense/scientific-skills.git
cp -r scientific-skills/skills/* ~/.claude/skills/Install Specific Domains Only
# RNA-seq skills only
cp -r scientific-skills/skills/rna-seq/* ~/.claude/skills/
# scRNA-seq + spatial transcriptomics
cp -r scientific-skills/skills/scrna-seq/* ~/.claude/skills/
cp -r scientific-skills/skills/spatial-transcriptomics/* ~/.claude/skills/Verify Installation
ls ~/.claude/skills/ | wc -l
# Should show 170+K-Dense vs Orchestra Research Skills
| Aspect | K-Dense Scientific Skills | Orchestra AI Research Skills |
|---|---|---|
| Focus | Domain-specific science (biology, chemistry, clinical) | ML/AI research (training, benchmarking, paper writing) |
| Skills count | 170+ across 17 domains | 86 across 22 categories |
| Best for | Bioinformaticians, wet lab researchers, clinical scientists | ML researchers, PhD students, research engineers |
| Overlap | Minimal — domain science focus | Minimal — ML methodology focus |
| Combined use | ✓ Recommended — they're complementary | ✓ Recommended — they're complementary |
FAQ
Q1: Do these skills work with agents other than Claude Code?
Yes. Any agent supporting SKILL.md format — including Codex, Gemini, and custom agents — can use these skills. They are plain Markdown files.
Q2: How do K-Dense skills differ from asking Claude directly?
Skills encode domain-expert knowledge — proper normalization methods, statistical test selection, tool parameters. Without skills, Claude may hallucinate parameters or use outdated methods. Skills ensure best-practice workflows.
Q3: Can I contribute new skills?
Yes. Follow the skill template format and submit a PR. Skills undergo domain-expert review before merging.
Q4: Which projects use K-Dense skills?
MedgeClaw (140 skills) is the largest user. Several research labs have adopted them for internal bioinformatics workflows.
Summary
K-Dense Scientific Skills is the most comprehensive domain-science skill library for AI coding agents. With 170+ skills across 17 scientific domains, it covers the breadth of modern research — from RNA-seq to clinical trials. Install it alongside Orchestra Research Skills for complete coverage of both domain science and ML methodology.