Claude Scientific Skills: The Complete Guide to AI-Powered Scientific Research
What if your AI coding agent could instantly become a bioinformatics expert, a drug discovery chemist, a clinical research analyst, and a materials scientist — all at the same time? That's what Claude Scientific Skills delivers. With 13,400+ GitHub stars and 170+ pre-built scientific skills, it transforms any compatible AI agent into an "AI Scientist" capable of executing complex multi-step scientific workflows across biology, chemistry, medicine, physics, and beyond.
Claude Scientific Skills on GitHub
What Is Claude Scientific Skills?
Claude Scientific Skills is a comprehensive, open-source collection of 170+ ready-to-use Agent Skills for scientific research, engineering, analysis, finance, and writing. Created by K-Dense AI, these skills follow the open Agent Skills standard and work with:
- Claude Code
- Cursor
- OpenAI Codex
- Any agent supporting the Agent Skills standard
Each skill includes comprehensive documentation (SKILL.md), practical code examples, use cases, best practices, and integration guides.
What's Included
The numbers are impressive:
- 170+ Scientific Skills across 17 domains
- 250+ Scientific & Financial Databases — PubMed, ChEMBL, UniProt, COSMIC, ClinicalTrials.gov, SEC EDGAR, Alpha Vantage, and more
- 60+ Optimized Python Package Skills — RDKit, Scanpy, PyTorch Lightning, scikit-learn, BioPython, PennyLane, Qiskit, OpenMM, scVelo, TimesFM
- 15+ Scientific Integration Skills — Benchling, DNAnexus, LatchBio, OMERO, Protocols.io
- 35+ Analysis & Communication Tools — Literature review, scientific writing, peer review, posters, slides, infographics
- 10+ Research & Clinical Tools — Hypothesis generation, grant writing, clinical decision support
Scientific Domains Covered
🧬 Bioinformatics & Genomics (20+ skills)
Sequence analysis with BioPython, single-cell RNA-seq with Scanpy, RNA velocity with scVelo, gene regulatory networks with Arboreto, variant annotation with Ensembl VEP, phylogenetic analysis with ETE Toolkit.
🧪 Cheminformatics & Drug Discovery (13+ skills)
Molecular property prediction with RDKit, virtual screening with DiffDock, ADMET analysis with DeepChem, molecular dynamics with OpenMM + MDAnalysis, cloud quantum chemistry with Rowan, drug-target binding with BindingDB.
🏥 Clinical Research & Precision Medicine (16+ skills)
Clinical trials via ClinicalTrials.gov, cancer genomics via cBioPortal, disease-gene associations via Monarch Initiative, variant interpretation with ClinVar/COSMIC/ClinPGx, healthcare AI with PyHealth.
🔬 Proteomics & Mass Spectrometry
LC-MS/MS processing with matchms, spectral analysis with pyOpenMS.
🖼️ Medical Imaging & Digital Pathology
DICOM processing with pydicom, whole slide imaging with histolab and PathML, NCI Imaging Data Commons.
🤖 Machine Learning & AI (16+ skills)
Deep learning with PyTorch Lightning & Transformers, reinforcement learning with Stable Baselines3, time series forecasting with TimesFM, Bayesian methods with PyMC, graph ML with Torch Geometric.
🔮 Materials Science, Chemistry & Physics (7 skills)
Crystal structure analysis with Pymatgen, metabolic modeling with COBRApy, astronomy with Astropy, quantum computing with Cirq/PennyLane/Qiskit/QuTiP.
⚙️ Engineering & Simulation (4 skills)
Discrete-event simulation with SimPy, computational fluid dynamics with FluidSim, numerical computing with MATLAB/Octave.
📊 Data Analysis & Visualization (17+ skills)
Matplotlib, Seaborn, Plotly, GeoPandas, NetworkX, SymPy, EDA workflows, statistical analysis, publication-quality figures.
🌍 Geospatial Science & Remote Sensing
Satellite imagery, GIS analysis, spatial statistics with GeoMaster (500+ examples).
🧪 Laboratory Automation (4 skills)
Liquid handling with PyLabRobot, cloud lab with Ginkgo, protocol management with Protocols.io, LIMS with Benchling.
🔬 Multi-Omics & Systems Biology (5+ skills)
Pathway analysis with KEGG/Reactome/STRING, multi-modal data integration with Denario.
🧬 Protein Engineering & Design (3 skills)
Protein language models with ESM, glycoengineering, cloud lab platforms with Adaptyv.
📚 Scientific Communication (24+ skills)
Literature review with OpenAlex/PubMed/bioRxiv, peer review, scientific writing, citation management with Zotero, posters, slides, schematics, infographics, Mermaid diagrams.
🎓 Research Methodology
Hypothesis generation, scientific brainstorming, critical thinking, grant writing, scholar evaluation.
Quick Examples
Drug Discovery Pipeline
Query ChEMBL for EGFR inhibitors (IC50 < 50nM), analyze structure-activity
relationships with RDKit, generate improved analogs with datamol, perform
virtual screening with DiffDock against AlphaFold EGFR structure, search
PubMed for resistance mechanisms, check COSMIC for mutations.
Skills Used: ChEMBL, RDKit, datamol, DiffDock, AlphaFold DB, PubMed, COSMIC
Single-Cell RNA-seq Analysis
Load 10X dataset with Scanpy, perform QC and doublet removal, integrate
with Cellxgene Census data, identify cell types using NCBI Gene markers,
run differential expression with PyDESeq2, infer gene regulatory networks
with Arboreto, enrich pathways via Reactome/KEGG.
Skills Used: Scanpy, Cellxgene Census, NCBI Gene, PyDESeq2, Arboreto, Reactome, KEGG
Multi-Omics Biomarker Discovery
Analyze RNA-seq with PyDESeq2, process mass spec with pyOpenMS, integrate
metabolites from HMDB, map proteins to pathways (UniProt/KEGG), find
interactions via STRING, build predictive model with scikit-learn,
search ClinicalTrials.gov for relevant trials.
Skills Used: PyDESeq2, pyOpenMS, HMDB, UniProt, KEGG, STRING, scikit-learn, ClinicalTrials.gov
Getting Started
Step 1: Clone the Repository
git clone https://github.com/K-Dense-AI/claude-scientific-skills.git
Step 2: Copy Skills to Your Agent's Skills Directory
Copy the skills you need to your agent's skills directory. Each skill is self-contained with its own SKILL.md, examples, and documentation.
Prerequisites
- Python 3.x
uvpackage manager (recommended)- A compatible AI agent (Claude Code, Cursor, Codex, etc.)
The agent automatically discovers and uses relevant skills when you ask it to perform a task in a supported domain.
Claude Scientific Skills vs Alternatives
Category: This repo is a collection of pre-built AI agent skills for scientific research and computing.
| Feature | Claude Scientific Skills | Anthropic Skills | AI-Scientist |
|---|---|---|---|
| Focus | 170+ scientific research skills | General-purpose agent skills | Automated scientific discovery |
| Stars | 13.5K ⭐ | 86K ⭐ | 12.3K ⭐ |
| License | MIT | No license specified | Custom |
| Scientific Domains | 17 domains (bio, chem, clinical, physics, materials, etc.) | General (web, data, coding) | ML/NeurIPS-style research |
| Database Access | 250+ scientific & financial databases | Limited | Custom experiment DBs |
| Python Package Skills | 60+ optimized skills (RDKit, Scanpy, PyTorch Lightning...) | General tools | Research-specific |
| Drug Discovery | ✅ Full pipeline (ChEMBL, DiffDock, ADMET, molecular dynamics) | ❌ | ❌ |
| Bioinformatics | ✅ 20+ skills (scRNA-seq, genomics, phylogenetics) | ❌ | ❌ |
| Clinical Research | ✅ 16+ skills (ClinicalTrials, ClinVar, COSMIC, FDA) | ❌ | ❌ |
| ML/AI Skills | ✅ 16+ skills | General coding | ✅ Fully automated ML research |
| Lab Automation | ✅ PyLabRobot, Ginkgo, Protocols.io, Benchling | ❌ | ❌ |
| Scientific Writing | ✅ 24+ skills (PubMed, OpenAlex, bioRxiv, Zotero) | ❌ | ✅ Auto-generates papers |
| Agent Compatibility | Claude Code, Cursor, Codex, any Agent Skills standard | Claude-specific | Standalone agent |
| Approach | Curated skill library (human-guided) | Official Anthropic platform | Fully autonomous research |
| Language | Python | Python | Jupyter Notebook |
When to choose Claude Scientific Skills: You're a researcher or scientist who needs your AI coding agent to execute complex, multi-step scientific workflows across biology, chemistry, clinical research, materials science, and more. The broadest coverage of scientific domains and databases of any agent skills collection.
When to choose Anthropic Skills: You want the official, general-purpose agent skills repository from Anthropic. Broadest community (86K stars), covers general coding and development tasks, but lacks specialized scientific tooling.
When to choose AI-Scientist: You want fully autonomous scientific discovery — AI that generates hypotheses, runs experiments, writes papers, and reviews them. Focused on ML research with NeurIPS-style paper generation. A different paradigm: fully autonomous vs. human-guided skills.
Why This Matters
The key insight behind Claude Scientific Skills is that while AI agents can use any Python package, they perform significantly better when given curated documentation, examples, and best practices for specific scientific tools. The skills don't limit the agent — they make it stronger and more reliable for the workflows that matter most in research.
Think of it as the difference between asking someone to "figure out RDKit" vs. giving them a comprehensive guide with examples, common pitfalls, and best practices.
Conclusion
Claude Scientific Skills bridges the gap between general-purpose AI coding agents and the specialized needs of scientific research. With 170+ skills across 17 domains, 250+ database connections, and 60+ optimized Python package skills, it's the most comprehensive open-source scientific skills collection available. Whether you're running drug discovery pipelines, analyzing single-cell RNA-seq data, performing clinical variant interpretation, or building predictive models — these skills transform your AI agent into a genuine research partner.
Explore Claude Scientific Skills on GitHub