What is LabClaw?
LabClaw is a skill library of 240 production-ready SKILL.md files for biomedical AI workflows. Created by researchers at Stanford (Le Cong Lab) and Princeton (Mengdi Wang Lab), it covers biology, drug discovery, clinical research, literature search, and scientific visualization.
Unlike monolithic software, LabClaw is modular — install the full collection or cherry-pick only the skill folders relevant to your research.
3-Second Quick Start
Just send this to OpenClaw:
install https://github.com/wu-yc/LabClawThat's it. All 240 skills are now available in your agent workspace.
The 7 Domains
| Domain | Skills | Focus |
|---|---|---|
| Biology & Life Sciences | 86 | Bioinformatics, single-cell, genomics, proteomics, multi-omics, databases |
| Pharmacy & Drug Discovery | 36 | Cheminformatics, molecular ML, docking, target research, pharmacology |
| General & Data Science | 54 | Statistics, machine learning, scientific writing, quality control |
| Literature & Search | 33 | Academic search, biomedical databases, patents, grants, citations |
| Medical & Clinical | 22 | Clinical trials, precision medicine, oncology, medical imaging |
| Vision & XR | 5 | Hand tracking, 3D pose estimation, segmentation |
| Visualization | 4 | matplotlib, seaborn, plotly, publication-ready figures |
Representative Workflows
Single-cell and Spatial Omics
Use skills like anndata, scanpy, and tooluniverse-spatial-transcriptomics to go from raw data to publication-ready spatial analysis:
You: Analyze my Visium spatial data and identify tissue domains.
Agent: [Uses scanpy skill for preprocessing]
[Uses spatial-transcriptomics skill for domain identification]
Found 8 distinct spatial domains. Generating publication figure...Drug Discovery Pipeline
Combine rdkit, diffdock, and tooluniverse-drug-repurposing for molecular design:
You: Find candidate molecules for EGFR inhibition and predict binding poses.
Agent: [Uses rdkit skill for molecular filtering]
[Uses diffdock skill for docking simulation]
Top 3 candidates with predicted binding affinities...Literature Review
Use pubmed-search, citation-management, and scientific-writing to accelerate reviews:
You: Find recent papers on CAR-T cell therapy for solid tumors
and write a summary paragraph.
Agent: [Searches PubMed, retrieves 47 papers from 2025-2026]
[Summarizes top 10 by impact factor]
[Generates citation-ready paragraph]Skill Format
Every SKILL.md follows a consistent structure:
# Skill Name
## Overview — what this skill enables
## When to Use — trigger conditions for the AI agent
## Key Capabilities — specific tools, APIs, parameters
## Usage Examples — concrete code or workflow examplesThis standardized format means any OpenClaw-compatible agent can immediately understand and use each skill.
Key Skills by Domain
Biology (86 skills)
Includes: scanpy, anndata, biopython, tooluniverse-genomics, tooluniverse-proteomics, tooluniverse-spatial-transcriptomics, tooluniverse-single-cell, and many more.
Drug Discovery (36 skills)
Includes: rdkit, diffdock, tooluniverse-drug-repurposing, tooluniverse-molecular-generation, chembl-database, drugbank-database, and more.
Literature (33 skills)
Includes: pubmed-search, google-scholar, arxiv-search, citation-management, scientific-writing, patent-search, grant-writing, and more.
Related Projects
| Project | Connection |
|---|---|
| OpenClaw | The runtime that loads LabClaw skills |
| ToolUniverse | Harvard AI-scientist tool ecosystem — many tooluniverse-* skills in LabClaw |
| Biomni | Stanford biomedical AI agent — LabClaw includes a biomni skill |
Links
- GitHub Repository
- Project Website
- Claw4Science — explore more OpenClaw ecosystem projects for science