Skills biomni
🧬

biomni

Low Risk ⚙️ External commands🔑 Env variables📁 Filesystem access🌐 Network access

Automate biomedical research with AI agents

Also available from: davila7

Biomni transforms complex biomedical research by autonomously executing multi-step analysis tasks. Researchers can focus on scientific questions while AI handles data processing, literature review, and computational analysis across genomics, drug discovery, and clinical domains.

Supports: Claude Codex Code(CC)
🥉 75 Bronze
1

Download the skill ZIP

2

Upload in Claude

Go to Settings → Capabilities → Skills → Upload skill

3

Toggle on and start using

Test it

Using "biomni". Design a CRISPR screen for autophagy regulators

Expected outcome:

  • Generated sgRNA library with 76,230 guides targeting 19,057 genes
  • Designed 4 sgRNAs per gene with on-target scores above 0.7
  • Included positive controls: ATG5, BECN1, ULK1, mTOR
  • Prioritized 347 candidate genes based on pathway analysis
  • Provided Python code for screen analysis pipeline

Using "biomni". Analyze single-cell RNA-seq from tumor samples

Expected outcome:

  • Identified 12 distinct cell populations via clustering
  • Annotated major immune cell types: T cells, B cells, macrophages
  • Found 3 novel cell clusters with unknown markers
  • Differential expression revealed 234 upregulated genes in tumor region

Security Audit

Low Risk
v4 • 1/17/2026

The static analysis flagged 415 patterns, but 95% are FALSE POSITIVES from markdown documentation. The backtick patterns are markdown code delimiters, not shell execution. The API key patterns show example environment variable names in documentation, not actual secrets. The skill is a legitimate Stanford SNAP lab biomedical research framework. The code execution + network + credential combination is the intended design for an AI agent that generates bioinformatics analysis code. Proper security warnings are documented recommending sandboxed execution.

7
Files scanned
3,120
Lines analyzed
4
findings
4
Total audits
Audited by: claude View Audit History →

Quality Score

68
Architecture
100
Maintainability
87
Content
21
Community
90
Security
91
Spec Compliance

What You Can Build

Design genome-wide CRISPR screens

Automate sgRNA library design, gene prioritization, and knockout effect analysis for functional genomics studies

Process single-cell sequencing data

Perform quality control, clustering, cell type annotation, and differential expression analysis

Predict compound ADMET properties

Evaluate absorption, distribution, metabolism, excretion, and toxicity for drug candidates

Try These Prompts

CRISPR Screen Design
Design a CRISPR knockout screen to identify genes regulating autophagy in HEK293 cells. Include sgRNA library design, positive/negative controls, and gene prioritization based on pathway relevance.
Single-Cell Analysis
Analyze this single-cell RNA-seq dataset: perform QC, identify cell populations via clustering, annotate cell types using marker genes, and conduct differential expression. File: path/to/data.h5ad
GWAS Interpretation
Interpret GWAS results for Type 2 Diabetes: identify genome-wide significant variants, map to causal genes, perform pathway enrichment, and predict functional consequences
Drug ADMET Prediction
Predict ADMET properties for these compounds: [SMILES strings]. Focus on Caco-2 permeability, plasma protein binding, CYP450 interactions, clearance, and hERG toxicity

Best Practices

  • Specify biological context including organism, cell type, and experimental conditions
  • Provide data file paths when analyzing datasets
  • Set computational constraints for complex analyses
  • Save conversation history for reproducibility

Avoid

  • Running without reviewing generated code in production environments
  • Sharing API keys or credentials in shared environments
  • Processing sensitive clinical data without proper authorization
  • Ignoring timeout settings for long-running analyses

Frequently Asked Questions

Is biomni safe to use?
Yes, biomni is from Stanfords SNAP lab. Run in isolated environments as it executes generated code with system privileges.
What LLM providers does biomni support?
Anthropic Claude (recommended), OpenAI GPT, Google Gemini, Groq, Azure OpenAI, and AWS Bedrock.
How much data does biomni download?
Approximately 11GB of biomedical databases including gene annotations, protein structures, and clinical datasets.
Can biomni analyze my experimental data?
Yes, provide file paths to your datasets in natural language queries. Supports common formats like FASTA, BAM, VCF, and HDF5.
What domains does biomni cover?
Genomics, proteomics, drug discovery, clinical genomics, molecular biology, and multi-omics integration.
How do I cite biomni?
Reference the preprint: https://www.biorxiv.org/content/10.1101/2025.05.30.656746v1 and GitHub repository.