Skill Data Limited info

scientific-agent-skills

A library of 177 validated scientific skills enabling AI agents to perform complex research workflows across biology, chemistry, medicine, physics, and engineering.

  • 48.2k GitHub stars
View on GitHub (opens in a new tab) Setup guide
npx skills add K-Dense-AI/scientific-agent-skills
scientific-agent-skills — illustration

What it is

Scientific Agent Skills is a large-scale skill library from K-Dense AI containing 177 validated skills that enable coding agents to perform complex scientific research workflows. The skills span bioinformatics, cheminformatics, proteomics, clinical research, microscopy, neuroscience, machine learning, materials science, physics, engineering simulation, data analysis, and scientific communication. Each skill includes a SKILL.md file documenting its purpose, workflows, code examples, and validation requirements. The library provides curated guidance for 70+ Python packages (RDKit for chemistry, Scanpy for single-cell genomics, PyTorch Lightning for ML, scikit-learn for general data science) and documented retrieval methods for over 100 scientific databases. Nine specialized integration skills connect to platforms like Benchling and DNAnexus. Security is maintained through weekly automated scans using the Cisco AI Defense Skill Scanner, with results published in a security report. The library follows the agentskills.io standard with progressive disclosure -- agents scan YAML frontmatter efficiently (about 30 tokens) and load full workflows only as needed. The project is MIT licensed and includes a published research paper on arXiv.

Who it's for

  • research scientists
  • bioinformaticians
  • computational chemists
  • data scientists in scientific fields

Requirements

Requirements

  • Claude Code, Cursor, Codex, or compatible agent platform
  • Python with relevant scientific packages (RDKit, Scanpy, PyTorch, scikit-learn, etc.)

Setup

  1. Step 1

    bash
    npx skills add K-Dense-AI/scientific-agent-skills

مكتبة من 177 مهارة علمية موثقة تمكن وكلاء الذكاء الاصطناعي من أداء سير عمل البحث العلمي المعقد.

Pros & cons

Pros

  • Pro:177 skills across biology, chemistry, medicine, physics, and engineering
  • Pro:Access to 100+ scientific databases with documented retrieval methods
  • Pro:70+ optimized Python package integrations
  • Pro:Weekly security scans using Cisco AI Defense Skill Scanner

Cons

  • Con:Requires domain-specific Python packages to be installed for each scientific field
  • Con:Some skills require access to specialized platforms like Benchling or DNAnexus