learningmatter-mit/AtomisticSkills

README preview

# AtomisticSkills

![AtomisticSkills Logo](site/logo/atomisticskills_logo.svg)

[![arXiv](https://img.shields.io/badge/arXiv-2605.24002-b31b1b.svg)](https://arxiv.org/abs/2605.24002)

## Overview
**AtomisticSkills** is a composable framework for AI-driven atomistic materials research. Built on the **hierarchical decomposition** of complex scientific tasks into **Workflows** → **Skills** → **Tools**, it enables coding AI agents to autonomously conduct multi-stage materials, chemistry, and drug discovery research by combining modular, reusable capabilities.

The framework integrates state-of-the-art Machine Learning Interatomic Potentials (MLIPs), DFT calculations, generative AI, database APIs, and advanced simulation methods through the Model Context Protocol (MCP) tools and Skills, making advanced materials research accessible to AI copilots like [Google Antigravity](https://antigravity.google), [Cursor](https://www.cursor.com/), [Claude Code](https://code.claude.com/docs/en/overview), and [OpenAI Codex](https://openai.com/codex/).

<div align="center">

🌐 **[Documentation Website](https://learningmatter-mit.github.io/AtomisticSkills/)** &nbsp;|&nbsp; 📄 **[Preprint](https://arxiv.org/abs/2605.24002)**

</div>

<div align="center">

🎬 **Video Demo: Using AtomisticSkills in Google Antigravity** — **[Watch on YouTube](https://www.youtube.com/watch?v=oiyZ52RS5oo&feature=youtu.be)**

View full repository on GitHub →