AtomisticSkills
Autonomous atomistic simulations for materials, chemistry, and drug discovery
All Skills in This Repository (121)
Pure Emerald Level Indicators
learningmatter-mitmat-defect-energy-dft
Compute charged point-defect formation energies and Fermi-level transition diagrams from VASP DFT outputs.
learningmatter-mitmat-surface-energy
Calculate surface energies for (hkl) planes and construct equilibrium Wulff shapes.
learningmatter-mitml-fairchem-finetune
Fine-tune Fairchem interatomic potentials on labeled structure datasets.
learningmatter-mitdrug-pocket-detection
Detect and rank ligandable protein pockets using fpocket or P2Rank.
learningmatter-mitmat-dft-lobster
Generate and run VASP-to-LOBSTER projection workflows for COHP bonding analysis.
learningmatter-mitml-property-predict-scd
Train atomistic property prediction models from SelfConditionedDenoisingAtoms checkpoints.
learningmatter-mitml-mlip-automl
Automate MLIP hyperparameter tuning via LLM-driven iterative search over learning rate, freezing, and loss weighting.
learningmatter-mitmat-solid-free-energy
Calculate solid Helmholtz free energy via Frenkel-Ladd thermodynamic integration.
learningmatter-mitml-generative-adit
Generate periodic crystal and molecular structures as CIF or XYZ outputs.
learningmatter-mitdrug-db-pdb
Search RCSB Protein Data Bank and export structure metadata as JSON.
learningmatter-mitmat-dft-electronic-transport
Compute carrier mobility, conductivity, and Seebeck coefficient from DFT band structures using AMSET.
learningmatter-mitchem-bond-dissociation
Calculate homolytic and heterolytic bond dissociation energies for cleavable single bonds.
Frequently Asked Questions
FAQPage SchemaHow to install AtomisticSkills?▼
Run `npx skills add learningmatter-mit/AtomisticSkills --all -g -y` in your terminal to install all skills globally. Then open the repository in your agent IDE and ask it to follow the setup guide to create conda environments and register MCP servers.
What can AtomisticSkills do for materials research?▼
It automates structure relaxation, molecular dynamics, phonon and defect calculations, phase diagrams, XRD refinement, and generative crystal design using MLIPs and DFT. It also queries databases like Materials Project and OPTIMADE for structures and properties.
Can AtomisticSkills help with drug discovery?▼
Yes. It covers protein preparation, pocket detection, molecular docking with AutoDock Vina, pose validation, protein-ligand MD, MM-GBSA rescoring, ADMET prediction, and retrosynthesis planning.
Which AI coding tools work with AtomisticSkills?▼
It works with Claude Code, Cursor, Google Antigravity, OpenAI Codex, and any agent that reads the standard SKILL.md format. The included CLAUDE.md and AGENTS.md files route your requests to the right skill automatically.
Do I need DFT or HPC experience to use AtomisticSkills?▼
No. You describe the research goal in plain English, and the agent selects skills, prepares inputs, submits HPC jobs via atomate2, and parses results. GPU access is recommended for fast MLIP inference but not strictly required.
Related Repositories in Education & Research
View All in Education & Research→
shareAI-lablearn-claude-code
Learn to build AI agent harnesses from scratch
microsoftai-agents-for-beginners
Learn to build, deploy, and secure AI agents step by step
leonardomso33-js-concepts
Master 33 essential JavaScript concepts with tested, verified guides