How do I install the Mat Elemental Energies skill?

Run npx skillmds@latest add learningmatter-mit/mat-elemental-energies in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Mat Elemental Energies safe to use?

This skill has not completed SkillMD's automated safety review yet. Independent scanners report: SkillSpector: PASS, Skill Scanner: PASS. Capability flags: executes scripts, makes network calls. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Mat Elemental Energies?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Mat Elemental Energies free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Mat Elemental Energies?

learningmatter-mit (@learningmatter-mit) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Mat Elemental Energies by learningmatter-mit

Name: Mat Elemental Energies
Rating: 4.666666666666667 (3 reviews)
Author: learningmatter-mit

File contents

Elemental Energies

Goal

To provide a centralized library of the most stable phases (ground states) for each element and their corresponding reference energies calculated using different Machine Learning Interatomic Potentials (MLIPs). This avoids redundant relaxations and ensure consistency in thermodynamics calculations (e.g., formation energy, stability).

Instructions

1. Retrieve Elemental Energies

To get the energies for a list of elements from a specific checkpoint:

# Env: base-agent
python .agents/skills/mat-elemental-energies/scripts/get_elemental_energies.py --elements Li Fe O --checkpoint mace-mp-medium

Library Status

As of 2026-01-30, the library is fully expanded and contains 89 elements (Ground states from H to Lr) across 36 MLIP variants, including:

MACE: MACE-MP (S/M/L), MACE-MH-0/1 (all heads), MACE-OMAT (S/M), MACE-MATPES (PBE/R2SCAN).
MatGL: CHGNet (MPtrj, MatPES), M3GNet (MP, MatPES), TensorNet (MatPES).
FairChem: UMA (S/M) with all heads (omat, omol, oc20).

Constraints

Materials Project: Structures must be queried from Materials Project to ensure they represent the true ground-state phases.
Naming: Library files must follow the <checkpoint_name>_energies.json format.
Units: Energies are stored as eV/atom (total potential energy of the relaxed structure divided by the number of atoms).
Relaxation: The stored value comes from relaxing both cell and coordinates (relax_cell=True). A positions-only relaxation leaves the reference above the potential's own minimum and biases every formation energy that uses it.
Molecular-crystal elements are sensitive: for O, N, F, Cl, H the MP ground state is a van der Waals-bound molecular crystal whose MLIP-relaxed cell differs a lot from the DFT cell, so the cell relaxation matters more here than for metals. With TensorNet-PES-MatPES-PBE-2025.2, mp-12957 (O2) gives -5.118 eV/atom fully relaxed versus -4.968 positions-only. More than one MP entry can also tie at energy_above_hull = 0 for these elements (O2: mp-12957 and mp-1524462), so pin the material_id rather than relying on a formula lookup returning a stable order.

Author: Bowen Deng Contact: GitHub @learningmatter-mit

---name: mat-elemental-energiesdescription: A library of ground-state element structures and their energies calculated from MLIPs. Used to calculate formation energies of compounds.---# Elemental Energies## GoalTo provide a centralized library of the most stable phases (ground states) for each element and their corresponding reference energies calculated using different Machine Learning Interatomic Potentials (MLIPs). This avoids redundant relaxations and ensure consistency in thermodynamics calculations (e.g., formation energy, stability).## Instructions### 1. Retrieve Elemental EnergiesTo get the energies for a list of elements from a specific checkpoint:```bash# Env: base-agentpython .agents/skills/mat-elemental-energies/scripts/get_elemental_energies.py --elements Li Fe O --checkpoint mace-mp-medium```## Library StatusAs of **2026-01-30**, the library is fully expanded and contains **89 elements** (Ground states from H to Lr) across **36 MLIP variants**, including:- **MACE**: MACE-MP (S/M/L), MACE-MH-0/1 (all heads), MACE-OMAT (S/M), MACE-MATPES (PBE/R2SCAN).- **MatGL**: CHGNet (MPtrj, MatPES), M3GNet (MP, MatPES), TensorNet (MatPES).- **FairChem**: UMA (S/M) with all heads (omat, omol, oc20).## Constraints- **Materials Project**: Structures must be queried from Materials Project to ensure they represent the true ground-state phases.- **Naming**: Library files must follow the `<checkpoint_name>_energies.json` format.- **Units**: Energies are stored as **eV/atom** (total potential energy of the relaxed structure divided by the number of atoms).- **Relaxation**: The stored value comes from relaxing **both cell and coordinates** (`relax_cell=True`). A positions-only relaxation leaves the reference above the potential's own minimum and biases every formation energy that uses it.- **Molecular-crystal elements are sensitive**: for O, N, F, Cl, H the MP ground state is a van der Waals-bound molecular crystal whose MLIP-relaxed cell differs a lot from the DFT cell, so the cell relaxation matters more here than for metals. With `TensorNet-PES-MatPES-PBE-2025.2`, mp-12957 (O2) gives -5.118 eV/atom fully relaxed versus -4.968 positions-only. More than one MP entry can also tie at `energy_above_hull = 0` for these elements (O2: mp-12957 and mp-1524462), so pin the `material_id` rather than relying on a formula lookup returning a stable order.---**Author:** Bowen Deng**Contact:** [GitHub @learningmatter-mit](https://github.com/learningmatter-mit)

Frequently asked questions

Mat Elemental Energies