ml-mlip-benchmark

ml-mlip-benchmark is a skill for Claude Code, Codex from learningmatter-mit/AtomisticSkills. It costs 34 tokens per session (1,734 once invoked), scanned A, original, MIT.

Benchmark MLIP accuracy against a labeled dataset — compute MAE/RMSE for energy/atom and forces, and generate parity plots.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/learningmatter-mit/atomisticskills/ml-mlip-benchmark
Any agent
npx skills add learningmatter-mit/AtomisticSkills --skill ml-mlip-benchmark
Clone the repo
git clone --depth 1 https://github.com/learningmatter-mit/AtomisticSkills

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for ml-mlip-benchmark

README.md
[![agentmods](https://agentmods.dev/badge/skills/learningmatter-mit/atomisticskills/ml-mlip-benchmark.svg)](https://agentmods.dev/skills/learningmatter-mit/atomisticskills/ml-mlip-benchmark)
Your own site
<a href="https://agentmods.dev/skills/learningmatter-mit/atomisticskills/ml-mlip-benchmark"><img src="https://agentmods.dev/badge/skills/learningmatter-mit/atomisticskills/ml-mlip-benchmark.svg" alt="Measured on agentmods" height="20"></a>
Per session 34 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,734 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin unknown No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00034 $0.01734
Opus 5 $0.00017 $0.00867
Sonnet 5 $0.00007 $0.00347
Haiku 4.5 $0.00003 $0.00173

Measured today against content hash cc57169a6200, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

ml-mlip-benchmark scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured today.

The scan reads SKILL.md. This mod also ships 3 executable files (examples/fetch_r2scan.py, scripts/plot_benchmark.py, scripts/run_benchmark.py), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.agents/skills/ml-mlip-benchmark/SKILL.md · 117 lines

How it starts

The opening of the file, as written. The whole thing — 117 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Benchmark Machine Learning Interatomic Potentials (MLIP)

This skill evaluates the accuracy of a given MLIP against an existing ground-truth dataset (e.g., DFT calculations or a higher-fidelity foundation potential). It computes the Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) for both energy (per atom) and atomic forces, and optionally stress. It also generates parity plots for visual inspection of the model's correlation.

Prerequisites

  1. Model Loaded: An MLIP must be currently active via a load_model MCP tool call (e.g., mcp_mace_load_model, mcp_fairchem_load_model, mcp_matgl_load_model).
  2. Labeled Data: A JSON dataset where each entry contains a structural dictionary under "structure", along with scalar/vector ground truth values for "energy", "forces", and optionally "stress". This is identical to the format used in ml-mlip-training. (Data can be generated using Atomate2 MongoDB queries or MD sampling + labeling).

Instructions

1. Run Benchmark metrics

Use the .agents/skills/ml-mlip-benchmark/scripts/run_benchmark.py script to perform inference across the dataset and compute global error metrics.

Environment requirement: This script instantiates the MLIP models directly and thus must be executed within the target model's conda environment (e.g., mace-agent, fairchem-agent, or matgl-agent). Run this using the run_command via conda run -n <model_agent> python ....

conda run -n <MODEL-AGENT-ENV> python .agents/skills/ml-mlip-benchmark/scripts/run_benchmark.py \
    --data_path <path_to_labeled_data.json> \
    --model <model_name_or_path> \
    --backend <mace|fairchem|matgl> \
    --output <path_to_save_benchmark_results.json>

Note: The script utilizes src.utils.mlips.loader.load_wrapper to abstract backend details.

2. Generate Parity Plots

Once run_benchmark.py finishes, it writes a comprehensive JSON file containing original targets alongside the model's predictions and numerical metrics. Visualize these using the plotting script.

Read the full file on GitHub · 117 lines

Files

What ships with it

8 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. today First seen · 117 lines · 34 tokens per session scan A cc57169a6200

Subscribe to this mod's changes

ml-mlip-benchmark is a skill published in the GitHub repository learningmatter-mit/AtomisticSkills (158 stars, last pushed yesterday), licensed MIT. It adds 34 tokens to every session and 1,734 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.

Related

Other skills, from other repositories

hugging-face-trackio

Track and visualize ML training experiments with Trackio. Use when logging metrics during training (Python API) or retrieving/analyzing logged metrics (CLI). Supports real-time dashboard visualization, HF Space syncing, and JSON output for automation.

patchy631/ai-engineering-hub · 52 tokens

pysr

Use when fitting equations to data with PySR or SymbolicRegression.jl, when a user wants an interpretable formula, symbolic model, scaling law, or empirical relation discovered from numeric data, or when debugging a PySR search that is slow, stuck, or giving poor equations.

astroautomata/PySR · 61 tokens

datamol

Pythonic wrapper around RDKit with simplified interface and sensible defaults. Preferred for standard drug discovery including SMILES parsing, standardization, descriptors, fingerprints, clustering, 3D conformers, parallel processing. Returns native rdkit.Chem.Mol objects. For advanced control or custom parameters…

synthetic-sciences/openscience · 67 tokens

deepchem

Molecular ML with diverse featurizers and pre-built datasets. Use for property prediction (ADMET, toxicity) with traditional ML or GNNs when you want extensive featurization options and MoleculeNet benchmarks. Best for quick experiments with pre-trained models, diverse molecular representations. For graph-first…

synthetic-sciences/openscience · 78 tokens

binding-affinity

Hybrid ML + physics binding affinity prediction. Empirical scoring, MM/GBSA rescoring, multi-method consensus, and batch virtual screening for protein-ligand complexes.

synthetic-sciences/openscience · 38 tokens

drug-design

End-to-end drug discovery pipeline orchestration. Deterministic Python script that auto-chains structure prediction, pocket detection, de novo design, docking, scoring, and ADMET filtering into reproducible workflows.

synthetic-sciences/openscience · 44 tokens