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.
npx skills add learningmatter-mit/AtomisticSkills --skill ml-fairchem-finetunegit clone --depth 1 https://github.com/learningmatter-mit/AtomisticSkillsWrote 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.
[](https://agentmods.dev/skills/learningmatter-mit/atomisticskills/ml-fairchem-finetune)<a href="https://agentmods.dev/skills/learningmatter-mit/atomisticskills/ml-fairchem-finetune"><img src="https://agentmods.dev/badge/skills/learningmatter-mit/atomisticskills/ml-fairchem-finetune.svg" alt="Measured on agentmods" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium Agent Snooping · line 56 Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.Fix: Remove all code or instructions that list or read other skills' files or directories. Skills should operate independently; cross-skill access is a privilege escalation.
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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00028 | $0.01699 |
| Opus 5 | $0.00014 | $0.00849 |
| Sonnet 5 | $0.00006 | $0.00340 |
| Haiku 4.5 | $0.00003 | $0.00170 |
Grade A, and why
ml-fairchem-finetune 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 4d ago.
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.
How it starts
The opening of the file, as written. The whole thing — 90 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Fairchem Fine-tuning
Goal
To evaluate and improve the accuracy of a foundation Fairchem potential (e.g., UMA, ESEN) for a specific chemical system or physical property using the provided Python fine-tuning script.
Instructions
- Prepare Labeled Dataset: Obtain diverse structures with high-fidelity labels (energy, forces, stress). See the
/benchmark-finetuningworkflow for details. - Custom Data Conversion: Read the source data format and write a customized conversion script if needed, formatting it for the subsequent preparation step.
- Benchmarking: Predict results on the new labels and benchmark the foundation model using ml-mlip-benchmark.
- Data Preparation: Execute
scripts/prepare_fairchem_data.pyto convert JSON structures to extxyz, generate native LMDB databases, compute dataset references, and configure a templateduma_sm_finetune_template.yaml. - Fine-Tuning: Execute
fairchem -c uma_sm_finetune_template.yaml job.run_dir=XXXnatively. - Validation: Run
scripts/extract_fairchem_logs.pyto extract curves and verify convergence against the benchmarked foundation metrics. - Registration: Use the
register_modeltool to register the newly fine-tuned model checkpoint into the local registry so future research tasks can discover and reuse it.
Training Configuration
Fairchem fine-tuning relies heavily on the fairchem CLI, which uses Hydra for configuration. The script scripts/prepare_fairchem_data.py bridges standard data into the complex Fairchem directory structure and generates .aselmdb dataset formats automatically.
Basic Arguments (Data Prep Script)
| Key | Type | Default | Description |
|---|---|---|---|
--data |
str | (Required) | Path to JSON file containing ASE/pymatgen structure dictionaries |
--val-data |
str | None | Path to JSON file containing validation split. (Optional, otherwise --val-split is used) |
--val-split |
float | 0.1 | Validation split if --val-data is not provided |
--seed |
int | 42 | Random seed for data splitting and initialization |
--model |
str | uma-s-1p1 |
Base model name or path to a checkpoint |
--task-name |
str | omat |
The specific multi-task context to run against (omat, omol) |
--epochs |
int | 10 | Number of training epochs |
--lr |
float | 4e-4 | Peak learning rate for training |
--batch-size |
int | 2 | Training batch size |
--freeze-backbone |
flag | N/A | Add flag to mathematically freeze OCP/UMA interaction layers |
--weight-decay |
float | 1e-3 | Weight decay parameter |
--warmup-factor |
float | 0.2 | LR warmup factor |
--warmup-epochs |
float | 0.01 | Epochs to perform LR warmup |
--lr-min-factor |
float | 0.01 | Minimum LR factor after decay |
--clip-grad-norm |
float | 100.0 | Gradient clipping threshold |
--evaluate-every-n-steps |
int | 100 | Steps frequency for validation evaluation |
--checkpoint-every-n-steps |
int | 1000 | Steps frequency for model checkpointing |
--ema-decay |
float | 0.999 | Exponential moving average decay parameter |
--linref-coeff |
str | None | JSON array of elemental energy linear references. If None, it auto-computes it over the data. |
--vasp-stress-conversion |
flag | N/A | Add flag to automatically convert kB to eV/ų for VASP inputs |
--output-dir |
str | ./fairchem_finetuning |
Directory to save the lmdb_output intermediate data and run configs |
What ships with it
11 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.
- examples/fairchem-wbm-finetune/_freeze_backbone_helper.py 458 B runs code
- examples/fairchem-wbm-finetune/dataset_metadata.json 2.0 KB
- examples/fairchem-wbm-finetune/README.md 3.3 KB
- examples/fairchem-wbm-finetune/run.sh 1.5 KB runs code
- examples/fairchem-wbm-finetune/training_history.json 2.1 KB
- examples/fairchem-wbm-finetune/training_history.png 337 KB
- examples/fairchem-wbm-finetune/uma_sm_finetune_template.yaml 3.3 KB
- scripts/extract_fairchem_logs.py 6.3 KB runs code
- scripts/generate_fairchem_config.py 9.9 KB runs code
- scripts/prepare_fairchem_data.py 12 KB runs code
- test_fairchem_finetuning.sh 1.2 KB runs code
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.
- 4d ago First seen · 90 lines · 28 tokens per session scan A 55def108c575
ml-fairchem-finetune is a skill published in the GitHub repository learningmatter-mit/AtomisticSkills (161 stars, last pushed 4d ago), licensed MIT. It adds 28 tokens to every session and 1,699 once invoked, about $0.0001 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.
Other skills, from other repositories
physicsnemo-discover
Official NVIDIA-authored guidance for navigating PhysicsNeMo — pick the model, datapipe, or example for a SciML/AI4Science task (surrogates, forecasting, downscaling, physics-informed, inverse, generative). Points at existing files via live repo search; never writes code. Do NOT use for installation or environment…
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.
cellxgene-census-query
Query CZ CELLxGENE Census (61M+ cells). Filter by cell type/tissue/disease, retrieve expression data, and integrate with scanpy/PyTorch for population-scale single-cell analysis. Use this skill when: (1) Querying single-cell expression data by cell type, tissue, or disease, (2) Exploring available single-cell datasets…
ml-research-lab
Machine-learning research loop for dataset curation, fine-tuning, evaluation, inference deployment, experiment tracking, and model explainability. Use when working on ML experiments, training data, model benchmarks, RunPod/GPU runs, classifier quality, vLLM/GGUF serving, SHAP-style model explanations, or…
ml-feedback-ladder
Use when planning how to verify an ML experiment cheaply before expensive runs - design the R0-R7 ladder from local sanity checks to full study, with promotion/stop criteria.
ml-for-aec
Computer vision for buildings, image-to-floorplan, generative ML models, performance prediction, structural analysis ML, energy prediction, natural language to design, and point cloud ML for AEC computational design.