メニュー
Generate candidate crystal structures from a chemical formula and assess their stability using ggen. Use when you need trial structures for a composition with no known experimental structure, or want stability-ranked polymorph candidates before running MD/DFT.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Get fully up to date on a project, or build its weekly / PI-meeting update bundle, from INSIDE the project directory. Use when an agent lands cold in a project and needs the current state (open workstreams, what moved, open PI decisions, blockers), or when asked to "build the weekly update", "prep the PI meeting", or "aggregate the relevant files into the update dir". Mines only THIS repo, driven by its own .sync/manifest.yaml. No dependency on LabSync.
Run Quantum ESPRESSO DFT calculations. Use when asked to perform first-principles calculations, SCF, structural relaxation, band structure, DOS, phonons, or any ab initio quantum mechanical calculation.
Decide where and how to run compute jobs across available backends (HPC, cloud GPUs, local). Use when starting any compute-intensive task, choosing a backend or partition, deciding job parameters (walltime, GPU type, parallelism), or debugging compute infrastructure issues. Embeds the smoke-first iteration pattern — validate cheaply before scaling.
Orchestrate one or more long-running compute campaigns (HPC simulations, multi-stage MD pipelines, ensemble runs) where stages take hours to days and failures must not halt the pipeline. Use when running multiple parallel campaigns that share a stage structure (smoke → equilibration → production → analysis), when waiting on long jobs makes interactive driving impractical, or when designing a self-healing tick-based supervisor for compute work. Reads/writes per-campaign WORKFLOW.md files as durable state.
Analyze simulation data and compute properties. Use when asked to parse LAMMPS/QE output, calculate diffusion coefficients, RDF, MSD, energies, or generate plots and visualizations.
Run compute jobs (QE DFT, LAMMPS/MD, MLIP) that may outlive a single agent turn, without parking-and-dying. Use whenever a task launches work that could take more than a few minutes — decide block-vs-detach, detach long jobs durably, keep state in files, and let the asw-loop supervisor re-invoke you across ticks. Read BEFORE launching any simulation whose runtime you are unsure of.
| name | ggen |
| description | Generate candidate crystal structures from a chemical formula and assess their stability using ggen. Use when you need trial structures for a composition with no known experimental structure, or want stability-ranked polymorph candidates before running MD/DFT. |
Crystal structure generation and stability assessment using ggen.
Validated: 2026-01-24 - API tested and working.
Use ggen when you need to:
Use Materials Project / materials-database skill when:
ggen generates NEW candidate structures. Materials databases provide KNOWN structures.
IMPORTANT: Install from GitHub, NOT PyPI:
pip install git+https://github.com/ourofoundation/ggen.git
WARNING: pip install ggen installs a different package (grid mesh generator).
Requires GPU for efficient relaxation.
python -c "from ggen import GGen; print('ggen:', GGen)"
ORB models cached in: ~/.cache/science-agent/ggen/
See: showcases/ggen-integration/workspace/demo_ggen.py
High-level systematic exploration:
from ggen import ChemistryExplorer
explorer = ChemistryExplorer(output_dir="./runs")
result = explorer.explore(
chemical_system="Li-P-S", # Elements to explore
max_atoms=20, # Max atoms per unit cell
num_trials=25, # Generation attempts per stoichiometry
compute_phonons=False # Skip phonons for speed
)
Lower-level structure generation:
from ggen import GGen
ggen = GGen()
result = ggen.generate_crystal(
formula="BaTiO3",
num_trials=10,
optimize_geometry=True
)
# result is a dict with these keys:
# - 'formula': str
# - 'structure': pymatgen Structure (best structure)
# - 'final_space_group': int
# - 'final_space_group_symbol': str
# - 'best_crystal_energy': float (eV)
# - 'optimization_steps': int
# - 'final_fmax': float (eV/A)
# - 'all_relaxed_trials': list of dicts (all generated structures)
# - 'cif_content': str (CIF file content)
print(f"Best: {result['final_space_group_symbol']}, E={result['best_crystal_energy']:.4f} eV")
structure = result['structure'] # pymatgen Structure
from ggen import ChemistryExplorer
explorer = ChemistryExplorer(output_dir="./ggen_runs")
result = explorer.explore(
chemical_system="Li-Co-O",
max_atoms=24,
num_trials=30,
min_fraction={"Li": 0.2}, # At least 20% Li
compute_phonons=True # Check stability
)
# Access results
print(f"Generated {result.total_structures} structures")
print(f"On-hull phases: {result.on_hull_count}")
# Get stable structures
for structure in result.stable_structures:
print(f"{structure.formula}: {structure.energy_above_hull:.3f} eV/atom")
# Explore a system
python -m ggen.scripts.explore Fe-Mn-Si --max-atoms 24 --num-trials 25
# Run phonon calculations on candidates
python -m ggen.scripts.phonons --system Li-P-S --e-above-hull 0.05
# Export top candidates
python -m ggen.scripts.export Li-Co-O -n 10
# Generate report
python -m ggen.scripts.report Li-P-S
from ggen import ChemistryExplorer
explorer = ChemistryExplorer(output_dir="./runs")
result = explorer.explore(chemical_system="Li-P-S", max_atoms=16)
# Get CIF files for stable candidates
for structure in result.get_structures_near_hull(e_above_hull=0.05):
cif_path = structure.to_cif(f"{structure.formula}.cif")
# Now use these structures for MD, DFT, etc.
# with lammps-simulation, quantum-espresso, or mlip-simulation skills
ggen generates structures. Other skills analyze them:
ggen (structure generation)
│
├── mlip-simulation → Quick ML potential checks
├── torch-sim → High-throughput screening
├── lammps-simulation → Classical MD
├── quantum-espresso → DFT validation
└── data-analysis → Property analysis
# 1. Generate candidates with ggen
from ggen import ChemistryExplorer
explorer = ChemistryExplorer()
result = explorer.explore("Li-P-S", max_atoms=20)
# 2. Get promising structures
candidates = result.get_structures_near_hull(e_above_hull=0.1)
# 3. Further analyze with other tools (e.g., torch-sim for MD)
from torch_sim import integrate
from torch_sim.models import MACEModel
model = MACEModel.from_pretrained("mace-mp-0-medium")
for candidate in candidates:
structure = candidate.to_pymatgen()
# Run MD to check ionic conductivity...
| Parameter | Default | Description |
|---|---|---|
max_atoms | 20 | Maximum atoms per unit cell |
min_atoms | 2 | Minimum atoms per unit cell |
num_trials | 15 | Generation attempts per stoichiometry |
max_stoichiometries | 100 | Limit search space |
compute_phonons | False | Run phonon stability checks |
min_fraction | {} | Minimum element fractions |
preserve_symmetry | True | Maintain space group during relaxation |
min_fraction / max_fractionmaterials-database skill - Query known structuresmlip-simulation skill - ML potential calculationstorch-sim skill - High-throughput screening