skills/vasp-ase/SKILL.md
Professional skill for setting up, executing, and debugging VASP DFT calculations using the Atomic Simulation Environment (ASE).
npx skillsauth add hkqai/MatClaw vasp-aseInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
This skill allows the agent to interface with VASP on high-performance computing (HPC) systems like the DGX A100. It focuses on using ASE as the wrapper for clean Pythonic control.
VASP_PP_PATH must point to the directory containing potpaw_PBE, etc.ase and pymatgen (optional but recommended for analysis) must be installed in the (base) conda environment.vasp_std, vasp_gam, or vasp_ncl.Always initialize the VASP calculator using the ase.calculators.vasp.Vasp class.
xc='PBE', kpts=(4, 4, 4), and encut=520.mpirun or srun. Example command:
command='mpirun -np 16 vasp_std > vasp.out'from ase.io import read
from ase.calculators.vasp import Vasp
from ase.optimize import BFGS
atoms = read('POSCAR')
calc = Vasp(directory='run_dir',
command='mpirun -np 16 vasp_std',
xc='PBE',
encut=520,
ismear=0,
sigma=0.05,
lreal='Auto',
nsw=100,
ibrion=2)
atoms.set_calculator(calc)
energy = atoms.get_potential_energy()
If a calculation fails, the agent should parse OUTCAR or stdout and apply these fixes:
| Error/Issue | Diagnostic | Fix Strategy |
| :--- | :--- | :--- |
| Electronic Convergence | NELM reached without dE < ediff | Set ALGO = Normal or Fast; increase NELM to 100; try AMIX = 0.2. |
| Ionic Convergence | Max steps reached | Restart from the last CONTCAR; check if forces are oscillating. |
| Memory/A100 Crash | Segmentation fault | Reduce NCORE or KPAR; ensure LREAL = Auto. |
| Missing POTCAR | RuntimeError: POTCAR not found | Verify VASP_PP_PATH is exported in ~/.bashrc. |
Before running large productions, the agent should run a "Dry Run" by setting LSTOP = .TRUE. in the INCAR or running a single-point energy on a 2-atom cell to verify the environment pathing.
energy_free, forces, and stress.tools
Analyze thermodynamic stability of inorganic materials by routing between two workflows: a cheap Materials Project-backed lookup path for known compositions, and a custom self-consistent MLIP hull workflow for novel or structure-specific materials. Use this skill whenever the user asks whether a material is stable, requests energy above hull, decomposition products, polymorph context, or wants to include stability as a screening criterion. This skill is intended to become the single orchestration layer for stability analysis, with workflow branching handled in the skill rather than inside an MCP tool.
tools
Pre-import validation and auto-fix for URDF files targeting Isaac Sim / USD
tools
Intelligent synthesis route planning for inorganic materials. Use this skill whenever the user needs a synthesis protocol.
development
Optimize CUDA/GPU simulation code using NVIDIA Nsight Systems (nsys) profiling. Use this skill whenever the user mentions performance problems, slow simulations, profiling, nsys, Nsight Systems, kernel optimization, GPU bottlenecks, or wants to speed up CUDA code. Also trigger when the user compares two scenes and one is unexpectedly slower, or asks "why is this slow?" about GPU code. This skill covers the full optimization loop: profiling, bottleneck diagnosis, targeted optimization, verification, and iterative measurement.