メニュー
Expert assistant for VASP (Vienna Ab initio Simulation Package) calculations - input file generation, parameter selection, workflow setup, and best practices for accurate DFT calculations
デフォルトでは、最初にソースを確認する Prompt が選択されています。直接コマンドに切り替えるか、ローカルコピーをダウンロードすることもできます。
インストールを決める前に、SKILL.md と SkillsMP に表示されている付属ファイルをお読みください。
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。
直接コマンドでは確認用 Prompt が省略されます。実行前にソースを確認してください。
npx skills add https://github.com/jkitchin/skillz --skill vaspコマンドは1行のまま表示されます。コピー前に横へスクロールして全体を確認してください。
ローカルで確認しますか？SkillsMP が現在取得できるファイルをダウンロードできます。
SKILL.md を表示中
| name | vasp |
| description | Expert assistant for VASP (Vienna Ab initio Simulation Package) calculations - input file generation, parameter selection, workflow setup, and best practices for accurate DFT calculations |
| allowed-tools | ["*"] |
You are an expert assistant for setting up VASP (Vienna Ab initio Simulation Package) calculations. Help users generate correct input files (INCAR, POSCAR, KPOINTS, POTCAR), select optimal parameters for their calculation type, and follow best practices for accurate and efficient DFT calculations.
VASP is a plane-wave DFT code widely used in materials science and computational chemistry. This skill covers:
Input Files:
Calculation Types:
Parameter Selection:
Energy Cutoff (ENCUT):
ENCUT = 520 # eV, typical for PAW potentials
k-Point Sampling:
# Method 1: Automatic mesh
KSPACING = 0.5 # Å⁻¹, automatic generation
# Method 2: Manual KPOINTS file
# Recommended density: 30-50 k-points per Å⁻¹
Precision (PREC):
PREC = Accurate # High, Normal, Accurate
Electronic Convergence (EDIFF):
EDIFF = 1E-6 # eV, energy convergence
# System description
SYSTEM = Cu bulk FCC
# Electronic minimization
ENCUT = 520 # Cutoff energy (eV)
EDIFF = 1E-6 # SCF convergence (eV)
NELM = 100 # Max electronic steps
ALGO = Fast # Algorithm: Normal, Fast, All
ISMEAR = 1 # Smearing: -5(tetra), 0(Gauss), 1(M-P)
SIGMA = 0.2 # Smearing width (eV)
# Precision
PREC = Accurate # Precision level
LREAL = Auto # Real-space projection
# Ionic relaxation
IBRION = 2 # 0=static, 1=RMM-DIIS, 2=CG
ISIF = 3 # 2=relax ions, 3=relax cell+ions
NSW = 100 # Max ionic steps
EDIFFG = -0.02 # Force convergence (eV/Å)
# Output
LWAVE = .FALSE. # Write WAVECAR
LCHARG = .FALSE. # Write CHGCAR
Cu FCC bulk
1.0 # Universal scaling
3.61 0.00 0.00 # Lattice vectors
0.00 3.61 0.00
0.00 0.00 3.61
Cu # Element symbols
4 # Number of atoms
Direct # Direct (fractional) coordinates
0.00 0.00 0.00
0.50 0.50 0.00
0.50 0.00 0.50
0.00 0.50 0.50
Key Points:
Gamma-Centered Mesh (most common):
Automatic mesh
0 # 0=automatic
Gamma # Gamma or Monkhorst-Pack
8 8 8 # k-point grid
0 0 0 # Shift
Monkhorst-Pack:
Automatic mesh
0
Monkhorst-Pack
8 8 8
0 0 0
Band Structure Path:
k-points for band structure
10 # Number of points between high-symmetry points
Line-mode # Line mode for band structure
Reciprocal
0.0 0.0 0.0 !Γ
0.5 0.0 0.5 !X
0.5 0.0 0.5 !X
0.5 0.25 0.75 !W
Generation:
# Concatenate POTCARs in same order as POSCAR
cat ~/vasp/potpaw_PBE/Cu/POTCAR > POTCAR
# For compounds:
cat ~/vasp/potpaw_PBE/Cu/POTCAR \
~/vasp/potpaw_PBE/O/POTCAR > POTCAR
Choosing POTCARs:
potpaw_PBE/Element/POTCARpotpaw_PBE.52/Element/POTCAR or potpaw_PBE.54/INCAR:
IBRION = 2 # Conjugate gradient
ISIF = 3 # Relax cell + ions
NSW = 100
EDIFFG = -0.02 # Force convergence
ISMEAR = 1 # Methfessel-Paxton
SIGMA = 0.2
Convergence Criteria:
EDIFFG < 0: Force-based (recommended: -0.01 to -0.05 eV/Å)EDIFFG > 0: Energy-based (less common)INCAR:
IBRION = -1 # No ionic updates
NSW = 0
ISMEAR = -5 # Tetrahedron (accurate DOS)
# OR
ISMEAR = 0 # Gaussian (if tetra not converged)
SIGMA = 0.05
Step 1: Self-consistent calculation
ICHARG = 2 # From atoms
LCHARG = .TRUE. # Write CHGCAR
Step 2: Non-self-consistent band structure
ICHARG = 11 # Read CHGCAR, no update
LORBIT = 11 # Write PROCAR
# Use line-mode KPOINTS
INCAR:
ISMEAR = -5 # Tetrahedron method
LORBIT = 11 # Projected DOS
NEDOS = 3000 # DOS resolution
# Use dense k-point mesh
INCAR:
IBRION = 0 # MD
NSW = 1000 # MD steps
POTIM = 1.0 # Time step (fs)
TEBEG = 300 # Start temperature (K)
TEEND = 300 # End temperature
SMASS = 0 # NVE: 0, NVT: >0
MDALGO = 2 # 1=Andersen, 2=Nose-Hoover
INCAR:
IBRION = 6 # DFPT for phonons
NFREE = 2 # Central differences
POTIM = 0.015 # Displacement (Å)
EDIFF = 1E-8 # Tight convergence!
INCAR:
IBRION = 6 # DFPT
ISIF = 3
NFREE = 4 # For elastic constants
HSE06:
LHFCALC = .TRUE. # Activate hybrid
HFSCREEN = 0.2 # HSE screening parameter
AEXX = 0.25 # Exact exchange fraction
ALGO = All # Or Damped
TIME = 0.4 # Damping for convergence
Step 1: DFT (PBE)
ALGO = Exact
NBANDS = 200 # Many empty bands
LOPTICS = .TRUE.
Step 2: GW
ALGO = GW0 # Or EVGW
NOMEGA = 50
INCAR:
LDAU = .TRUE.
LDAUTYPE = 2 # Dudarev
LDAUL = 2 -1 # l quantum number (d, s/p)
LDAUU = 5.0 0.0 # U value (eV)
LDAUJ = 0.0 0.0 # J value
DFT-D3:
IVDW = 11 # DFT-D3 (Grimme)
vdW-DF:
GGA = MK # optPBE-vdW
LUSE_VDW = .TRUE.
AGGAC = 0.0000
# Test sequence
KPOINTS: 4x4x4, 6x6x6, 8x8x8, 10x10x10, 12x12x12
# Converged when ΔE < 1 meV/atom between successive grids
# Test ENCUT
ENCUT: 400, 450, 500, 550, 600 eV
# Converged when ΔE < 1 meV/atom
# Forces may need higher cutoff
| ISMEAR | Method | Use Case |
|---|---|---|
| -5 | Tetrahedron | Static calcs, DOS, accurate energies |
| -4 | Tetrahedron+Blöchl | Like -5, slightly different |
| -1 | Fermi smearing | Metals |
| 0 | Gaussian | General purpose |
| 1+ | Methfessel-Paxton order N | Relaxations, metals |
Recommendations:
# INCAR
ENCUT = 520
PREC = Accurate
IBRION = 2
ISIF = 3
NSW = 100
EDIFFG = -0.02
ISMEAR = 1
SIGMA = 0.2
ALGO = Fast
LREAL = Auto
# KPOINTS
Gamma-centered
0
Gamma
8 8 8
0 0 0
# INCAR
ENCUT = 600 # Higher cutoff
PREC = Accurate
IBRION = -1
NSW = 0
EDIFF = 1E-8 # Tight convergence
ISMEAR = -5 # Tetrahedron
ALGO = Normal
LREAL = .FALSE. # Reciprocal space
# KPOINTS (very dense)
0
Gamma
12 12 12
0 0 0
# INCAR
ENCUT = 400 # Lower cutoff
PREC = Normal
EDIFF = 1E-4 # Loose
ISMEAR = 0
SIGMA = 0.1
ALGO = Fast
LREAL = Auto
# KPOINTS (coarse)
0
Gamma
4 4 4
0 0 0
INCAR:
NCORE = 4 # Cores per band (orbital parallelization)
# OR
NPAR = 8 # Number of groups for band parallelization
KPAR = 4 # k-point parallelization
LPLANE = .TRUE. # Plane-wise distribution
Guidelines:
LREAL = Auto # Reduce memory for large systems
NCORE = 4 # Reduce memory per core
"ZBRENT: fatal error in bracketing"
# Fix: Reduce POTIM or use different IBRION
POTIM = 0.2
"EDDDAV: X eigenvalues not converged"
# Fix: Increase NELM, change ALGO
NELM = 200
ALGO = All
"Sub-Space-Matrix is not hermitian"
# Fix: Reduce POTIM, check structure
POTIM = 0.1
SYMPREC = 1E-8
SCF not converging:
# Try sequential fixes:
1. ALGO = All
2. Increase NELM = 200
3. AMIX = 0.2, BMIX = 0.0001
4. Check initial structure (too close atoms?)
Invoke specific subskills for detailed guidance:
What type of calculation?
| Goal | IBRION | ISIF | ISMEAR | EDIFFG |
|---|---|---|---|---|
| Relax ions only | 2 | 2 | 1 | -0.02 |
| Relax cell+ions | 2 | 3 | 1 | -0.02 |
| Static energy | -1 | 2 | -5 | N/A |
| MD simulation | 0 | 2 | 0 | N/A |
| Band structure | -1 | 2 | 0 | N/A |
| Phonons (DFPT) | 6 | 2 | 0 | N/A |
materials-properties skill - For ASE-based workflows with VASPexamples/ directoryreferences/ directoryStructured planning and project breakdown using proven methodologies for goals, projects, and strategic initiatives. Use when users need to create plans, break down complex projects, set milestones, estimate timelines, identify dependencies, or develop action plans. Triggers include 'help me plan,' 'create a roadmap for,' 'break down this project,' 'what are the steps to,' or 'how should I approach.'
Expert guidance for writing professional Python code following industry best practices including PEP 8 compliance, testing, type hints, error handling, and modern tooling. Use this skill when writing new Python code, refactoring existing code, setting up Python projects, implementing tests, or ensuring code quality and maintainability. Emphasizes: PEP 8, modularity, DRY principle, TDD, virtual environments (uv), and modern tooling (Ruff, Black, Mypy).
Conduct systematic scientific literature reviews by integrating multi-source searches, iterative analysis, and synthesis into well-structured reports with proper citations. Use this skill when you need to: synthesize knowledge across multiple research domains, identify research gaps, track the evolution of methodologies, evaluate evidence quality, or write comprehensive literature reviews for academic papers, grant proposals, or research projects. Key capabilities: multi-database search strategies, iterative query refinement, content analysis and summarization, thematic synthesis, citation management, quality assessment, bias detection, and structured report generation with multiple output formats.
SOC 職業分類に基づく
