Openfoam Agent

name: openfoam-agent description: OpenFOAM CFD workflow playbook for AI coding assistants. Use to plan and implement OpenFOAM cases, choose solvers, generate/validate meshes, run/debug simulations, and standardize post-processing. argument-hint: "[task] [case_path(optional)]" license: Complete terms in LICENSE.txt

OpenFOAM Agent Skill

This skill is an AI-facing OpenFOAM manual.

When invoked, follow the workflow below.

0) First clarify (don’t guess)

Ask for the minimum set of facts required to avoid non-physical or unstable setups:

• Physics: incompressible/compressible, single/multiphase, heat transfer/CHT, reacting?
• Regime: laminar/RANS/LES, steady/transient
• Geometry & scale: key dimensions, 2D vs 3D assumption
• Fluid properties: ρ, ν (or μ), cp, k, etc.
• Boundary conditions: inlets/outlets/walls/symmetry, turbulence quantities if RANS
• Outputs: forces, pressure drop, temperature, y+, probes
• Resources: target runtime, cores, memory

If any item is unknown, propose 1-3 concrete options and request confirmation.

1) Use standard case structure

Enforce the canonical OpenFOAM layout:

• 0/: initial conditions and boundary fields
• constant/: mesh and physical properties
• system/: numerics and runtime controls
• Allrun / Allclean: reproducible run scripts (recommended)

Detailed reference: reference/case_layout.md

Boundary condition baselines: reference/boundary_conditions.md

2) Mesh workflow (always validate)

• Pick the path:
  • blockMesh (structured / simple)
  • snappyHexMesh (STL-based)
  • gmshToFoam (Gmsh .msh)
• Always run checkMesh and address quality issues before solver runs.
• Ensure boundary patch names match what 0/ expects.

Detailed reference: reference/meshing.md

3) Solver & turbulence selection (engineering defaults)

• Prefer the simplest solver that matches physics.
• For RANS, default to k-omega SST unless there’s a reason not to.
• Target version is ESI/OpenCFD OpenFOAM v2412. Ensure run commands match the installed environment.

Detailed reference: reference/solver_selection.md

Numerics templates: reference/fv_solution_patterns.md

4) Run, monitor, and iterate

• Start with conservative numerics (time step / relaxation) and then tighten.
• Track:
  • residual trends
  • continuity errors
  • CFL / Co number (transient)
  • forces / Δp stabilization
• Prefer small, explainable changes per iteration.

Debug playbook: reference/troubleshooting.md

Parallel workflow: reference/parallel_run.md

5) Post-processing as first-class output

• Use functionObjects / postProcess for repeatable outputs.
• Output concise, decision-oriented summaries (e.g., drag coefficient, Δp, max T, y+ stats).

Cheatsheet: reference/command_cheatsheet.md

6) Repository-local helper scripts

• scripts/case_audit.py: static audit of a case directory (structure + presence of key dictionaries)
• scripts/skill_quality_check.py: validate skill docs (broken links + length limits)

Additional resources

• reference/windows_wsl.md
• reference/version_notes_v2412.md
• reference/dictionary_syntax.md
• reference/solver_log_reading.md
• reference/openfoam_cpp_dev.md
• reference/boundary_conditions.md
• reference/fv_solution_patterns.md
• reference/parallel_run.md
• reference/quality_and_maintenance.md