Hpc Fenics

Build, review, and debug FEniCS or DOLFINx PDE scripts for finite-element workflows. Use when translating PDEs into UFL, selecting function spaces, applying...

Security Scan

The skill's files, instructions, and templates are consistent with its stated purpose of building, reviewing, and debugging FEniCS/DOLFINx PDE scripts; it does not request unrelated credentials or install arbitrary code.

Purpose & Capability

The name/description match the included reference docs and example templates for FEniCS/DOLFINx workflows; nothing required by the skill (no env vars, no unusual binaries) is unrelated to the stated HPC FEM purpose.

Instruction Scope

SKILL.md directs the agent to read internal reference files and use bundled templates; instructions do not request reading arbitrary system files, credentials, or sending data to external endpoints. The guidance is narrowly scoped to PDE formulation, solver patterns, and cluster execution.

Install Mechanism

No install spec is provided (instruction-only), so nothing is downloaded or installed by the skill itself. The included example scripts assume the user environment has dolfinx/mpi4py/petsc4py available — a runtime dependency but not an install action by the skill.

Credentials

The skill requests no environment variables or credentials. It references common cluster SLURM environment variables in the example job script (expected for cluster use) and relies on Python packages (dolfinx, mpi4py, petsc4py) being present — these are runtime dependencies, not secrets.

Persistence & Privilege

always is false and the skill does not request persistent privileges or modify other skills or system-wide settings. It is user-invocable and uses normal autonomous invocation defaults, which is expected.

Assessment

This skill appears coherent and focused on FEniCS/DOLFINx help. Before using: (1) review the provided Python and SLURM templates (do not run them unreviewed on production systems), (2) ensure your environment has the required libraries (dolfinx, mpi4py, petsc4py) and MPI/PETSc stack, and (3) when submitting job scripts to a cluster, confirm paths and module/container setup so you don't accidentally run with missing dependencies or write outputs to unintended storage. If you need higher assurance, run the examples in a disposable container or development partition first.

Treat FEniCS as a family with two main stacks: classic FEniCS and DOLFINx.

Start

1. Read references/runtime-selection.md first.
2. Read references/dolfinx-boundary-workflow.md when targeting DOLFINx boundary conditions.
3. Read references/ufl-and-solver-patterns.md when translating PDEs, choosing spaces, or building forms.
4. Read references/pde-template-cookbook.md when mapping common PDE classes to known formulation patterns.
5. Read references/time-dependent-and-nonlinear-patterns.md when building transient or nonlinear solvers.
6. Read references/implementation-skeletons.md when you need a concrete script shape for classic FEniCS or DOLFINx.
7. Read references/petsc-solver-playbook.md when choosing linear, nonlinear, or block solver settings.
8. Read references/mixed-problems-and-output.md when working with mixed spaces, XDMF, or post-processing.
9. Read references/cluster-execution-playbook.md when staging a FEniCS or DOLFINx script for scheduler-backed cluster execution.
10. Read references/boundary-io-and-errors.md when repairing runtime failures or resolving IO issues.

Work sequence

1. Choose one stack and stay consistent:
   • classic FEniCS: fenics or dolfin
   • modern stack: dolfinx
2. Translate the PDE into the correct weak form before writing code.
3. Match unknown type and space:
   • scalar field -> scalar space
   • vector field -> vector space
   • mixed formulation -> mixed space with explicit subspace handling
4. For linear problems, use TrialFunction and TestFunction.
5. For nonlinear problems, use a Function for the unknown and solve a residual form.
6. Write outputs in a format the expected post-processor can open, usually XDMF for mesh-coupled fields.

Guardrails

• Do not mix classic FEniCS imports with DOLFINx APIs in one script.
• Do not use TrialFunction for a nonlinear unknown.
• Do not guess a boundary condition if the PDE is under-constrained; say what is missing.
• Do not ignore shape and rank mismatches in UFL expressions.

Additional References

Load these on demand:

• references/gmsh-meshtags-and-refinement.md for imported meshes, physical groups, markers, and refinement transfer
• references/io-visualization-and-writers.md for XDMF, VTK, VTX, and visualization-safe output
• references/parallel-and-mpi-caveats.md for ownership, boundary marking, and parallel consistency
• references/space-boundary-output-matrix.md for function-space, boundary-condition, and writer-selection matrices
• references/cluster-execution-playbook.md for MPI launch, environment pinning, and cluster-side IO planning

Reusable Templates

Use assets/templates/ when a concrete script scaffold is needed, especially:

• poisson_dolfinx.py
• transient_diffusion_dolfinx.py
• fenics-dolfinx-slurm.sh

Outputs

Report:

• chosen stack and version family if known
• PDE form and space selection
• boundary conditions applied
• expected output files
• the exact failure class if the script is being repaired