Open-source Agent Skills for computational materials science and numerical simulation workflows.

Give your AI coding agent domain expertise in numerical methods, simulation best practices, and scientific computing -- without re-explaining the same concepts every session. Skills are portable across Claude Code, Codex, Gemini CLI, Cursor, VS Code Copilot, and 20+ other compatible tools.

• The Problem
• The Solution
• What's Inside
  • Core Numerical Skills
  • Simulation Workflow Skills
  • HPC Deployment Skills
  • Ontology Skills
• How Skills Work
• Quick Start
• Adding Skills to Your Agent
  • Claude Code
  • Gemini CLI
  • OpenAI Codex
  • VS Code / GitHub Copilot
  • Cursor
  • Other Agents
• Repository Layout
• Contributing
• License
• Acknowledgements

Simulation engineers repeat the same guidance to AI agents constantly: "Check the CFL number before running," "Use Richardson extrapolation for grid convergence," "Exit code 2 means bad input." General-purpose agents lack the domain knowledge to run reliable numerical simulations without heavy hand-holding.

This project packages that domain knowledge into skills -- structured folders of instructions, scripts, and references that agents discover automatically and load on demand. Each skill teaches an agent a specific simulation competency, backed by validated Python scripts that produce reproducible results.

You: Check if dt=0.001 is stable for my advection problem with v=2.0 m/s and dx=0.01.

Agent: I'll use the numerical-stability skill to check this.
       [runs cfl_checker.py --dx 0.01 --dt 0.001 --velocity 2.0 --json]

       CFL number = 0.2 (limit: 1.0) -- stable.
       Recommended max dt = 0.005 with safety factor 1.0.

No prompt engineering. No copy-pasting formulas. The agent finds the right skill, runs the script, and interprets the results.

17 skills | 67 scripts | 857 tests | Cross-platform CI on Python 3.10-3.12

Foundational numerical methods and analysis tools.

End-to-end simulation management and automation.

Deployment and job submission tooling for running simulations on HPC systems.

Materials science ontology understanding, mapping, and validation.

Skills follow the open Agent Skills standard. Each skill is a folder with three tiers of content, loaded progressively to keep context efficient:

skills/core-numerical/numerical-stability/
    SKILL.md              # Instructions + YAML metadata (loaded when skill triggers)
    scripts/              # Python CLI tools (executed for reproducible results)
        cfl_checker.py
        von_neumann_analyzer.py
        matrix_condition.py
        stiffness_detector.py
    references/           # Deep domain knowledge (loaded only when needed)
        stability_reference.md

1. Discovery -- The agent sees each skill's name and description at startup (~100 tokens per skill)
2. Activation -- When a task matches, the agent loads the full SKILL.md with decision guidance, workflows, and CLI examples
3. Execution -- Scripts run as subprocesses with --json output for structured, parseable results

All scripts are standalone CLI tools with --help, a pure-function core for testing, and consistent error handling (exit code 2 for bad input, 1 for runtime errors).

git clone https://github.com/heshamfs/materials-simulation-skills.git
cd materials-simulation-skills
pip install -r requirements-dev.txt

python -m pytest tests/ -v --tb=short          # All 857 tests
python -m pytest tests/unit -v --tb=short       # Unit tests only
python -m pytest tests/integration -v           # Integration tests only

These skills follow the open Agent Skills standard and work across 20+ AI coding tools. Choose your agent below.

Copy individual skills (or the whole skills/ tree) into your personal or project skills directory:

# Personal (available across all projects)
cp -r skills/core-numerical/numerical-stability ~/.claude/skills/numerical-stability

# Project-level (committed to version control)
cp -r skills/core-numerical/numerical-stability .claude/skills/numerical-stability

Or clone the whole repo and point Claude Code at it with --add-dir:

claude --add-dir /path/to/materials-simulation-skills/skills

Verify with: What skills are available? or type / to see skills in the autocomplete menu.

See the Claude Code skills docs for more details.

Install directly from the repo, or copy skills into your Gemini skills directory:

# User-scoped (available across all workspaces)
cp -r skills/core-numerical/numerical-stability ~/.gemini/skills/numerical-stability

# Workspace-scoped (project-specific)
cp -r skills/core-numerical/numerical-stability .gemini/skills/numerical-stability

Verify with: gemini skills list

See the Gemini CLI skills docs for more details.

Copy skills into one of the Codex skills directories:

# User-scoped
cp -r skills/core-numerical/numerical-stability ~/.agents/skills/numerical-stability

# Repository-scoped
cp -r skills/core-numerical/numerical-stability .agents/skills/numerical-stability

Restart Codex after adding skills. Use /skills or $ to invoke skills by name.

See the Codex skills docs for more details.

Copy skills into your workspace or personal skills directory:

# Workspace (committed to version control)
cp -r skills/core-numerical/numerical-stability .github/skills/numerical-stability

# Personal (across all workspaces)
cp -r skills/core-numerical/numerical-stability ~/.copilot/skills/numerical-stability

Type /skills in the chat input to see and invoke available skills.

See the VS Code skills docs for more details.

Copy skills into a skills/ directory at your project root:

cp -r skills/core-numerical/numerical-stability skills/numerical-stability

Cursor's MCP server auto-discovers skills from the skills/ directory.

Any agent that supports the Agent Skills standard can use these skills. The general pattern:

1. Copy the skill directory (containing SKILL.md, scripts/, references/) into your agent's skills folder
2. The agent discovers the skill by its name and description in the YAML frontmatter
3. Mention the skill by name or ask a task that matches its description

Use numerical-stability to check a proposed dt for my phase-field run.

The agent loads the skill's instructions, runs the appropriate scripts, and interprets the results.

skills/
    core-numerical/          # 8 skills: stability, solvers, meshing, convergence, ...
    simulation-workflow/     # 5 skills: validation, optimization, orchestration, ...
    hpc-deployment/          # 1 skill: SLURM job script generation
    ontology/                # 3 skills: ontology exploration, mapping, validation
tests/
    unit/                    # Pure-function tests via load_module()
    integration/             # Subprocess + JSON schema validation
    fixtures/                # Sample data files for CI smoke tests
.github/
    workflows/ci.yml         # Cross-platform CI (Ubuntu, Windows, macOS)
    ISSUE_TEMPLATE/          # Bug reports, skill proposals
    PULL_REQUEST_TEMPLATE.md # PR checklist

We welcome contributions of all kinds -- new skills, bug fixes, documentation, and tests. The project is designed to grow from 17 skills across 4 categories into a broader collection spanning materials physics, verification & validation, HPC deployment, and more.

See CONTRIBUTING.md for:

• Step-by-step guide to creating a new skill
• Script and test templates
• Skill taxonomy and open categories for community contributions
• PR guidelines and checklists

Apache 2.0

• Agent Skills standard -- Open specification for portable agent capabilities
• Anthropic -- Original developer of the Agent Skills format
• agentskills/agentskills -- Reference implementation and validation library