Computational modeling requires scientific and methodological judgment that extends beyond software engineering, but general-purpose AI coding agents do not consistently apply established modeling practices. OMF Skills encode community-developed modeling methodology, research software practices, and scientific workflows, helping computational modelers at any career stage build models that are transparent, reviewable, reproducible, and FAIR-aligned.
Status: Alpha — We are actively seeking community feedback to refine these workflows.
This repository is designed for researchers, research software engineers, students, and practitioners at any career stage developing computational models.
Agent Skills capture specialized methodological knowledge that foundation models do not consistently possess. This repository follows a "shrink-to-fit" philosophy: as foundation models improve, skills are simplified or removed, preserving only frontier community knowledge, standards, workflows, and practices that models do not consistently possess or reliably execute.
- A coding-capable AI agent (e.g., Cursor, Claude Code, Windsurf, etc.)
- Node.js LTS (see docs/install.md)
npx skills add omf/skillsAsk your agent: "Install all skills from https://github.com/openmodelingfoundation/skills"
Confirm your agent has loaded the skills:
List all installed skills and summarize when each should be used.
An animated walkthrough will be added here in a future release.
Example: /omfa help me build a spatially explicit ABM of wildfire evacuation
The /omfa skill guides you through a structured modeling lifecycle and produces:
- Conceptual model: entities, processes, and system boundaries.
- Assumptions list: explicit, reviewable modeling choices.
- Implementation plan: modular structure with validation activities.
- Recommendations: when to use specialist skills for documentation, FAIR metadata, HPC orchestration, or peer-review readiness.
.
├── skills/ # Published agent skills (e.g., omfa, document, fair)
├── docs/ # Contributor documentation
├── evals/ # Evaluation framework
├── scripts/ # Repository tooling
└── AGENTS.md # Repository operating contract
We welcome new skills, improvements to existing workflows, and evaluation cases.
- Contribution Workflow: CONTRIBUTING.md
- Authoring Reference: docs/agent-skills-creation-reference.md
- New Skill Template: docs/SKILL-TEMPLATE.md
| Available | Planned |
|---|---|
Modeling Methodology (/omfa) |
Statistical Analysis |
Narrative Documentation (/document) |
Calibration & Validation |
FAIR Metadata (/fair) |
Reproducibility Automation |
HPC/HTC Orchestration (/hpc, /ospool) |
Domain-specific Modeling Templates |
Peer Review Readiness (/peer-review) |
Visualization Standards |
- Open Modeling Foundation: openmodelingfoundation.org
- Agent Skills Specification: agentskills.io
- Skills CLI: github.com/vercel-labs/skills
All skills in this repository are licensed under the MIT License.