We've just released a free set of agent skills that teach AI coding assistants—Claude Code, Codex or any agent with command-line access—to work with Wolfram System Modeler.

With the skills installed, your assistant can create a model from a plain-language description, validate that it compiles, simulate it, plot the results, diagnose why a model is slow or failing and even answer Modelica questions with citations from the language specification and the System Modeler documentation. For post-processing—parameter sweeps, calibration, machine learning on results—it hands off to Wolfram Language via Wolfram MCP.

The following video shows the idea end to end: from a hand-drawn sketch to a validated, simulation-ready model, with a single prompt.

Everything runs against your local System Modeler installation.

Get the skills and setup instructions here: https://github.com/WolframResearch/system-modeler-ai-toolkit