Simulink Agentic Toolkit gives AI agents the tools and the expertise to work effectively with Simulink models. It builds on top of the MATLAB MCP Core Server and connects AI agents to Simulink using the Model Context Protocol (MCP). Purpose-built MCP tools let AI agents drive interaction with Simulink. Curated skills encode best practices for Model-Based Design and guide AI agents to the right tools for design, debugging, and testing. The toolkit ships configurations for coding agents such as Claude® Code, GitHub Copilot®, OpenAI® Codex, Gemini CLI®, and Sourcegraph Amp.
Draft and Trace Requirements
Your AI agent can draft requirements and link each one to the model elements that implement it—whether you are starting from a high-level specification or documenting an existing model. The AI agent helps maintain the traceability as requirements and model evolve, so you spend less time updating links and documents and more time on the design. Connect to Requirements Toolbox™ to carry each requirement through to implementation and verification.
Specify and Implement Designs
Your AI agent can modify existing models with a targeted fix or create new systems such as a control algorithm or plant model. For complex work, the AI agent can propose a specification and plan for your review, so design choices are settled early instead of discovered late through rework. Build multi-domain systems, including Stateflow® for decision logic, Simscape™ for physical systems, and System Composer™ for system and software architectures.
Simulate and Analyze Results
Your AI agent can orchestrate simulations so you can focus on what the results mean, not on how to set them up. The AI agent handles inputs, configurations, and performance best practices like fast restart and parallel execution with Parallel Computing Toolbox™. From a single scenario to thousands of runs, the AI agent returns logged signals as plots and comparisons ready for review—so that you stay focused on the design tradeoffs.
Troubleshoot and Verify Changes
Your AI agent can reproduce a reported issue, isolate the root cause, and write a test using Simulink Test™ that captures the issue. The AI agent then proposes a fix, applies it once you approve, and reruns the test to verify the change. You review findings at each step, keeping engineering decisions in your hands while the agent takes on the testing and verification cycles.
Purpose-Built for AI Agents
Simulink Agentic Toolkit is built for how AI agents work. MCP tools give the AI agent structured access to your model — architecture, signal flow, block parameters, and simulation behavior — so it operates on the model, not just on text. Skills encode the right workflow for each task, so the AI agent follows a recommended path instead of improvising. Since the AI agent reads only what it needs at the right depth, the toolkit scales to the models you work with.
Bring Your Own AI Agent
Use your preferred AI agent with an automated setup process, which configures MATLAB® paths, enables MCP attachment, adds tools and skills, and validates your installation. Clone the repository, ask your AI agent to set up the toolkit, and it will be ready to work with Simulink.
Download the Simulink Agentic Toolkit
