AI Agent Hub
Back to plugins
🖥️

dsh-matlab-mcp

Client Updated 2026.09.05

Run the following command in DeepSeek Harness:

dsh plugin install James-web11/dsh-matlab-mcp

Paste the following prompt into your AI chat to install this plugin:

Run dsh plugin install James-web11/dsh-matlab-mcp in DeepSeek Harness to install this plugin; the source repository is https://github.com/James-web11/dsh-matlab-mcp

About this plugin

In DeepSeek Harness agent sessions, MATLAB and Simulink have always been a black box: an agent could generate code snippets but had no way to actually execute, validate, or modify a running model. dsh-matlab-mcp plugs the official MathWorks MATLAB MCP Server into the dsh GUI, giving every agent session a native set of mcp__matlab__* tools that turn MATLAB computation and Simulink modeling into first-class, callable primitives.

The capability spans two layers. On the MATLAB side, an agent can run arbitrary code, execute scripts and unit tests, perform static analysis on .m files, and query the running version along with installed toolboxes. On the Simulink side, it can inspect the hierarchical model view, read block topology and expression notation, make structural edits and parameter changes, validate connectivity, query any block configuration parameter on demand, resolve workspace variables to concrete values, and even run Gherkin-based behavioral tests. The bundle is a pure configuration package with paths resolved at config time, so Windows, macOS, and Linux all work without any per-machine editing.

It is aimed at engineers and researchers already running MATLAB R2021a or later with Simulink installed, especially those who want their dsh agent workflow to drive MATLAB computations, batch-edit Simulink models, or let an agent autonomously perform code review and behavioral verification.

Use Cases

  • Execute MATLAB code, run scripts, and perform static analysis directly within an agent session
  • Let an agent read, edit, and validate Simulink model structure, resolving workspace variables on demand
  • Autonomously run unit tests and Gherkin-based behavioral tests to close the verification loop

Best For

  • Engineers and researchers working with MATLAB R2021a+ and Simulink
  • Teams integrating scientific computing and simulation into DeepSeek Harness agent workflows
  • Automation practitioners who want AI agents to drive MATLAB and Simulink independently