DeepSeek Harness Plugin Hub

Publish and manage complete Harness Profiles. Discover Plugins for your next setup.

Explore

PluginsPresetsDocsNews

Community

Publish a pluginContactReport an issue

Resources

Plugin Hub on GitHubDeepSeek HarnessSystem statusPrivacy notice
© 2026 DeepSeek Harness Plugin HubPowered byPaxTech

Independent and unofficial. Not affiliated with, authorized by, or endorsed by DeepSeek.

DeepSeek Harness Plugin Hub
ProfilesPluginsCategoriesNewsDocsSign inManage Profiles
ProfilesPluginsCategoriesNewsDocsSign in
← Plugins
M

dsh-matlab-control

Matlab Control

MATLAB/Simulink control-engineering assistant for DeepSeek Harness: reuses the MathWorks MATLAB MCP Server (with the Simulink Agentic Toolkit tool registry) and adds Simulink model checks, simulation result summaries, App Designer UI generation, and control-design analysis.

The plugin will be installed here. Keep web if you are unsure.

npx -y @deepseek-ai/dsh plugin --profile web add github:reisen0327-sudo/dsh-matlab-control#bcd2c7b1ac6672b0516fc401e7a3bac4c0932649
READMECompatibilityVersions

Compatibility and provenance

Matlab Control is published as dsh-matlab-control and currently resolves to version 0.1.0. The Hub verifies its manifest and preserves the exact installation source for reproducible installs.

DSH compatibility
*
Runtime surfaces
any
Release source
github
Registry updated
11/4/2026

Versions

0.1.0stable
11/4/2026

Related plugins

Loading related plugins…

Latest
0.1.0
DSH
*
HMR
Process restart
Tree shaking
Safe tree shaking not declared
Unpacked size
Unavailable
Files
Unavailable
Surface
any
License
MIT
Source
github
GitHub
★ 1
Weekly downloads
0
Last push
11/4/2026
View source ↗
README badge

Click the badge to copy Markdown for your README.

Do you maintain this Plugin?Claim benefit · Priority security scan

Verify the GitHub repository declared in package.json to manage this listing. After you claim it, Hub will prioritize a security scan of the current version and publish the result when it passes.

Claim this Plugin →
Report an issue

Related plugins

More verified plugins in developer-tools.

Web App@deepseek-ai/dsh-web-appThe dsh browser-surface bundle: the web patch layer over dsh-base plus the runtime glue plugin (frontend dist serving, web-surface prompt, bash runtime variables, URL line)Sdk Minimal@deepseek-ai/dsh-sdk-minimalThe standalone minimal SDK profile bundle: JSON-RPC, one DeepSeek adapter, persistent shell, and JSONL sessionsSdk App@deepseek-ai/dsh-sdk-appThe dsh SDK profile bundle: stdio JSON-RPC serving and process lifecycle over dsh-baseExperimental Inspector@deepseek-ai/dsh-experimental-inspectorExperimental cross-realm CDP hub for Host debugging and Client Runtime inspection

README

dsh-matlab-control

MATLAB/Simulink control-engineering assistant for DeepSeek Harness (DSH). English | 中文

This plugin does not implement a MATLAB/Simulink transport. It reuses the official MathWorks MATLAB MCP Server as the bridge and the official Simulink Agentic Toolkit tool registry for model tooling, then adds four agent skills that specialise the workflow for Simulink simulation, App Designer, and control-system design.

  • 14 model-facing tools: 5 MATLAB (evaluate_matlab_code, run_matlab_file, check_matlab_code, run_matlab_test_file, detect_matlab_toolboxes) + 9 Simulink (model_overview, model_read, model_edit, model_check, model_query_params, model_resolve_params, model_read_diagnostics, model_scan, model_test).
  • All tools appear to the agent as mcp__matlab__<tool>.
  • Domain knowledge lives in skill files, not in a system prompt.

Features

AreaWhat the agent does
Simulink simulationReads model structure (model_overview, model_read), validates connectivity and Edit-Time checks (model_check), runs simulations and reports signal/metric summaries instead of raw command output
App Designer / GUIGenerates modern uifigure/uiaxes/uidropdown/uibutton app classes, wires callbacks that call sim() and plot into uiaxes, and verifies them with check_matlab_code
Control engineeringRoutes an intent to bode/nyquist/margin/allmargin, step/impulse/lsim, ss/tf/zpk, pidtune/pidTuner, robuststab/robgain, and returns gain/phase margin, overshoot, settling time, steady-state error and stability as a structured summary

How it works

cordis.patch.yml is the entire runtime — two rows, no JavaScript:

RowPackagePurpose
mcp-matlab@deepseek-ai/dsh-mcp-clientLaunches the MathWorks MATLAB MCP Server over stdio and registers its tools as mcp__matlab__*
matlab-control-skills@deepseek-ai/dsh-skill-filesystemRegisters skills/ as a skill source for every agent session

The Simulink tool surface comes from the official Simulink Agentic Toolkit registry, passed to the same MCP server with --extension-file. Neither the toolkit nor the MCP server binary is bundled in this repository; both are MathWorks deliverables and stay under their own terms.

Requirements

  • MATLAB R2023a or later with Simulink (the toolkit requires R2023a+; tested with R2024b).
  • MATLAB MCP Server binary, v0.13.0 or later. The MATLAB Agentic Toolkit installs it at <user>\.matlab\agentic-toolkits\bin\matlab-mcp-server.exe on Windows.
  • MATLAB MCP Server Toolbox, installed once per MATLAB release: matlab-mcp-server.exe --setup-matlab.
  • Optional products: Control System Toolbox (pidtune, allmargin), Simulink Test (model_test), Robust Control Toolbox (robuststab, robgain).

Installation

  1. Get the official Simulink Agentic Toolkit into the plugin directory (not tracked here):

    git clone --depth 1 https://github.com/matlab/simulink-agentic-toolkit.git \
      thirdparty/simulink-agentic-toolkit
    
  2. Point the bundle at your machine. Edit these values in cordis.patch.yml:

    FieldSet it to
    commandabsolute path of matlab-mcp-server.exe
    --matlab-root=your MATLAB installation folder (the one that contains bin), no trailing \bin
    --extension-file=<plugin dir>/thirdparty/simulink-agentic-toolkit/tools/tools.json
    customSkillDirs<plugin dir>/skills
  3. Install the bundle into your DSH profile:

    plugin_manager { action: "install_bundle", target: "<absolute path to this directory>" }
    

    A local-directory install links the checkout instead of copying it, so the absolute paths from step 2 stay valid as long as the directory does not move.

  4. Start a new session; the mcp__matlab__* tools and the four skills are then available.

Session mode

The example configuration omits --matlab-session-mode, which selects the MathWorks default auto: attach to a MATLAB session that ran shareMATLABSession(), otherwise start one. To bind strictly to your desktop MATLAB session, add --matlab-session-mode=existing to args after running shareMATLABSession() there.

Skills

SkillTriggers onDoes
simulink-model-check"check this model", "what's wrong with it"model_overview → model_check → per-issue explanation and a verdict
simulink-simulation-run"run the simulation", "show the response"Simulink.SimulationInput + sim() via evaluate_matlab_code, structured signal and metric summary
app-designer-ui"generate .mlapp code", "build a UI"App class with modern UI components, sim() callback plotting into uiaxes, verified with check_matlab_code
control-design"tune a PID", "Bode plot", "state space", "robustness"Intent routing to the right analysis functions, structured closed-loop metrics

Each skill body names the official SKILL.md to read before acting; see reference/official-skills-index.md. The toolkit's own skills are not copied into this repository — read them from your clone.

Example prompts

Check this Simulink model and tell me what is wrong with it: <path>.slx
Run demo_plant for 10 s and summarise overshoot, settling time and final value.
Generate a small App Designer app that runs a model and plots the result in a uiaxes.
Tune a PID for tf(1,[1 1]) and verify the closed loop in Simulink.

Repository layout

dsh-matlab-control/
├── cordis.patch.yml                 # the whole bridge + skill registration
├── package.json                     # bundle manifest (dsh.bundle.patch)
├── icon.svg, locale/{en,zh}.json    # plugin-list title, description, icon
├── skills/<name>/SKILL.md           # four specialisation skills
├── reference/official-skills-index.md
├── README.md, README.zh.md
└── LICENSE

thirdparty/ (your toolkit clone) and local test artifacts are ignored by git.

Troubleshooting

SymptomCause / fix
First tool call takes 1–3 minutesThe MCP server is starting MATLAB. Later calls are fast.
model_* tools report "function not found"The toolkit is not on the MATLAB path for this session. Run addpath(genpath('<plugin>/thirdparty/simulink-agentic-toolkit')) in MATLAB (evaluate_matlab_code also works).
pidtune / robuststab undefinedMissing Control System / Robust Control Toolbox. Check with detect_matlab_toolboxes.
No mcp__matlab__* tools after installThe MATLAB MCP Server binary path in cordis.patch.yml is wrong, or the MCP Server Toolbox is not installed for that MATLAB release.
application: restart-required after installDSH hot-applies the rows; restart once to clear the flag and load the bundle through the normal plugin list.

Status and roadmap

Tested on MATLAB R2024b with Simulink and Control System Toolbox. This is a minimal first version: model search/batch processing, report export and multi-session MATLAB pooling are intentionally not included yet.

License

MIT — see LICENSE. The MATLAB MCP Server, the MathWorks MATLAB MCP Server Toolbox and the Simulink Agentic Toolkit are MathWorks software and are not covered by this license; they are distributed under their own terms and are not bundled here.