Mechatronics Engineering plugin

Plugin version 0.1.1, containing the unchanged 0.1.0 skill. MIT licensed. The package includes all 52 skill files: engineering references, Python calculations, browser workbench, block editor, OpenUSD export and Scilab/Xcos/ATOMS/ROS 2 interface templates.

Install from GitHub in ChatGPT desktop / Codex

With the Codex CLI installed, run:

codex plugin marketplace add ShaunPrice/mechatronics-skill --ref v0.1.1

Restart the ChatGPT desktop app. In Work or Codex, open the Plugins Directory, select the Personal marketplace source, find Mechatronics Engineering, and select Install. Start a new conversation after installation.

The repository uses the plugin creator's default marketplace identifier, personal. If a marketplace with that identifier is already configured, inspect codex plugin marketplace list before adding this one; do not replace an unrelated catalog. The local-folder route below provides the same package for inspection or integration into your existing catalog.

On CLI versions offering codex plugin add, the installation step can also be run as:

codex plugin add mechatronics-engineering@personal

Check codex plugin list --json for the installed/enabled state. Marketplace registration, plugin installation, and successful use are separate checks. Commands do not require API keys or paid API billing.

Download and install from a local folder

1. Download mechatronics-marketplace-0.1.1.zip and extract it. It contains the catalog and complete plugin.
2. In a terminal, change into the extracted mechatronics-marketplace folder.
3. Run codex plugin marketplace add . and use the Plugins Directory installation step above.

For tooling that asks for just a plugin folder, download mechatronics-engineering-plugin-0.1.1.zip and extract it. Its mechatronics-engineering folder contains plugin.json, compatibility manifests, skills/, README and LICENSE. The plugin ZIP and the standalone .skill file have different layouts; choose the artifact requested by your installer.

ChatGPT web and mobile

GitHub distribution does not automatically publish a plugin into ChatGPT's universal directory or upload it to your account. Local/repository marketplace availability varies by client. The documented GitHub installation route above is for supported desktop clients.

For workspace-wide access, a workspace admin can publish a locally added plugin using Plugins → Personal → plugin menu → Publish, then select eligible workspace roles, subject to workspace policy. Public-directory availability requires a separate OpenAI submission and approval. This repository has not been submitted to that directory.

If your web account exposes Skills → Create → Upload from your computer, you can use the standalone .skill bundle. This installs the skill through that surface, rather than registering the plugin. A Project/custom GPT fallback is documented in the installation guide.

Use and verify

In ChatGPT, type @ and select Mechatronics Engineering. In Codex, use the skill picker; the plugin loader exposes the qualified skill name $mechatronics-engineering:mechatronics-engineering. If you have already installed the standalone skill, disable one copy if duplicate entries appear.

Try these prompts and check the actual output:

• Teach and calculate: “Use Mechatronics Engineering. Can a two-link arm with 0.3 m and 0.2 m links reach (0.6, 0) m?” Expected: unreachable; maximum reach is 0.5 m.
• Build a simulation: “Create a mass–spring–damper browser simulation with adjustable PID gains, live charts, and root-locus, Bode and Nyquist explanations.” Expected: runnable HTML/code with stated equations and model limits.
• Build a diagram: “Open the included block editor. Connect Constant → Gain → Scope and explain how to drag an OUT port onto an IN port.” With constant 3 and gain 2, expect scope output 6.
• Check a boundary: “Use the 3D arm animation to certify motor torque.” Expected: explain that kinematics alone cannot determine or certify torque; request masses, inertia, acceleration and load data.

The included workbench is at skills/mechatronics-engineering/assets/browser-lab/index.html inside the plugin. The block editor is builder.html in that same directory. Open the HTML files in a browser from an extracted package, or ask the assistant to generate downloadable standalone pages. Installing this plugin does not itself embed a persistent simulator panel in ChatGPT.

Calculations need Python; browser numerical tests need Node.js. Native Scilab, USD SDK and ROS 2 execution require their own environments. The plugin bundles instructions and source files; it does not install these runtimes or connect hardware. There are no bundled MCP servers, hooks, credentials, external service connections or telemetry.

Updates, removal and package verification

This guide pins v0.1.1 for repeatable installation. For a newer release, inspect its release notes and register the chosen ref explicitly; upgrading a catalog pinned to an old tag does not select a new version. Reinstall/refresh the plugin and start a new conversation. Preserve project-specific models and dossiers outside the installed plugin cache. Use the Plugins Directory to disable/remove the plugin, or codex plugin remove mechatronics-engineering@personal on supported CLI versions.

Compare downloaded files with the release's SHA256SUMS. In the source repository, run:

python3 scripts/build_packages.py
python3 scripts/validate_package.py

The builder regenerates the plugin's skill directory from canonical skills/mechatronics-engineering/; edit that canonical source. Validation checks source parity, manifest identity, archive contents, safe paths and release checksums. Packaging checks do not establish hosted account installation or hardware validation.

Official documentation

Installation guidance checked on 15 September 2026 against OpenAI plugin packaging and marketplaces, plugin testing, and skill invocation. UI labels and availability can change. Source and licence.