Guide · Agent skill

How to use the EEZ Studio Command-Line Control agent skill

Updated 2026-10-29 · written with AI from the project's own documentation and checked against our catalogue data

The cli-anything-eez-studio skill teaches an AI agent to work with EEZ Studio, a visual design tool for embedded device screens, through a command-line program instead of the app's interface. It is written for embedded developers, lab engineers and teams who want scripted, repeatable edits to .eez-project files. After this guide you will know what the skill does, how to install it, how to ask for things in plain language, and which safety settings to check first.

EEZ Studio Command-Line ControlHKUDS/CLI-Anything · Apache-2.0 Download the skill

Quick start

1Download and unzip the skill. Download the skill ZIP from this page and unzip it into ~/.claude/skills/ or a project's .claude/skills/ folder, or upload it under Settings → Capabilities → Skills in the Claude apps.
2Install the Python package. Open a terminal in the agent-harness folder and run pip install -e . to install cli-anything-eez-studio.
3Set up the backend if needed. For backend build features only, clone the EEZ Studio source, run npm install and npm run build, and set EEZ_STUDIO_SOURCE to its absolute path.
4Ask the agent for a task. Describe the project file and the change you want in plain language, such as adding a label to a screen.
5Preview risky edits. Ask for --dry-run on one-shot project commands when you want to review changes before they are saved.
6Review the result. Check the project widgets or SCPI metadata the agent reports, and use version control to diff the .eez-project file.

What the skill lets the agent do

With this skill, an agent can create native .eez-project files, add labels and buttons to LVGL screens, list widgets in a project, and define SCPI instrument subsystems, commands and parameters. The skill describes these as a command-line program named cli-anything-eez-studio.

Some build tasks call the real EEZ Studio backend. Those commands need a separate local copy of the EEZ Studio source, built with Node.js.

Create and edit .eez-project files
Add LVGL labels and buttons to screens and list widgets
Add SCPI subsystems, commands and parameters
Check the backend and run LVGL simulator builds (requires the EEZ Studio source)

Install the skill in Claude

Download the skill ZIP from this page and unzip it into one of these locations: ~/.claude/skills/ for your user account, or .claude/skills/ inside a project folder for that project only. Alternatively, in the Claude apps, open Settings, go to Capabilities, and upload the skill under Skills.

The skill's own install step requires Python with pip. From the agent-harness folder, run the install command below.

Option A: unzip into ~/.claude/skills/
Option B: unzip into a project's .claude/skills/ folder
Option C: upload through Settings → Capabilities → Skills in the Claude apps

cd eez-studio/agent-harness
pip install -e .

Set up the EEZ Studio backend (only for build features)

Commands such as lvgl backend-inspect and lvgl simulator-build need the real EEZ Studio source tree. Clone the studio repository, install dependencies, build it, and point the EEZ_STUDIO_SOURCE environment variable at the folder. Basic project and widget editing does not need this step.

If the backend is missing, these commands are designed to fail loudly rather than guess.

Clone github.com/eez-open/studio
Run npm install and npm run build inside it
Set EEZ_STUDIO_SOURCE to the absolute path of that folder

git clone https://github.com/eez-open/studio.git
cd studio
npm install
npm run build
export EEZ_STUDIO_SOURCE=/absolute/path/to/studio

Ask for work in natural language

You do not need to type commands yourself. Describe the outcome you want, name the project file, and say whether the agent should preview changes. The agent will translate the request into cli-anything-eez-studio calls. The skill tells the agent to always use --json so output is machine-readable.

Good requests name the file, the screen or widget, and the text or instrument detail involved.

"Create a new EEZ project called Panel in panel.eez-project."
"Add a Ready label and a Run button to the panel project, then list the widgets."
"Add a SCPI subsystem named SOURCE with a :VOLTage? command and an optional channel parameter of type nr1."
"Check whether the EEZ Studio backend is set up, then build the simulator into build/sim."

Run the commands the skill uses

The skill uses these patterns. Each command runs against a project file with --project, and --json makes output readable by the agent. Use these as reference when reviewing what the agent did.

Create a project: project new with -o and --name
Add LVGL items: lvgl add-label and lvgl add-button with --text
List widgets: project widgets
Define SCPI metadata: scpi subsystem-add, scpi command-add, scpi parameter-add

cli-anything-eez-studio --json project new -o panel.eez-project --name Panel
cli-anything-eez-studio --json --project panel.eez-project lvgl add-label --text "Ready"
cli-anything-eez-studio --json --project panel.eez-project scpi parameter-add SOURCE :VOLTage? channel --type nr1 --optional

Safety, undo and review

Runs scripts: yes. Uses network: yes. The network use is for cloning and building the EEZ Studio source during setup; review the skill files before installing if that matters to you.

One-shot commands that use --project auto-save changes unless you add --dry-run. Preview with --dry-run first when you want to review edits before they are saved. During interactive REPL or scripted sessions, session undo, session redo and session status let you step back through changes. Keep a copy of important project files under version control.

Use --dry-run to preview project edits before saving
Use session undo and session redo in REPL or scripted sessions
Keep project files under version control so you can diff agent changes

cli-anything-eez-studio --json --project panel.eez-project lvgl add-button --text "Run" --dry-run

What people use it for

Scaffolding a new LVGL panel with a title label and action buttons for an embedded device
Listing widgets in an existing .eez-project file to document a screen
Defining SCPI subsystems, commands and parameters for lab instruments in bulk
Running a repeatable LVGL simulator build from a script after the backend is set up

Frequently asked questions

Is the EEZ Studio command-line skill free to use?

The source describes the skill as part of the HKUDS/CLI-Anything repository under the Apache-2.0 licence. Check the repository's licence file for the terms that apply to your use. EEZ Studio itself is a separate project at github.com/eez-open/studio.

What do I need to install before using this skill?

You need Python with pip to install the package from the agent-harness folder. For backend features you also need a local clone of the EEZ Studio source, built with npm install and npm run build, plus the EEZ_STUDIO_SOURCE environment variable set.

Do I need EEZ Studio installed to use the skill?

Basic project and widget editing works through the command-line tool. Backend commands such as lvgl backend-inspect and lvgl simulator-build require the built EEZ Studio source tree and will fail if it is missing.

Will the agent overwrite my project file?

One-shot commands with --project auto-save changes by default. Add --dry-run to preview edits without saving, and keep your .eez-project file under version control.

Where do I put the skill so Claude can find it?

Unzip it into ~/.claude/skills/ to make it available across your projects, or into a project's .claude/skills/ folder to limit it to that project. In the Claude apps you can instead upload it under Settings → Capabilities → Skills.

Why does the backend command fail with a missing backend error?

The backend commands call EEZ Studio's built Node modules and are designed to fail loudly when they are missing. Confirm that you cloned the studio repository, ran npm install and npm run build, and exported EEZ_STUDIO_SOURCE with an absolute path.

New agent skills, repos and AI tools — every week

The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.

Our verdict

Use this skill if you are an embedded developer or engineer who is comfortable with Python and a command line and wants an agent to scaffold LVGL screens, edit .eez-project files or define SCPI metadata in repeatable steps. Pick something else if you only want a visual design tool or cannot set up a Python package and build a large Node.js project, since the backend features depend on that setup.

Download the skill