AI Coding Assistants

Use the Slint skill to give your AI coding agent project-specific guidance for working with Slint. The skill covers the .slint language, layouts, common gotchas, host-language interop, and the embedded MCP server.

The plugin and skills are available from the slint-ui/ai-plugins ↗ repository.

Claude Code

Add the Slint marketplace from GitHub and install the plugin:

/plugin marketplace add slint-ui/ai-plugins
/plugin install slint@slint

bash

Claude Code can then use the bundled Slint skill when working with Slint projects.

Codex

Enter the following in the Codex prompt to install the skill from GitHub:

$skill-installer install https://github.com/slint-ui/ai-plugins/tree/master/skills/slint

bash

Restart Codex after the install so it picks up the new skill.

GitHub CLI

If you have the GitHub CLI ↗ installed, gh skill install works for any supported agent — pick the one matching your tool:

# Codex
gh skill install slint-ui/ai-plugins slint --agent codex

# Claude Code
gh skill install slint-ui/ai-plugins slint --agent claude-code

bash

See gh skill install ↗ for the full list of supported agents (Cursor, Gemini CLI, Windsurf, and others). Restart your agent afterward so it picks up the new skill.

npx skills

Install the Slint skill into the current project with the skills ↗ CLI:

npx skills add slint-ui/ai-plugins --skill slint

bash

This works for any agent supported by the skills CLI (Claude Code, Cursor, and others) and installs into the project’s agent skills directory by default. Pass -g to install globally for all projects.

What Your Assistant Can Do

With the skill installed, the assistant follows Slint’s conventions when it writes, debugs, and reviews .slint code, and wires the UI to Rust, C++, JavaScript, or Python. Assistants also learn to verify rendering by taking screenshots of the scene before calling the work done.

The skill also connects the assistant to the Model Context Protocol (MCP) server embedded in your running Slint application. Once connected, the assistant can inspect the UI through its accessibility tree, drive it with mouse, touch, and keyboard input, and capture screenshots to check the result — built on Slint’s existing accessibility and testing APIs, with no debugging hooks in your application.