Design Kit logo

Design Kit

Community
daveebbelaar
design-kit

Custom design system for a macOS SwiftUI app with its own parts, not Apple's, covering three token layers, type, an icon set, every part in every state, custom window chrome, a parts gallery, and checks. Use when the user wants the app to have its own look and feel, a brand, or custom buttons, rows, and sidebars, when design-from-app has measured a reference app, or when the user asks for a design system or component library.

Overview

Publisherdaveebbelaar
Repositorybama
Skill namedesign-kit
Stars
952
Forks
136
Bundled files
3
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 3 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by daveebbelaar on GitHub. Read the source before you install it.

Installation

Install the Design Kit AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/daveebbelaar/bama/tree/main/skills/design/design-kit
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/daveebbelaar/bama.git /tmp/bama
mkdir -p .claude/skills
cp -r /tmp/bama/skills/design/design-kit .claude/skills/design-kit
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Design Kit in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Design Kit on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Design Kit is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Design kit

A native look comes from Apple's parts. A custom look comes from your own parts: your sidebar row, your button, your checkbox, your icons. SwiftUI draws all of them. A kit is those parts, the numbers behind them, and one page that shows every part in every state.

Most apps keep Apple's parts and only pick an accent color (design-system skill). Build a kit when the parts themselves must change. To copy the look of an existing app, call the Skill tool with design-from-app first: it measures the app and then comes back here.

Everything lives in Sources/<App>/DesignKit/. Code for every step: parts.md.

1. Write the spec

Write design/KIT.md from kit-spec.md: three words for the feel, the tokens, the type scale, the icon rules, and one section per part with its sizes and states. Agents build from this page and review against it.

Check: every part the app's screens need has a section with its sizes and its states.

2. Tokens in three layers

DesignKit/Tokens.swift:

  1. Palette: raw colors, named by hue and step (gray50, blue600). Only this file reads it.
  2. Meaning: Ink.text, Ink.accent, Ink.selection, Ink.divider, with a light and a dark value each. Parts read only this layer.
  3. Scales: KitType (type roles), Metric (sizes), Motion (named intents like hover and check).

A rebrand changes the palette. A dark mode fix changes the meaning layer. Neither touches a part.

Check: grep -rn 'Palette\.' Sources/ | grep -v Tokens.swift finds nothing.

3. Type

Name styles by role (title, body, label, section, meta), never by size. Use the system font unless the brand needs its own; a bundled font goes in Resources/Fonts/ with ATSApplicationFontsPath in Info.plist.in. Half-point sizes are fine when a reference needs them.

Check: views set fonts only through .kitType(...).

4. Icons

Icons decide the feel more than color does. Views name a KitIcon, never an SF Symbol or a file, so the whole set can change in one place.

  • Start from SF Symbols at a fixed size and weight. Light weights read less like Apple.
  • For a set of your own, draw each icon as a SwiftUI Shape from its SVG path, or make custom symbols in the SF Symbols app.
  • One size, one stroke weight, one corner style for the whole set.
  • A selected row may switch to the filled form.

Check: grep -rn 'systemName:' Sources/ | grep -v DesignKit/ finds nothing.

5. Parts

One file per family in DesignKit/Components/: sidebar, rows, buttons, fields, menus. Each part:

  • Is a real control. Style a Button with a ButtonStyle, a Toggle with a ToggleStyle, a field with a TextFieldStyle. Never a shape with a tap gesture: it loses the keyboard, focus, and VoiceOver.
  • Has every state: rest, hover, pressed, selected, focused, disabled, in light and dark. Hover uses onHover, pressed uses configuration.isPressed, disabled reads isEnabled.
  • Takes a forced state for the gallery, so hover and pressed show up in a snapshot.
  • Reads only meaning tokens, KitType, Metric, Motion, and KitIcon.

Check: the part matches its section in design/KIT.md in every state.

6. Window chrome

For a fully custom look, hide the title bar and draw your own top bar, and build the sidebar from kit rows instead of List. Keep the traffic lights, menus, context menus, sheets, and alerts native: users depend on them. SwiftUI cannot move the traffic lights; a small NSViewRepresentable reaches the NSWindow (parts.md, Window chrome).

A custom sidebar matches the macOS 27 system one where people would notice: the selected row's text turns semibold, the selection fill dims when the window is inactive (@Environment(\.appearsActive)), and row corners stay concentric with the sidebar (mac-design patterns.md, "Custom chrome that follows the window"). Custom glass bounces on click only on controls.

Check: the window opens at its default size, the traffic lights sit where the spec says, the custom sidebar moves with the arrow keys, and its selection dims when another window is in front.

7. Gallery

DesignKit/KitGallery.swift shows every part in every state, in light and dark, the way a storybook does. The app shows it instead of its pages when APP_KIT_GALLERY=1 is set, so the normal snapshot command captures it. Review a new part here before it goes into a screen.

Check: the gallery snapshot shows every part from design/KIT.md, and the user has seen it.

8. Checks

bash
tools/grid-lint.sh
grep -rnE 'Color\(hex|Color\(red|\.font\(\.system|systemName:' Sources/ | grep -v DesignKit/

Both come back clean. Then build, snapshot, and run the layout-grid checks on the screens (build-and-verify skill).

Done when design/KIT.md covers every part, every part is a real control with every state, the gallery shows them all in light and dark, and the checks above come back clean.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Design Kit AI skill do?

Custom design system for a macOS SwiftUI app with its own parts, not Apple's, covering three token layers, type, an icon set, every part in every state, custom window chrome, a parts gallery, and checks. Use when the user wants the app to have its own look and feel, a brand, or custom buttons, rows, and sidebars, when design-from-app has measured a reference app, or when the user asks for a design system or component library.

Why use Design Kit on TypingMind?

Because you install it once and use it with any model. Design Kit is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Design Kit in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/daveebbelaar/bama/tree/main/skills/design/design-kit. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Design Kit?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Design Kit?

As many as you like. As long as a model supports skills, you can use Design Kit with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Design Kit AI skill free?

Yes. It is published on GitHub by daveebbelaar under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇