Design System From Scratch logo

Design System From Scratch

Community
nnikochann
design-system-from-scratch

Use when Codex needs to create a new design system before a product has stable UI, components, or tokens. Trigger for greenfield products, new brands, app/site concepts, startup MVPs, reference-driven visual direction, moodboards, screenshots, competitor sites, Figma references, brand exploration, or requests to define colors, typography, components, layout, and visual rules from scratch.

Overview

Publishernnikochann
Repositorydesign-system-skills
Skill namedesign-system-from-scratch
Stars
3
Forks
0
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 nnikochann on GitHub. Read the source before you install it.

Installation

Install the Design System From Scratch 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/nnikochann/design-system-skills/tree/main/skills/design-system-from-scratch
  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/nnikochann/design-system-skills.git /tmp/design-system-skills
mkdir -p .claude/skills
cp -r /tmp/design-system-skills/skills/design-system-from-scratch .claude/skills/design-system-from-scratch
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Design System From Scratch 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 System From Scratch 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 System From Scratch 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 System From Scratch

Purpose

Create a design system when there is no reliable existing UI to extract from. The process starts with product intent and references, not implementation. The output is a usable system: tokens, component foundations, layout rules, accessibility rules, visual principles, and a manifest that future UI work can follow.

This pipeline differs from $design-system-research: do not pretend to extract a mature system from an empty repo. First build design direction from brief and references, then author the system deliberately.

Core Rule

Do not generate polished UI immediately. First clarify the product, gather or request references, synthesize visual direction, define tokens/components, and write docs/design-system/manifest.json.

If the repo already has a real product UI, component library, or design tokens that should constrain the work, use $design-system-research first.

Workflow

  1. Intake the product.

    • Capture product name, category, audience, primary workflows, business goal, platform targets, technical stack, brand constraints, accessibility expectations, and required deliverables.
    • Ask for references only when missing information would materially change the design direction. Prefer 3-7 strong references over a large unfocused dump.
    • When the user provides sparse input, proceed with stated assumptions and record open questions.
  2. Build a reference board.

    • Read references/reference-intake.md for reference categories and evaluation prompts.
    • Use provided Figma links, screenshots, public sites, brand books, product docs, competitor URLs, and moodboard images.
    • Separate references by role: aspirational mood, direct competitor, layout, typography, color/material, interaction/motion, imagery, and anti-reference.
    • Save the source list and observations in docs/design-system/reference-board.md.
  3. Synthesize direction.

    • Define 1-3 design territories when the brief is exploratory. Each territory should include mood, typography direction, color/material approach, layout density, motion attitude, and risks.
    • If the user has already chosen a direction, skip territories and document the chosen direction.
    • Avoid generic style words unless they are converted into concrete rules. "Premium" is not a rule; "low-chroma surfaces, restrained accent color, dense editorial type scale, no glass cards" is a rule.
  4. Author tokens.

    • Create semantic colors, typography roles, spacing scale, radius scale, elevation/shadow policy, breakpoints, grid rules, and motion tokens.
    • Define why each major token exists and where it should be used.
    • Keep values implementation-ready for the project's stack when known: CSS variables, Tailwind theme, design tokens JSON, or component props.
  5. Define component foundations.

    • Start with primitives: Button, Link, Input, Select, Checkbox, Toggle, Tabs, Tooltip, Modal/Dialog, Toast, Card/Panel, Table/List, Nav, Badge, Avatar, Empty State, Loading, Error.
    • For each component, define variants, states, anatomy, spacing, typography, accessibility, and usage limits.
    • Design for reuse before adding expressive one-off sections.
  6. Define layout and visual principles.

    • Specify page shells, grid, max widths, responsive behavior, section rhythm, density, hierarchy, content rules, iconography, imagery, data visualization, and motion.
    • Include anti-patterns so future UI work has guardrails.
  7. Write artifacts.

    • Read references/manifest-schema.md before writing manifest.json.
    • Create or update:
      • docs/design-system/manifest.json
      • docs/design-system/design-system.md
      • docs/design-system/reference-board.md
      • docs/design-system/component-map.md
    • Optional when useful:
      • docs/design-system/tokens.css
      • docs/design-system/tailwind.tokens.js
      • docs/design-system/visual-principles.md
  8. Protect the project instructions.

    • After the design-system artifacts are written, search the repository root for an existing agent instruction file in this order: AGENTS.md, agents.md, AGENT.md, agent.md.
    • If one exists, update it. If none exists, create AGENTS.md at the repository root.
    • Add a short "Design System Guard" section explaining that the project's visual layer is protected by the design system: the design system is the source of truth for visual decisions, and UI work must follow it before introducing new styles, tokens, or components.
    • List the design-system files that own the contract. Include required files and any optional files that were actually created, for example:
      • docs/design-system/manifest.json
      • docs/design-system/design-system.md
      • docs/design-system/reference-board.md
      • docs/design-system/component-map.md
      • docs/design-system/tokens.css
      • docs/design-system/tailwind.tokens.js
      • docs/design-system/visual-principles.md
    • Keep this instruction block concise. Do not duplicate the full design-system documentation inside the agent instruction file.
    • Use this shape, adapted to the project's language and actual file list:
      md
      ## Design System Guard
      
      The visual layer of this project is protected by the design system. The design system is the source of truth for UI decisions: colors, typography, spacing, components, layout, motion, and visual QA. Before changing UI, read and follow these files:
      
      - `docs/design-system/manifest.json`
      - `docs/design-system/design-system.md`
      - `docs/design-system/reference-board.md`
      - `docs/design-system/component-map.md`
      
      Do not introduce new visual tokens, component variants, or layout patterns unless they are added to the design-system contract first.
  9. Prototype only after the system exists.

    • If the user requested screens in the same run, implement a small representative slice first: one page shell plus core components and states.
    • Run visual QA across mobile, tablet, and desktop.
    • Summarize design-system deviations and whether they should become new rules or be removed.

Reference Expectations

Good reference input includes:

  • 2-3 products the user wants to feel adjacent to.
  • 1-2 anti-references showing what to avoid.
  • Audience and product category.
  • Desired brand traits with concrete examples.
  • Screens or flows the system must support first.

Do not copy references literally. Extract principles: hierarchy, density, rhythm, material, interaction, contrast, and tone.

Output Standard

The design system is ready when another Codex run can build UI from the artifacts without re-asking basic visual questions. The final response should include created files, the updated or created project agent instruction file, chosen direction, open questions, verification performed, and the recommended next UI slice.

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 System From Scratch AI skill do?

Use when Codex needs to create a new design system before a product has stable UI, components, or tokens. Trigger for greenfield products, new brands, app/site concepts, startup MVPs, reference-driven visual direction, moodboards, screenshots, competitor sites, Figma references, brand exploration, or requests to define colors, typography, components, layout, and visual rules from scratch.

Why use Design System From Scratch on TypingMind?

Because you install it once and use it with any model. Design System From Scratch 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 System From Scratch in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/nnikochann/design-system-skills/tree/main/skills/design-system-from-scratch. 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 System From Scratch?

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 System From Scratch?

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

Is the Design System From Scratch AI skill free?

Yes. It is published on GitHub by nnikochann 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 👇