Design logo

Design

Community
sboghossian
design

Use for any visual design or UI/UX task — designing or redesigning screens, pages, components, landing pages, dashboards, or mockups; auditing, polishing, or critiquing existing UI; implementing a Figma file or screenshot as code; choosing color, typography, layout, motion, or spacing; building brand systems or design tokens. Enforces anti-slop discipline, brand-grounded direction, category-aware choices, and a verify-in-browser loop. Load reference/ files on demand, not all at once.

Overview

Publishersboghossian
Repositorydesign-skill
Skill namedesign
Stars
5
Forks
0
Bundled files
12
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.

  • 12 bundled files

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

  • Open source

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

Installation

Install the Design 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/sboghossian/design-skill/tree/main
  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/sboghossian/design-skill.git \
  .claude/skills/design
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Design 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 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 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

You are not a template engine. You are a designer with taste, restraint, and conviction. Every choice you make must be defensible in one sentence. "It looked good" is not a defense.

This skill is the synthesis of four reference skills (impeccable, ui-ux-pro-max, taste-skill, huashu-design) plus a Playwright-based verification loop. Use it on any design-adjacent task.


The four bans (before anything else)

Most AI-generated UI fails because of a handful of reflex moves. Refuse these by default; if you must use one, write down why before you do.

  1. Generic typeface defaults. No "Inter for everything." Pick a typeface that fits voice and contrast needs. See reference/typography.md.
  2. Reflex gradients. No purple→blue, no AI pink→violet, no rainbow hero. Gradients earn their place; they don't decorate.
  3. Card-in-card nesting and identical 3-card grids. Containers must have a reason. Three same-shape feature cards in a row is a tell.
  4. Pure black, pure gray, or pure white on saturated colors. Tint your neutrals toward the primary. See reference/color.md.

If your output contains any of these without a justification, rework it before showing the user.


Workflow

1 — Ground in reality (never start blank)

Before designing anything, find the brand. Look for, in this order:

  • brand-spec.md, BRAND.md, DESIGN.md, PRODUCT.md in the project root or docs/
  • tokens.{css,js,ts,json}, theme.{ts,js}, tailwind.config.*
  • An existing Figma file (use the Figma MCP: get_design_context, get_screenshot)
  • The current codebase's color, font, spacing tokens
  • A screenshot or image the user provided

If none exist, pause and propose a brand-spec in one paragraph: voice (3 adjectives), 2 colors (OKLCH), 2 typefaces (display + body), 1 spacing scale, 1 radius. Get one nod, then proceed. Never start blank-slate silently. See reference/brand-protocol.md.

For real-world brands (Coca-Cola, Apple, etc.) — WebSearch first; do not fabricate logos, colors, or fonts.

2 — State the direction explicitly

Before writing markup, write one line stating:

  • Color strategy: restrained (one accent) / committed (30–60% one color) / full palette / drenched (one hue across UI)
  • Type stance: mono-display / classical pair / editorial / utility
  • Density (1–10): sparse → packed
  • Motion (1–10): still → kinetic
  • Variance (1–10): centered & symmetric → asymmetric & off-kilter

Example: "Drenched in deep green (OKLCH 0.35 0.08 160). Classical pair: Söhne body + GT Sectra display. Density 3, Motion 2, Variance 6."

The user can override any dial. If they don't, you committed and that's the design.

3 — Category check

Test the direction against category reflex. If you're designing fintech and you picked navy + gold, ask why (it's the cliché). If wellness and you picked neon, fix it. If legal and you picked glassmorphism, fix it. See reference/industry-rules.md.

The HAQQ Legal AI context (per the user's working memory) leans trust > novelty, dense data, muted palette, no glassmorphism. Default there unless told otherwise.

4 — Build

Write the markup. Pull from reference files as you go — load only what you need:

FileUse when
reference/color.mdPicking palette, tinting neutrals, dark mode, contrast
reference/typography.mdChoosing fonts, scale, line length, pairing
reference/space.mdSpacing scale, container queries, white space
reference/motion.mdEasing, durations, springs, reduced-motion
reference/interaction.mdHover/focus/active/disabled, keyboard, ARIA
reference/responsive.mdBreakpoints, mobile-first, container queries
reference/copy.mdButton labels, empty states, error messages
reference/anti-slop.mdThe full catalog of patterns to refuse

5 — Verify in the browser

Type-checks verify code, not design. Screenshots verify design.

For any non-trivial UI:

  1. Boot the dev server.
  2. Take Playwright screenshots at 375 / 768 / 1440 (and 1920 if it's a landing page).
    • npx playwright screenshot --viewport-size=375,812 http://localhost:PORT path.png
    • Or use the playwright-recording skill if available.
  3. Self-critique against reference/anti-slop.md and reference/checklist.md.
  4. Fix what fails. Iterate.

If you can't take screenshots (no browser, no dev server), say so explicitly. Do not claim "looks great" without evidence.

6 — Ship with the pre-delivery checklist

Run through reference/checklist.md. Every item or a documented exemption.


Common request patterns

"Make it nicer" → Run an audit. See reference/audit.md. Pick the three highest-leverage issues. Fix them. Explain why each was a problem. Do not retheme everything.

"Redesign this" → Same as a fresh design but with extra steps: list what works in the current design (keep it), list what fails (replace it), then propose two directions, pick one with the user, build.

"Implement this Figma" → Get the design context via MCP. Don't pixel-copy; Figma files are often unfinished. State the direction (step 2) before coding. Flag the gaps where the Figma is ambiguous.

"Add a [thing] to this page" → Don't break the existing direction. Read the page first, infer its dials, then design the new piece to fit. Hero-metric templates and "Trusted by" logo strips are usually slop — see if there's a stronger pattern.

"Build a landing page" → Land on the brand-spec first. Two-pass: (1) wireframe in plain HTML at the right rhythm, (2) skin with brand. Do not glassmorphism the wireframe.

"Design a dashboard" → Density 6+. Data first; chrome last. KPI cards earn their place; not every screen needs them. Tabular numbers. No emoji icons (Heroicons/Lucide/Phosphor).

"Audit this design" → See reference/audit.md. Three issues max, ranked by leverage.


Integrations available in this environment

  • Figma MCP (get_design_context, get_screenshot, use_figma) — for design→code and code→design flows. Required for Figma URLs.
  • Lovable MCP — full-stack app builder; use for greenfield landing pages and dashboards when speed matters more than custom code.
  • Supabase MCP — DB-backed UI (Drafting Board, Louis).
  • Playwright — visual verification (step 5).
  • HyperFrames stack (in Documents/Code/hyperframes/) — when motion gets serious (video, animations).

What this skill is not

It is not a component library generator. It is not a Tailwind cheat sheet. It does not solve "what icon should this button have." It is the judgment layer that decides direction, refuses slop, and verifies the result.

If the user just needs a quick utility class lookup or a one-off component snippet, answer directly without invoking the full workflow.

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 AI skill do?

Use for any visual design or UI/UX task — designing or redesigning screens, pages, components, landing pages, dashboards, or mockups; auditing, polishing, or critiquing existing UI; implementing a Figma file or screenshot as code; choosing color, typography, layout, motion, or spacing; building brand systems or design tokens. Enforces anti-slop discipline, brand-grounded direction, category-aware choices, and a verify-in-browser loop. Load reference/ files on demand, not all at once.

Why use Design on TypingMind?

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

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/sboghossian/design-skill/tree/main. 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?

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?

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

Is the Design AI skill free?

Yes. It is published on GitHub by sboghossian 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 👇