N
nuuxcode-arkx/ai-frontend-design-kit
Skills, repos and rules that make AI coding agents (Claude Code, Codex, Cursor) produce designed frontend work instead of AI slop.
updated 2026-11-19
View on GitHub ↗Direct link · no redirect page
- Star count
- 0 ★ · nuuxcode-arkx/ai-frontend-design-kit · read from the GitHub search API · (checked 2026-11-19 20:00:48)
- Source
- read directly from its GitHub page · type: open-source project
- Harvest history
- first seen 2026-11-18 · last seen 2026-11-20 · seen 3× · listed by 1 sources
- Score
- score 103 · 1 sources · repo
FAQ (answers come from our records)
Who is it for?
We filed it under these tags: Agent skills. The tags come from the entry itself, not a taxonomy we imposed.
Who maintains it?
repository
nuuxcode-arkx/ai-frontend-design-kit;first catalogued 2026-11-18;last seen 2026-11-20。There is no official certification or quality rating here — we record sources and dates only.README excerpt (English, untranslated) · Read the full README on GitHub →
# AI FRONTEND DESIGN KIT
Excerpted from the repository README (abridged)
Install once. After that, your AI coding agent is instructed to follow a design workflow each time it builds or changes a screen, without you asking. It is an instruction plus a reminder, not a guarantee: in our tests on 17 Nov 2026 fresh agents followed most of it on plain requests such as "make me a landing page" (details under LIMITS), and each task ends with a receipt that says what was checked and what was not. Made for Claude Code and Codex. Other agents (Cursor, Gemini CLI) can use the same files, with less automation.
## INSTALL (FOR PEOPLE)
Open your agent and send it this one message:
text
Read https://github.com/nuuxcode-arkx/ai-frontend-design-kit and follow its INSTALL.md.
That is all. The agent installs the workflow, the design skills and a browser it can look through, for all your projects, then tells you what worked. You do not need to understand any of the tools. You need Node.js 22.20 or newer on the computer; if it is missing, the agent tells you.
## EVERY JOB GETS ITS BEST TOOL
The kit knows about 350 tools. Installing all of them would flood the agent, and installing six would waste the rest. So the agent does a tool sweep on every task: it walks a list of 43 jobs (direction, references, colors, fonts, components, icons, images, charts, motion, 3D, scroll, text effects, copy, checks and more), switches each job on or off from your words and your project, and fetches the right tool at the moment a job needs it: a skill is read into the project for that task, never installed for all projects. In three real tests the agent answered all 43 jobs. The number of tools depends on the jobs the page really needs. One tool per job, never two fighting over the same one. The receipt shows which tools ran, where the browser or the project's own code was enough, and which checks could not run.
## WHAT CHANGES AFTER INSTALL
You ask for anything with a screen: a website, an app page, a dashboard, a quick HTML demo. You do not mention design. The agent now does this on its own:
| Step | What the agent does | Why it matters |
| 1. Brief and content | Writes down what the screen is for and the real words of each section | Design follows the message, not the other way round |
| 2. Identity | Reads or writes a DESIGN.md file: colors, fonts, spacing, tone | Every screen looks like the same product |
| 3. Direction | Uses one design skill and real reference sites to choose a style | No generic template look |
| 4. Build | Uses real components, real icons, chosen fonts, current web practice, and builds every state (loading, empty, error) | Solid, accessible parts |
| 5. Motion and effects | Adds smooth interaction motion everywhere. On pages that deserve it (landing pages, launches, "futuristic" briefs) it also adds signature effects by itself: 3D scenes (three.js), scroll storytelling, parallax, diagrams with light flowing along the lines, text reveals, animated backgrounds | Impressive where it helps, calm where people work or read, and still fast on a phone |
| 6. Copy | Cleans the text on screen | Does not sound like a chatbot |
| 7. See | Opens the page in a browser at phone, tablet and desktop size, looks at it, uses it like a visitor, fixes what is wrong, repeats | The agent sees its own mistakes before you do |
## INSTALL (FOR PEOPLE)
Open your agent and send it this one message:
text
Read https://github.com/nuuxcode-arkx/ai-frontend-design-kit and follow its INSTALL.md.
That is all. The agent installs the workflow, the design skills and a browser it can look through, for all your projects, then tells you what worked. You do not need to understand any of the tools. You need Node.js 22.20 or newer on the computer; if it is missing, the agent tells you.
## EVERY JOB GETS ITS BEST TOOL
The kit knows about 350 tools. Installing all of them would flood the agent, and installing six would waste the rest. So the agent does a tool sweep on every task: it walks a list of 43 jobs (direction, references, colors, fonts, components, icons, images, charts, motion, 3D, scroll, text effects, copy, checks and more), switches each job on or off from your words and your project, and fetches the right tool at the moment a job needs it: a skill is read into the project for that task, never installed for all projects. In three real tests the agent answered all 43 jobs. The number of tools depends on the jobs the page really needs. One tool per job, never two fighting over the same one. The receipt shows which tools ran, where the browser or the project's own code was enough, and which checks could not run.
## WHAT CHANGES AFTER INSTALL
You ask for anything with a screen: a website, an app page, a dashboard, a quick HTML demo. You do not mention design. The agent now does this on its own:
| Step | What the agent does | Why it matters |
| 1. Brief and content | Writes down what the screen is for and the real words of each section | Design follows the message, not the other way round |
| 2. Identity | Reads or writes a DESIGN.md file: colors, fonts, spacing, tone | Every screen looks like the same product |
| 3. Direction | Uses one design skill and real reference sites to choose a style | No generic template look |
| 4. Build | Uses real components, real icons, chosen fonts, current web practice, and builds every state (loading, empty, error) | Solid, accessible parts |
| 5. Motion and effects | Adds smooth interaction motion everywhere. On pages that deserve it (landing pages, launches, "futuristic" briefs) it also adds signature effects by itself: 3D scenes (three.js), scroll storytelling, parallax, diagrams with light flowing along the lines, text reveals, animated backgrounds | Impressive where it helps, calm where people work or read, and still fast on a phone |
| 6. Copy | Cleans the text on screen | Does not sound like a chatbot |
| 7. See | Opens the page in a browser at phone, tablet and desktop size, looks at it, uses it like a visitor, fixes what is wrong, repeats | The agent sees its own mistakes before you do |
Listed badge (optional)
Copy this line into the README if you want the listing mark. Optional, and it does not affect being listed.
[](https://anclum.com/en/s/nuuxcode-arkx-ai-frontend-design-kit/)
- Source
- github · https://github.com/nuuxcode-arkx/ai-frontend-design-kit
- Browse by field
- Agent skills
- added
- today
More entries
Filed under
Every tag opens the full list for that filter. Counts are the live catalogue figures.