Trystan-SA

Trystan-SA / claude-design-system-prompt

Public

Reverse-engineered system prompt and skill library that turns an LLM into an opinionated, accessibility-aware, AI-slop-resistant design collaborator.

12
5
100% credibility
Found Apr 28, 2026 at 12 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
AI Summary

A library of prompts and skills that configures a language model to act as an opinionated design collaborator emphasizing accessibility, quality, and avoidance of clichéd AI outputs.

How It Works

1
🔍 Discover the design helper

You find a free set of instructions on GitHub that turns any AI chat into a smart design partner avoiding boring, generic looks.

2
đź“– Explore the guide

You read the simple readme to see how it teaches AI to focus on real quality, accessibility, and thoughtful visuals.

3
✨ Grab the main instructions

You copy the core prompt and the handy skill guides that the AI can use for different design tasks.

4
đź’¬ Paste into your AI chat

You open your favorite AI conversation tool, like Claude or ChatGPT, and add the instructions at the start.

5
🎨 Start your design project

You tell the AI about your idea, and it asks smart questions, draws wireframes, builds prototypes, and suggests variations.

6
âś… Review and polish

You run quick checks for accessibility, style issues, and interactions to make everything perfect.

🏆 Get pro-level designs

You now have clean, accessible, human-feeling designs ready to use, feeling proud of the thoughtful results.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 12 to 12 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is claude-design-system-prompt?

This repo delivers a reverse-engineered system prompt from Anthropic's Claude Design, plus a library of procedural skills, that transforms any LLM—Claude, ChatGPT, Gemini, or local models—into a strict design collaborator. It enforces content discipline, visual hierarchy, WCAG accessibility, and real CSS techniques while rejecting AI slop like gradients and emoji spam. Drop the prompt into your LLM setup for opinionated outputs in HTML prototypes, audits, or token extraction.

Why is it gaining traction?

Unlike generic design prompts churning out SaaS clones, this one commits to a philosophy: no filler, purposeful palettes, interaction states, and system thinking—echoing reverse-engineered GitHub gems for ChatGPT, Claude code, or even GTA projects. Developers hook into its 14 skills for workflows like wireframing prototypes, accessibility audits, or AI-slop checks, chaining them for polished results. It's MIT-licensed and platform-agnostic, making LLMs viable for real design work without the tropes.

Who should use this?

Frontend devs prototyping UIs solo, needing quick interactive mocks or variations without Figma. Teams reverse-engineering design systems from legacy sites or auditing for WCAG and keyboard nav. LLM tinkerers building custom design agents resistant to low-effort outputs.

Verdict

Worth forking for LLM-driven design experiments—solid docs and principles punch above its 12 stars and 1.0% credibility score—but too early for production pipelines without your own testing. Pair it with a capable model like Claude for best results.

(178 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.