Skip to main content
Main content starts here
Shorts0:55

Same AI design? Fixed — free

In this video

  • One system prompt plus 14 procedural skills, free on GitHub
  • WCAG audit, design-token extractor, and an AI-slop check included
  • Community project, not an official Anthropic release

Gemini Summary

Ethan covers a community open-source release: a developer reverse-engineered the system prompt behind Claude Design and published it under MIT license. One system prompt plus 14 procedural skills, including a WCAG accessibility audit, a design-system token extractor, and an AI-slop check that hunts the gradient-and-emoji look.

Notes

Your AI keeps designing the same gradient-and-emoji template? A developer reverse-engineered the system prompt behind Claude Design and open-sourced it under MIT license, free even for commercial use.

It ships as one system prompt plus 14 procedural skills: a WCAG accessibility audit, a design-system token extractor, and an AI-slop check that hunts down the gradient, emoji, Inter-everywhere look. It works in Claude, GPT, Gemini, or a local model. Heads up: this is one developer's community open-source project, not an official Anthropic release.

More videos

Get your US business address today

No setup fees. Cancel anytime. Active within 24 hours.

Get your address