Taste Skill
Frontend framework that stops AI agents from generating generic UIs
Visit tasteskill.dev ↗What is Taste Skill?
Taste Skill is an open-source collection of skill files designed to improve the output quality of AI coding agents like Cursor, Claude Code, and Codex. Instead of generic templates, it guides these tools to generate interfaces with better layout, typography, motion, and spacing. The framework works across multiple AI coding platforms through a single install command.
The project includes both frontend design skills and image-generation skills for reference boards. Users can pair generated designs with AI image tools, then hand the frames to their preferred coding agent for implementation. Taste Skill is free and MIT-licensed, maintained by Leon Lin with support from 21 contributors.
Momentum
Daily activity over the last 30 days: upvotes, unique visits, new GitHub stars and Hacker News points.
How do we know it's vibe coded?
vibetrend discovered Taste Skill automatically. Evidence: GitHub Leonxlnx/taste-skill; repo has .github/copilot-instructions.md. View source.
Similar vibe-coded apps
Open source AI gateway routing across 350+ providers through one endpoint
Open-source AI job search agent that runs locally in your CLI
Context optimization layer that compresses LLM inputs to reduce token usage
CLI proxy that cuts noise from command output before it reaches AI
LLM-driven stock analysis system for A-share, Hong Kong, and US markets
Dependency-aware issue tracker for AI coding agents that survive context loss
Local code graph for AI-assisted reviews with 63x context reduction
Open-source orchestration layer for composing and governing AI agents