TrendWhat rose on GitHub, every day at 07:30 KST · 한국어

GitHub · as of November 8, 2026 · View on GitHub →

Leonxlnx/taste-skill

AI coding agent design skill

+4,359 stars in 7 days · trending 8 of the last 30 days

Give coding agents the design sense to build professional websites instead of generic AI pages

When you ask an agent to build a corporate website with a simple prompt, the result is usually a mix of similar layouts and careless text. This tool forces the agent to check the page in a real browser, enforce design rules, and verify quality. The output is a refined webpage with a palette and typography suited to a specific industry, along with actual photos and icons.

New verification system in version 2.5

The recently updated version 2.5 focuses on controlling quality in a real browser environment rather than just generating code. After writing the page, the agent runs rendering at multiple resolutions via headless Chrome. It automatically detects empty visual elements, broken shaders, insufficient contrast, alignment errors, and image loading issues on slow networks. A 'finish gate' applies up to three iterations of inspection until all fixable errors are resolved, preventing unfinished pages from being delivered.

Industry-specific design tokens and layouts

The tool classifies client industries into six categories: software, gaming, tech, product, service, and location. Depending on the industry, the palette, Google Fonts combination, and hero section layout are determined automatically. For example, shader effects and 3D objects are allowed for software or tech pages, while service pages like cafes or hotels are restricted to quiet designs centered on photos and typography. These rules control the agent from overusing excessive effects, maintaining an appropriate visual intensity for each page.

Real assets and offline support

Photos and icons are not left blank even if the client does not provide them. It uses a curated photo library from Wikimedia Commons and Unsplash, citing sources in the footer. Software pages use actual brand icons and interface SVGs resolved locally. Bundled photos work offline, and network access is only required when downloading additional photos or icons. It can run with Node 22 or higher and a Chrome-based browser, without API keys or additional installation steps.

Environment and limitations to check before applying

This tool is designed to work in agent-based coding environments such as Claude Code, Codex, and Cursor. It requires Node 22 or higher and one of Chrome, Edge, or Chromium installed, with ffmpeg optionally needed for WebP output or hero photo cropping. Existing React, Vue, and Svelte projects are built within their respective stacks, while Next.js is handled via static export. The license is MIT, and it is unrelated to official token or cryptocurrency projects. Since the agent's design decision process is automated, it is recommended that the final aesthetic judgment of the output still undergoes developer review.

Written by AI from this repository's README on November 8, 2026. For the facts, the original on GitHub is the reference.

By the numbers

GitHub trending history
26 days between Jun 19, when Trend began recording, and Nov 8, 2026 · 8 of the last 30 days · best daily rank #4 (Jul 5)

Read from GitHub, as of November 8, 2026.

Language
JavaScript
Topics
agent-skills · ai-agents · anti-slop · claude-code · codex · cursor · design-system · frontend
Stars
111,953
Stars gained in 7 days
+4,359
Contributors
27
Forks
7,529
Latest release
v2.5.1 · November 7, 2026
Commits
321
Last commit
November 7, 2026
Open issues
1
Open pull requests
1

Related repositories

View on GitHub → · Homepage

README badge

trend.ondo.sh · GitHub daily trending #4 · July 5, 2026

The badge above shows the highest this repository has ranked on GitHub's daily trending since June 19, 2026, and the day it did. Paste the Markdown below into your README. The badge opens this page, and it updates by itself when the repository climbs higher.

[![trend.ondo.sh](https://trend.ondo.sh/github/r/Leonxlnx--taste-skill/badge.en.svg)](https://trend.ondo.sh/github/r/Leonxlnx--taste-skill/en.html)