impeccable  by pbakaus

AI-powered frontend design enhancement for LLMs

Created 3 months ago
326 stars

Top 84.0% on SourcePulse

GitHubView on GitHub
Project Summary

Summary

Impeccable enhances Large Language Models (LLMs) for UX and frontend design tasks by providing specialized vocabulary, commands, and curated anti-patterns. It addresses the common issue of LLMs producing generic or flawed designs due to a lack of specific design expertise. This tool equips engineers and designers with the language and guidance needed to elicit high-quality, expert-level UI designs from AI, moving beyond predictable mistakes like poor typography, color choices, or layout issues.

How It Works

Impeccable expands upon Anthropic's foundational frontend-design skill by incorporating seven domain-specific reference files (covering typography, color, spatial, motion, interaction, responsive, and UX writing). It introduces seventeen steering commands, such as /audit, /polish, /simplify, and /animate, enabling users to instruct LLMs on specific design actions. Crucially, it includes curated anti-patterns, explicitly guiding the AI away from common design pitfalls and biases, thereby fostering more intentional and effective frontend development.

Quick Start & Requirements

Installation can be done via a universal command (npx skills add pbakaus/impeccable), which auto-detects the AI harness, or by downloading a universal ZIP file from impeccable.style and extracting it to the project root. Manual installation involves copying provider-specific directories (e.g., .cursor/, .claude/, .gemini/, .codex/, .agents/) into the project. Supported tools include Cursor, Claude Code, Gemini CLI, Codex CLI, VS Code Copilot, and Antigravity. Specific requirements include using the Cursor Nightly channel and enabling Agent Skills, or installing the Gemini CLI preview version and enabling Skills.

Highlighted Details

  • Features seven comprehensive reference files covering key design domains: typography, color-and-contrast, spatial-design, motion-design, interaction-design, responsive-design, and ux-writing.
  • Provides seventeen distinct commands for detailed design control, including /audit for quality checks, /critique for UX review, /normalize for design system alignment, and /animate for purposeful motion.
  • Includes a curated list of anti-patterns to prevent common design mistakes, such as using overused fonts, inappropriate text colors on backgrounds, or dated easing functions.
  • Offers before-and-after case studies and a command cheatsheet on impeccable.style.

Maintenance & Community

The project was created by Paul Bakaus. Version 1.1.0 (March 4, 2026) introduced a unified skills architecture, added support for VS Code Copilot and Antigravity, and refined the installation process. The DEVELOP.md file is referenced for contributing guidelines. No specific community channels (like Discord or Slack) or roadmap links were detailed in the provided text.

Licensing & Compatibility

The project is licensed under the Apache 2.0 license. It builds upon Anthropic's original frontend-design skill. The Apache 2.0 license is generally permissive for commercial use and integration into closed-source projects.

Limitations & Caveats

Setup requires specific AI coding environments, some of which may necessitate using preview or nightly builds (e.g., Cursor Nightly, Gemini CLI preview), potentially indicating a less stable or experimental integration. The effectiveness of Impeccable is contingent on the capabilities of the underlying LLM and the context provided.

Health Check
Last Commit

1 day ago

Responsiveness

Inactive

Pull Requests (30d)
1
Issues (30d)
5
Star History
250 stars in the last 30 days

Explore Similar Projects

Starred by Patrick von Platen Patrick von Platen(Author of Hugging Face Diffusers; Research Engineer at Mistral), David Cournapeau David Cournapeau(Author of scikit-learn), and
3 more.

codecompanion.nvim by olimorris

0.6%
6k
Neovim plugin for AI-powered coding assistance
Created 2 years ago
Updated 2 days ago
Feedback? Help us improve.