Repolode
Submit
pbakaus

pbakaus/impeccable

Design guidance for AI coding agents, with a design skill, 24 commands, live browser iteration and 61 deterministic detector rules for AI-generated frontend UI.

View on GitHubWebsite
Stars
83,060
Forks
5,022
Open issues
48
Last commit
17 hours ago

What is it?

Impeccable is a design skill and toolkit for AI coding agents such as Claude Code, Cursor, Codex CLI and Gemini CLI. It gives the agent a shared design vocabulary through commands like polish, audit, critique, distill and animate, and an init step that records product context in PRODUCT.md so later commands know the audience and constraints.

It also ships a standalone CLI and browser extension that run 61 deterministic detector rules against HTML, source directories or live URLs, with no LLM or API key required. The rules target common AI-generated frontend patterns such as overused fonts, purple gradients, nested cards and bounce easing, plus general issues like small touch targets and skipped headings.

On supported harnesses, an installable hook runs the detector after direct UI file edits and feeds findings back into the agent's workflow. Live mode lets you iterate on page elements in the browser against a local dev server or static HTML.

Key features

  • 01One /impeccable init step records product context in PRODUCT.md.
  • 0224 commands for critique, audit, polish, layout, typography, motion and more.
  • 0361 deterministic detector rules runnable via CLI, browser extension or agent hook.
  • 04Live browser mode for iterating on elements against a local dev server or static HTML.
  • 05Installer supports many AI coding tools, including Cursor, Claude Code, Codex CLI and Gemini CLI.

Use cases

  • Improving the visual quality of frontend code generated by an AI coding agent.
  • Auditing an existing web page for accessibility, responsive and layout issues before shipping.
  • Running the detector in CI on HTML or JSX to flag common generic AI design patterns.
  • Scanning a production URL for design anti-patterns without touching the source code.

Quick start

  1. 1

    Install the skill into your project from the project root

    npx impeccable install
  2. 2

    Inside your AI coding tool, run the one-time setup command

    /impeccable init

Things to know

  • Requires Node.js for the npx installer, though manual and Git install options work without it.
  • Some harnesses need extra steps such as hook approval or trust settings before the design hooks run.

Topics

#ai-agents#frontend#design#ui-ux#linter#claude-code#cursor#accessibility

FAQ

What is impeccable?
Design guidance for AI coding agents, with a design skill, 24 commands, live browser iteration and 61 deterministic detector rules for AI-generated frontend UI. Impeccable is a design skill and toolkit for AI coding agents such as Claude Code, Cursor, Codex CLI and Gemini CLI. It gives the agent a shared design vocabulary through commands like polish, audit, critique, distill and animate, and an init step that records product context in PRODUCT.md so later commands know the audience and constraints.
Is impeccable free to use? What license does it have?
impeccable is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
Is impeccable still maintained?
Actively maintained. The last commit to impeccable was on Oct 19, 2026.
What can I use impeccable for?
Improving the visual quality of frontend code generated by an AI coding agent. Auditing an existing web page for accessibility, responsive and layout issues before shipping. Running the detector in CI on HTML or JSX to flag common generic AI design patterns. Scanning a production URL for design anti-patterns without touching the source code.
How do I install impeccable?
Install the skill into your project from the project root: npx impeccable install Inside your AI coding tool, run the one-time setup command: /impeccable init
Where can I get impeccable?
The source code and installation instructions are on GitHub: https://github.com/pbakaus/impeccable

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects