addyosmani/web-quality-skills

Agent Skills for optimizing web quality based on Lighthouse and Core Web Vitals.

What it solves

It provides a set of specialized "skills" for AI coding agents to help developers optimize web projects for performance, accessibility, SEO, and general quality. Instead of generic advice, it enables agents to use measurement-first evidence—such as Lighthouse audits and Core Web Vitals—to diagnose and fix real issues.

How it works

These skills are instruction sets that can be installed into AI agents (like Claude Code, Codex, or Gemini CLI). They guide the agent to use a specific measurement model involving four types of evidence: CrUX field data (real user experience), first-party RUM, DevTools traces/Lighthouse lab runs, and static source inspection. If the agent has access to the Chrome DevTools MCP, it can perform live audits; otherwise, it falls back to CLI tools and static analysis.

Who it’s for

Web developers using AI coding agents who want to ensure their sites meet professional standards for speed, accessibility (WCAG 2.2), and search engine visibility regardless of the framework they use (e.g., React, Vue, Svelte, Astro).

Highlights

  • Measurement-First Approach: Prioritizes real-user data (CrUX) and lab traces over hypotheses.
  • Comprehensive Quality Suite: Includes dedicated skills for Core Web Vitals (LCP, INP, CLS), WCAG 2.2 accessibility, and technical SEO.
  • Framework Agnostic: Works across all major web frameworks and plain HTML.
  • Agent Integration: Compatible with multiple agent platforms including Claude Code, Codex, and Gemini CLI.

Related

  • Project
    zinja-coder/jadx-ai-mcp

    A plugin and MCP server that integrates the JADX decompiler with LLMs to provide real-time, context-aware reverse engineering and vulnerability analysis for Android APKs.

  • Project
    rusq/slackdump

    A tool to archive private and public Slack messages, channels, and files without admin privileges, featuring an MCP server for AI-powered data analysis.

  • Project
    bergside/awesome-design-skills

    A curated registry of design system skill files that provide AI coding agents with specific UI rules, tokens, and accessibility constraints to ensure consistent, high-quality design implementation.

  • Project
    teaql/teaql-agent-kit

    A model-mediated harness for coding agents that ensures reliable software development by placing a validated domain model and typed contracts between business requirements and final code.