core-vitals-fixer

What This Skill Does

The core-vitals-fixer is an intelligent diagnostic tool designed to optimize website performance by addressing Core Web Vitals (LCP, CLS, FID) through AI-powered analysis. Instead of just highlighting performance bottlenecks, this skill scans your local codebase, identifies specific architectural or resource-loading flaws, and generates actionable, prioritized code snippets to resolve them. It bridges the gap between raw Lighthouse audit scores and actual code implementation, allowing developers to improve their SEO and user experience by resolving common performance anti-patterns directly within their project structure.

Installation

To integrate this skill into your workflow, you must have the OpenClaw agent platform configured. Install the package directly from the central registry using the following command in your terminal:

clawhub install openclaw/skills/skills/lxgicstudios/core-vitals-fix

Ensure that you have Node.js 18 or higher installed on your system. Because this tool utilizes AI to analyze your source files, you must provide an OPENAI_API_KEY in your environment variables to allow the agent to process the code suggestions effectively.

Use Cases

This skill is ideal for developers and SEO specialists who need to improve their Search Console performance. Use it when:

• Your Lighthouse scores show poor rankings for mobile or desktop load times.
• Users report "janky" behavior or slow interactivity during page load.
• Your site is experiencing layout shifts (CLS) that negatively impact your search ranking.
• You are in the process of refactoring a legacy codebase and want to ensure performance best practices are applied without manual trial-and-error.

Example Prompts

1. "Analyze my ./src/components folder to identify why my Largest Contentful Paint (LCP) is consistently over 4 seconds and suggest refactors."
2. "I am seeing high Cumulative Layout Shift in my app directory. Please run core-vitals-fixer on those files and explain how I can reserve space for images to stop the jumping."
3. "Run a full performance audit on the current project directory and prioritize fixes that will have the biggest impact on my FID scores."

Tips & Limitations

• Fix Priority: Always address LCP first, as it is the most critical metric for perceived page speed.
• Scope: The tool works best when pointed at specific directories rather than the entire root, especially in large monoliths, to keep the context window focused.
• Environment: Since the tool relies on an external API (GPT-4o-mini), ensure your network allows calls to OpenAI endpoints.
• Limitations: While the tool provides excellent code suggestions, always review the generated patches manually before merging them into production. Performance optimization can sometimes involve trade-offs with complex state management libraries.