perf-auditor

What This Skill Does

The perf-auditor is a high-performance diagnostic tool designed to bridge the gap between abstract Lighthouse scores and actionable code improvements. Instead of providing generic advice on how to improve your Core Web Vitals, this skill runs a deep-dive audit on your target URL and synthesizes the findings with AI to generate specific, code-centric recommendations. It effectively acts as a senior performance engineer in your terminal, identifying render-blocking resources, image optimization candidates, and JavaScript execution bottlenecks. The tool prioritizes these issues based on their potential impact on your overall score, allowing developers to focus on the high-value 'quick wins' that typically account for the majority of latency issues.

Installation

To integrate this skill into your OpenClaw environment, execute the following command in your terminal:

clawhub install openclaw/skills/skills/lxgicstudios/perf-auditor

Ensure you have Node.js 18 or higher installed on your system. No further global installations or API key configurations are required, as the tool leverages the standard npx environment to run directly.

Use Cases

• SEO Optimization: Use this tool when your site's Core Web Vitals are failing to meet the threshold for search engine ranking requirements.
• Performance Regression: Quickly diagnose why a production site's performance score dropped after a recent deployment.
• Client Reporting: Generate a clear, prioritized list of technical fixes to present to stakeholders when site speed becomes a bottleneck for conversions.
• Pre-Launch Audits: Systematically identify and resolve performance debt before a major product release or marketing campaign.

Example Prompts

1. "Run a performance audit on https://ecommerce-site.com and identify the top 3 bottlenecks affecting mobile load times."
2. "I'm getting a 45 Lighthouse score on my homepage. Use the perf-auditor to analyze it and suggest specific code refactors for the main bundle."
3. "Perform a performance check on my local dev environment at http://localhost:3000 to see if my recent image optimization changes improved the LCP metric."

Tips & Limitations

• Production vs. Dev: Always run audits against a production-like environment. Audit results on local development servers often include debug overhead that doesn't reflect actual user experience.
• Prioritization: Resist the urge to fix every minor suggestion. Focus on the 'top 3' recommendations provided by the AI, which usually address 80% of the underlying performance friction.
• Contextual Awareness: Remember that performance is subjective to the device. Always run both mobile and desktop audits to ensure your fixes aren't introducing regressions for users on lower-powered devices.
• Network Conditions: While the tool provides excellent diagnostics, remember that network latency is variable. Perform multiple runs to get an average baseline.