Performance Audit

What it does

The Performance Audit skill guides AI agents through comprehensive performance audits of web applications. It helps identify bottlenecks, optimization opportunities, and potential performance regressions during development or after deployments. The audit process involves reconnaissance to understand application architecture and bundle analysis to pinpoint large dependencies and code inefficiencies. This is particularly useful when preparing for production releases or investigating user experience issues.

When to use it

• Conducting pre-deployment performance reviews
• Investigating slow page loads or poor user experience
• Performing periodic performance audits on existing applications
• Validating performance after major feature additions
• Preparing for traffic scaling or load testing
• Optimizing Core Web Vitals for SEO and user experience

Key capabilities

• Reconnaissance: Identifies performance-critical code using discover queries with patterns for bundle imports, database queries, component files, and API routes.
• Bundle Analysis: Focuses on identifying large dependencies, code splitting opportunities, and dead code within application bundles.
• Component Identification: Helps identify page entry points, heavy npm dependencies, database access patterns, image assets, API endpoints, and client-side rendering techniques.

Example prompts

• "Perform a performance audit of my web application."
• "Run the reconnaissance phase to map performance-critical code in my project."
• "Analyze the bundles for large dependencies and potential code splitting opportunities."

Tips & gotchas

• The skill relies on discover queries, so ensure your agent has access to the codebase.
• The verbosity: files_only setting limits output to file lists during reconnaissance; adjust as needed.
• This skill is designed for web application performance auditing and may not be applicable to other types of projects.