agentsclimarketplace

Nextjs performance

Skill jgamaraalv/delivery-loop/.claude/skills/nextjs-performance

Continuous fullstack delivery loops — orchestrates frontend, backend, and quality subagents (behaviour drivers, engineers, UI/UX specialist, code/security reviewers, architects) in a test → diagnose → fix → review → secure → re-test cycle until the work is production-ready

Install
npx -y skills add jgamaraalv/delivery-loop --skill nextjs-performance

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Next.js performance optimization for Core Web Vitals, image/font optimization, caching, streaming, bundle size, and Server Components. Use when fixing LCP/INP/CLS, implementing next/image or next/font, configuring unstable_cache, converting Client to Server Components, streaming with Suspense, or reducing bundle size. Covers Next.js 16 + React 19 patterns.

SKILL.md

5.5 KB, as published. Nobody here has run it

Next.js Performance Optimization

Before Starting

  1. Analyze current performance with Lighthouse
  2. Identify bottlenecks - check Core Web Vitals in Chrome DevTools or PageSpeed Insights
  3. Determine optimization priority:
    • LCP issues → Focus on images, fonts
    • INP issues → Reduce JS, use Server Components
    • CLS issues → Add dimensions, use next/font

Core Principles

  1. Prefer Server Components - Only use 'use client' when necessary (browser APIs, interactivity)
  2. Load components as low as possible - Keep Client Components at leaf nodes
  3. Use Suspense boundaries - Enable streaming and progressive loading
  4. Cache appropriately - Use tags for granular revalidation
  5. Measure before/after - Always verify improvements with real metrics

References

Each file is loaded on demand — read one only when the task needs that depth (progressive disclosure).

  • references/core-web-vitals.md — LCP/INP/CLS targets, diagnostics, and per-metric optimizations · read when measuring or fixing Core Web Vitals.
  • references/image-optimization.md — next/image sizing, priority, blur placeholder, remotePatterns config · read when fixing LCP or image-related CLS.
  • references/font-optimization.md — next/font setup, subsets, display swap, CSS variable integration · read when eliminating font-related CLS or FOUT.
  • references/caching-strategies.md — unstable_cache, revalidateTag/Path, ISR, cache tags, TTL guidance · read when configuring data or route caching.
  • references/server-components.md — Client-to-Server conversion patterns, RSC constraints, Suspense boundaries · read when moving components server-side or wiring streaming.
  • references/streaming-suspense.md — Suspense boundaries, progressive loading, fallback skeletons · read when implementing streaming or reducing TTFB.
  • references/bundle-optimization.md — dynamic(), code splitting, tree shaking, modularizeImports, bundle-analyzer · read when reducing initial JS bundle size.
  • references/metadata-seo.md — static/generateMetadata, Open Graph, canonical, sitemap · read when configuring metadata or SEO.
  • references/api-routes.md — Route Handler patterns, Edge runtime, input validation, error handling · read when optimizing API routes.
  • references/nextjs-16-patterns.md — async params Promise, use() hook, useOptimistic, React 19 patterns · read when adopting Next.js 15+/React 19 APIs.

Common Conversions

FromToBenefit
useEffect + fetchDirect async in Server Component-70% JS, faster TTFB
useState for dataServer Component with direct DB accessSimpler code, no hydration
Client-side fetchunstable_cache or ISRFaster repeated loads
img tagnext/imageOptimized formats, lazy loading
CSS font importnext/fontZero CLS, automatic optimization
Static import of heavy componentdynamic()Reduced initial bundle

Constraints and Warnings

Server Components Limitations

  • Cannot use browser APIs (window, localStorage, document)
  • Cannot use React hooks (useState, useEffect, useContext)
  • Cannot use event handlers (onClick, onSubmit)
  • Cannot use dynamic imports with ssr: false

Image Optimization Constraints

  • priority should only be used for above-the-fold images
  • External images require configuration in next.config.js
  • width and height are required unless using fill
  • Animated GIFs are not optimized by default

Caching Considerations

  • Cache tags must be manually invalidated
  • Data cache is per-request in development
  • Edge runtime has different caching behavior
  • Be careful caching user-specific data

Bundle Size Warnings

  • Dynamic imports can impact SEO if critical content
  • Tree shaking requires proper ES module usage
  • Some libraries cannot be tree shaken (avoid barrel exports)
  • Client Components increase bundle size - use sparingly

Performance Checklist

  • All images use next/image with proper dimensions
  • LCP images have priority attribute
  • Fonts use next/font with subsets
  • Server Components used where possible
  • Client Components at leaf level only
  • Suspense boundaries for data fetching
  • Caching configured for expensive operations
  • Bundle analyzed for duplicates
  • Heavy components lazy loaded
  • Lighthouse score verified before/after

Common Mistakes

// ❌ DON'T: Fetch in useEffect
'use client'
useEffect(() => { fetch('/api/data').then(...) }, [])

// ✅ DO: Fetch directly in Server Component
const data = await fetch('/api/data')

// ❌ DON'T: Forget dimensions on images
<Image src="/photo.jpg" />

// ✅ DO: Always provide dimensions
<Image src="/photo.jpg" width={800} height={600} />

// ❌ DON'T: Use priority on all images
<Image src="/photo1.jpg" priority />
<Image src="/photo2.jpg" priority />

// ✅ DO: Priority only for LCP
<Image src="/hero.jpg" priority />
<Image src="/photo.jpg" loading="lazy" />

// ❌ DON'T: Cache everything with same TTL
{ revalidate: 3600 }

// ✅ DO: Match TTL to data change frequency
{ revalidate: 86400 } // Categories rarely change
{ revalidate: 60 }     // Comments change often

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.