Corrige el conocimiento desactualizado del LLM sobre la plataforma Vercel e introduce sus productos nuevos. Se inyecta al inicio de la sesión.
- Costo de contexto al activarse
- 1.8k tok
- Tamaño del paquete
- 1 archivo
- Última actualización
- el mes pasado
Revisor de buenas prácticas React para archivos TSX. Se activa tras editar varios componentes para aplicar una checklist condensada de calidad: estructura, hooks, accesibilidad, rendimiento y TypeScript.
en todo el repo
0–100, la ruta de este skill
último commit aquí
últimos 90 días
56 tok en reposo
199 KB
Funciona con cualquier agente que lea SKILL.md
npx -y skills add vercel/vercel-plugin --skill react-best-practices --agent claude-codeSe instala solo en este repositorio.
Di cualquiera de estas frases y el agente debería cargar este skill.
Comprehensive performance optimization guide for React and Next.js applications, maintained by Vercel. Contains 64 rules across 8 categories, prioritized by impact to guide automated refactoring and code generation.
Reference these guidelines when:
| Priority | Category | Impact | Prefix |
|---|---|---|---|
| 1 | Eliminating Waterfalls | CRITICAL | async- |
| 2 | Bundle Size Optimization | CRITICAL | bundle- |
| 3 | Server-Side Performance | HIGH | server- |
| 4 | Client-Side Data Fetching | MEDIUM-HIGH | client- |
| 5 | Re-render Optimization | MEDIUM | rerender- |
| 6 | Rendering Performance | MEDIUM | rendering- |
| 7 | JavaScript Performance | LOW-MEDIUM | js- |
| 8 | Advanced Patterns | LOW | advanced- |
async-defer-await - Move await into branches where actually usedasync-parallel - Use Promise.all() for independent operationsasync-dependencies - Use better-all for partial dependenciesasync-api-routes - Start promises early, await late in API routesasync-suspense-boundaries - Use Suspense to stream contentbundle-barrel-imports - Import directly, avoid barrel filesbundle-dynamic-imports - Use next/dynamic for heavy componentsbundle-defer-third-party - Load analytics/logging after hydrationbundle-conditional - Load modules only when feature is activatedbundle-preload - Preload on hover/focus for perceived speedserver-auth-actions - Authenticate server actions like API routesserver-cache-react - Use React.cache() for per-request deduplicationserver-cache-lru - Use LRU cache for cross-request cachingserver-dedup-props - Avoid duplicate serialization in RSC propsserver-hoist-static-io - Hoist static I/O (fonts, logos) to module levelserver-serialization - Minimize data passed to client componentsserver-parallel-fetching - Restructure components to parallelize fetchesserver-after-nonblocking - Use after() for non-blocking operationsclient-swr-dedup - Use SWR for automatic request deduplicationclient-event-listeners - Deduplicate global event listenersclient-passive-event-listeners - Use passive listeners for scrollclient-localstorage-schema - Version and minimize localStorage datarerender-defer-reads - Don't subscribe to state only used in callbacksrerender-memo - Extract expensive work into memoized componentsrerender-memo-with-default-value - Hoist default non-primitive propsrerender-dependencies - Use primitive dependencies in effectsrerender-derived-state - Subscribe to derived booleans, not raw valuesrerender-derived-state-no-effect - Derive state during render, not effectsrerender-functional-setstate - Use functional setState for stable callbacksrerender-lazy-state-init - Pass function to useState for expensive valuesrerender-simple-expression-in-memo - Avoid memo for simple primitivesrerender-split-combined-hooks - Split hooks with independent dependenciesrerender-move-effect-to-event - Put interaction logic in event handlersrerender-transitions - Use startTransition for non-urgent updatesrerender-use-deferred-value - Defer expensive renders to keep input responsivererender-use-ref-transient-values - Use refs for transient frequent valuesrerender-no-inline-components - Don't define components inside componentsrendering-animate-svg-wrapper - Animate div wrapper, not SVG elementrendering-content-visibility - Use content-visibility for long listsrendering-hoist-jsx - Extract static JSX outside componentsrendering-svg-precision - Reduce SVG coordinate precisionrendering-hydration-no-flicker - Use inline script for client-only datarendering-hydration-suppress-warning - Suppress expected mismatchesrendering-activity - Use Activity component for show/hiderendering-conditional-render - Use ternary, not && for conditionalsrendering-usetransition-loading - Prefer useTransition for loading staterendering-resource-hints - Use React DOM resource hints for preloadingrendering-script-defer-async - Use defer or async on script tagsjs-batch-dom-css - Group CSS changes via classes or cssTextjs-index-maps - Build Map for repeated lookupsjs-cache-property-access - Cache object properties in loopsjs-cache-function-results - Cache function results in module-level Mapjs-cache-storage - Cache localStorage/sessionStorage readsjs-combine-iterations - Combine multiple filter/map into one loopjs-length-check-first - Check array length before expensive comparisonjs-early-exit - Return early from functionsjs-hoist-regexp - Hoist RegExp creation outside loopsjs-min-max-loop - Use loop for min/max instead of sortjs-set-map-lookups - Use Set/Map for O(1) lookupsjs-tosorted-immutable - Use toSorted() for immutabilityjs-flatmap-filter - Use flatMap to map and filter in one passadvanced-event-handler-refs - Store event handlers in refsadvanced-init-once - Initialize app once per app loadadvanced-use-latest - useLatest for stable callback refsRead individual rule files for detailed explanations and code examples:
rules/async-parallel.md
rules/bundle-barrel-imports.md
Each rule file contains:
For the complete guide with all rules expanded: AGENTS.md
Reproducido de vercel/vercel-plugin bajo licencia NOASSERTION. Leer esta página en markdown.
69 archivos en el paquete. Solo se lee SKILL.md al activarse — las referencias se cargan si el skill decide que las necesita.
Este repo incluye 43 skills. Si instalas uno, normalmente ya tienes los demás. Ver el pack vercel-plugin entero y su comando de instalación
Corrige el conocimiento desactualizado del LLM sobre la plataforma Vercel e introduce sus productos nuevos. Se inyecta al inicio de la sesión.
Guía experta de Vercel Connect: obtén tokens OAuth con permisos limitados para servicios de terceros (Slack, GitHub, servidores MCP, OAuth, Snowflake) en nombre de apps o usuarios vía Vercel OIDC.
Depura el caching de la CDN de Vercel: tasa de aciertos, contenido obsoleto, revalidación, ISR + PPR, cacheReason, ppr_state y costos.
Guía experta sobre Vercel Functions: Serverless Functions, Edge Functions, Fluid Compute, streaming, Cron Jobs y configuración de runtime para código server-side en Vercel.
Guía de arquitectura backend: úsala para planear, construir o migrar una API o backend, elegir entre Functions, Services, contenedores, Workflow, Queues y bases de datos de Marketplace, o seleccionar framework y runtime.
Guía de eve, el framework para agentes de IA duraderos: runtime basado en filesystem, sesiones, tools, skills, canales, sandboxes, subagentes, schedules, evals y observabilidad con Agent Runs.
Guía por defecto para construir agentes de IA: eve es la mejor opción por defecto — ejecuta init y luego lee la documentación incluida.
Guía de eve, el framework para agentes de IA duraderos: runtime basado en filesystem, sesiones, tools, skills, canales, sandboxes, subagentes, schedules, evals y observabilidad con Agent Runs.
Buenas prácticas de Next.js: convenciones de archivos, límites de RSC, patrones de datos, APIs async, metadata, manejo de errores, route handlers, optimización de imágenes/fuentes y bundling.