Skills pour agents IA

vercel-optimize · Vercel Labs

Vercel Optimize

S’appuie sur les métriques Vercel pour classer les gains de coût, performance, cache et fiabilité.

Instructions originales de la skill

Le modèle technique reste en anglais afin de fonctionner directement avec les agents compatibles. Les explications, critères de choix et vérifications sont localisés.

Installation

npx skills add vercel-labs/agent-skills --skill vercel-optimize

Agent Skills CLI

Idéal pour

  • Reducing function and bandwidth spend
  • Investigating slow or expensive routes
  • Finding caching, ISR, image, and build-minute opportunities

Ce que la skill apporte

  • Collects account, usage, and route metrics before deciding which code paths deserve attention.
  • Connects observed production behavior to specific routes, configuration, and implementation evidence.
  • Produces a ranked report that separates measured facts, documentation, and recommendations.

Workflow conseillé

  • Confirm the CLI version, authenticated scope, linked project, and analysis window.
  • Collect Vercel metrics and usage data, then inspect only the implicated routes and files.
  • Rank changes by expected impact, confidence, effort, and a concrete way to verify the result.

Points à vérifier avant validation

  • Metrics and billing data can be sensitive; use the narrowest account and project scope.
  • Do not apply caching or architecture changes until correctness and freshness requirements are understood.
  • Re-measure after changes instead of treating estimated savings as achieved results.

Source et maintenance

Vercel Labs · vercel-agent-skills