/ Verzeichnis / Playground / claude-skills
● Community secondsky ⚡ Sofort

claude-skills

von secondsky · secondsky/claude-skills

Production-ready Claude Code skills for Cloudflare, React, Tailwind v4, and AI integrations — the modern full-stack bundle.

A focused set of skills for the Cloudflare + React + Tailwind v4 + AI stack. Opinionated toward what actually ships in 2026: Workers, D1, R2, Pages on Cloudflare; React 19 + Tailwind v4 on the frontend; provider-agnostic AI integration patterns.

Warum nutzen

Hauptfunktionen

Live-Demo

In der Praxis

claude-skill-5.replay ▶ bereit
0/0

Installation

Repository des Autors · Offizielle Dokumentation zu Claude Skills

Anwendungsfälle

Praxisnahe Nutzung: claude-skills

Ship a new Cloudflare-hosted React app in one session

👤 Devs building on the CF stack ⏱ ~60 min intermediate

Wann einsetzen: You want Workers + Pages + D1 wired up without reading 5 Cloudflare docs.

Ablauf
  1. Scaffold
    Use claude-skills. Scaffold a React 19 + Tailwind v4 app deployed to Cloudflare Pages with a Workers API and D1 DB.✓ Kopiert
    → Monorepo structure with apps/web, apps/api (Worker), wrangler.toml
  2. Add a feature
    Add a 'notes' feature — D1 schema, CRUD Worker endpoints, React UI with Tailwind v4 styling.✓ Kopiert
    → Migration + Worker + React page
  3. Deploy
    Deploy to Cloudflare.✓ Kopiert
    → wrangler deploy output; URLs returned

Ergebnis: Live URL within an hour.

Fallstricke
  • Wrangler config drift between environments — Use wrangler environments from the start — dev/staging/prod sections
Kombinieren mit: cloudflare · cloudflare-api

Migrate a Tailwind v3 project to v4

👤 Teams on v3 hitting the v4 release ⏱ ~45 min intermediate

Wann einsetzen: You want the new engine + CSS-first config without breaking everything.

Ablauf
  1. Audit current
    Use claude-skills. Audit my tailwind.config.js for v4 breaking changes.✓ Kopiert
    → Change list with severity
  2. Migrate config
    Convert to v4 CSS-first config with @theme directives.✓ Kopiert
    → New globals.css with @theme block
  3. Fix utilities
    Find v3 utilities that changed or were removed; rewrite usage.✓ Kopiert
    → Grep + rewrite pass

Ergebnis: Project on v4 without visual regressions.

Fallstricke
  • Plugin ecosystem not v4-ready yet — Check each plugin's v4 status; stay on v3 if a critical one lags
Kombinieren mit: filesystem

Wrap multiple AI providers behind one adapter

👤 Devs who want to swap Claude / GPT / local models without rewriting ⏱ ~45 min advanced

Wann einsetzen: Locking into one provider feels wrong and you want a clean interface.

Ablauf
  1. Define interface
    Use claude-skills ai-adapter pattern. Define a TS interface that abstracts chat completion across Claude, OpenAI, and a local model.✓ Kopiert
    → Interface + 3 implementations
  2. Wire up streaming
    Add SSE streaming support to the adapter.✓ Kopiert
    → Streaming implementation consistent across providers

Ergebnis: A swappable AI layer.

Fallstricke
  • Over-abstracting and losing provider-specific features — Keep a provider-specific escape hatch alongside the shared interface

Kombinationen

Mit anderen MCPs für 10-fache Wirkung

claude-skill-5 + cloudflare

Deploy + manage the CF resources the scaffold creates

After scaffold, use cloudflare MCP to verify DNS and tail Worker logs.✓ Kopiert
claude-skill-5 + cloudflare-api

Bulk CF config operations

Configure cache rules and WAF for the deployed Pages project.✓ Kopiert

Werkzeuge

Was dieses MCP bereitstellt

WerkzeugEingabenWann aufrufenKosten
cf_scaffold app type, features New CF app 0
tailwind_migrate v3 config Tailwind upgrade 0
ai_adapter_scaffold providers list Multi-provider AI setup 0

Kosten & Limits

Was der Betrieb kostet

API-Kontingent
None in skill
Tokens pro Aufruf
Moderate
Kosten in €
Cloudflare has a generous free tier; paid features as you grow
Tipp
Use wrangler dev locally; reserve cloud deploys for staging/prod.

Sicherheit

Rechte, Secrets, Reichweite

Credential-Speicherung: Wrangler handles CF auth; AI provider keys via env / CF secrets
Datenabfluss: CF, AI providers you configure

Fehlerbehebung

Häufige Fehler und Lösungen

D1 migration fails

Wrangler D1 migrations are finicky; run with --local first and inspect the SQL

Prüfen: wrangler d1 migrations list <db>
Tailwind v4 styles not applying

Check the @source directive captures your content paths

Deploy succeeds but Worker returns 500

wrangler tail to see the actual error

Prüfen: wrangler tail <worker>

Alternativen

claude-skills vs. andere

AlternativeWann stattdessenKompromiss
Vercel / Next.js stackYou prefer Next ecosystemDifferent deployment model, pricing
cc-frontend-skillYou care most about visual identityNot stack-specific

Mehr

Ressourcen

📖 Offizielle README auf GitHub lesen

🐙 Offene Issues ansehen

🔍 Alle 400+ MCP-Server und Skills durchsuchen