cloudflare-worker-builder - Claude MCP Skill

# Cloudflare Worker Builder

Scaffold a working Cloudflare Worker project from a brief description. Produces a deployable project with Hono routing, Vite dev server, and Static Assets.

## Workflow

### Step 1: Understand the Project

Ask about the project to choose the right bindings and structure:

- What does the app do? (API only, SPA + API, landing page)
- What data storage? (D1 database, R2 files, KV cache, none)
- Auth needed? (Clerk, better-auth, none)
- Custom domain or workers.dev subdomain?

A brief like "todo app with database" is enough to proceed.

### Step 2: Scaffold the Project

```bash
npm create cloudflare@latest my-worker -- --type hello-world --ts --git --deploy false --framework none
cd my-worker
npm install hono
npm install -D @cloudflare/vite-plugin vite
```

Copy and customise the asset files from this skill's `assets/` directory:
- `wrangler.jsonc` — Worker configuration
- `vite.config.ts` — Vite + Cloudflare plugin
- `src/index.ts` — Hono app with Static Assets fallback
- `package.json` — Scripts and dependencies
- `tsconfig.json` — TypeScript config
- `public/index.html` — SPA entry point

### Step 3: Configure Bindings

Add bindings to `wrangler.jsonc` based on project needs. Wrangler 4.45+ auto-provisions resources on first deploy — always specify explicit names:

```jsonc
{
  "name": "my-worker",
  "main": "src/index.ts",
  "compatibility_date": "2025-11-11",
  "assets": {
    "directory": "./public/",
    "binding": "ASSETS",
    "not_found_handling": "single-page-application",
    "run_worker_first": ["/api/*"]
  },
  // Add as needed:
  "d1_databases": [{ "binding": "DB", "database_name": "my-app-db" }],
  "r2_buckets": [{ "binding": "STORAGE", "bucket_name": "my-app-files" }],
  "kv_namespaces": [{ "binding": "CACHE", "title": "my-app-cache" }]
}
```

### Step 4: Deploy

```bash
npm run dev           # Local dev at http://localhost:8787
wrangler deploy       # Production deploy
```

---

## Critical Patterns

### Export Syntax

```typescript
// CORRECT — use this pattern
export default app

// WRONG — causes "Cannot read properties of undefined"
export default { fetch: app.fetch }
```

Source: [honojs/hono #3955](https://github.com/honojs/hono/issues/3955)

### Static Assets + API Routes

Without `run_worker_first`, SPA fallback intercepts API routes and returns `index.html` instead of JSON:

```jsonc
"assets": {
  "not_found_handling": "single-page-application",
  "run_worker_first": ["/api/*"]  // CRITICAL
}
```

Source: [workers-sdk #8879](https://github.com/cloudflare/workers-sdk/issues/8879)

### Vite Config

```typescript
import { defineConfig } from 'vite'
import { cloudflare } from '@cloudflare/vite-plugin'

export default defineConfig({ plugins: [cloudflare()] })
```

Always set the `main` field in wrangler.jsonc — the Vite plugin needs it.

### Scheduled/Cron Handlers

When adding cron triggers, switch to explicit export:

```typescript
export default {
  fetch: app.fetch,
  scheduled: async (event, env, ctx) => { /* ... */ }
}
```

---

## Reference Files

Read these for detailed troubleshooting:

- `references/common-issues.md` — 10 documented issues with sources and fixes
- `references/architecture.md` — Route priority, caching, Workers RPC
- `references/deployment.md` — CI/CD, auto-provisioning, gradual rollouts