Use this skill when the user asks about "Cloudflare Workers", "Cloudflare Pages", "Wrangler configuration", "R2", "D1", "KV", "Durable Objects", "Workers AI", or any Cloudflare deployment/development question. It uses the cloudflare-docs-fetcher agent to fetch official documentation, keeping the main context clean.
Installs into .claude/skills of the current project.
Are you the author of Cloudflare Guide?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/david-li0406-cloudflare-guide)
---
name: cloudflare-guide
description: Use this skill when the user asks about "Cloudflare Workers", "Cloudflare Pages", "Wrangler configuration", "R2", "D1", "KV", "Durable Objects", "Workers AI", or any Cloudflare deployment/development question. It uses the cloudflare-docs-fetcher agent to fetch official documentation, keeping the main context clean.
version: 1.0.0
context: fork
allowed-tools:
- Read
- Write
- Edit
- Glob
- Grep
- Bash
- Task
---
# Cloudflare Development Assistant
## Core Principle
**Use the `cloudflare-docs-fetcher` agent to fetch and analyze official Cloudflare documentation.**
This keeps the main conversation context clean while ensuring answers are based on the latest official documentation.
## How to Use
When a user asks a Cloudflare-related question, spawn the `cloudflare-docs-fetcher` agent:
```
Task(
subagent_type: "general-purpose",
description: "Fetch Cloudflare docs for [brief topic]",
prompt: "User question: [THE EXACT USER QUESTION]"
)
```
The agent will:
1. Determine which official docs URL to fetch
2. Fetch and analyze the documentation
3. Return a concise answer with exact syntax from the docs
## When to Trigger
Trigger this skill when the user mentions:
- **Workers**: runtime, handlers, bindings, Wrangler
- **Pages**: deployment, build config, Pages Functions
- **Storage**: R2, D1, KV, Durable Objects, Queues
- **AI**: Workers AI, AI Gateway, Agents SDK, Vectorize
- **Config**: `wrangler.toml`, `wrangler.jsonc`, compatibility flags
- **Deployment**: `wrangler deploy`, `wrangler dev`, environments
## Quick Commands (No Fetch Needed)
These common commands don't require documentation lookup:
```bash
# Create new project
npm create cloudflare@latest
# Local development
npx wrangler dev
# Deploy
npx wrangler deploy # Workers
npx wrangler pages deploy # Pages
# Manage secrets
npx wrangler secret put <KEY>
# D1 database
npx wrangler d1 create <NAME>
npx wrangler d1 execute <DB> --command "SQL"
# KV namespace
npx wrangler kv:namespace create <NAME>
# R2 bucket
npx wrangler r2 bucket create <NAME>
```
## Example Interaction
**User**: "How do I add a D1 database binding to my Worker?"
**Your action**:
```
Task(
subagent_type: "general-purpose",
description: "Fetch D1 binding docs",
prompt: "User question: How do I add a D1 database binding to my Worker?"
)
```
**Agent returns** (concise answer with exact wrangler.toml syntax)
**You**: Provide the answer to the user, citing the official source.
## What NOT to Do
1. ❌ **Don't WebFetch in main process** - Use the agent to keep context clean
2. ❌ **Don't guess config syntax** - Let the agent fetch current docs
3. ❌ **Don't skip for "simple" questions** - Cloudflare APIs change frequently
## Additional Resources
- **`../agents/cloudflare-docs-fetcher.md`** - The agent that fetches official docs
- **`references/official-sources.md`** - Complete URL index