Cloudflare

Cloudflare Skill

Comprehensive assistance with Cloudflare's developer platform, including Workers, AI Agents, AI Gateway, Pages, R2 storage, D1 databases, Durable Objects, and AI Search (RAG).

When to Use This Skill

This skill should be triggered when:

• Building serverless applications with Cloudflare Workers
• Creating AI agents using the Agents SDK or Model Context Protocol (MCP)
• Integrating AI models through AI Gateway or Workers AI
• Implementing Retrieval Augmented Generation (RAG) with AI Search
• Deploying static sites with Cloudflare Pages
• Working with object storage via R2
• Managing serverless databases with D1
• Building stateful applications with Durable Objects
• Setting up CDN configurations or edge computing
• Implementing Zero Trust security patterns
• Troubleshooting Cloudflare API integrations
• Learning Cloudflare best practices and patterns

Quick Reference

1. Basic Cloudflare Worker

export default {
  async fetch(request, env, ctx) {
    return new Response('Hello from Cloudflare Workers!', {
      headers: { 'Content-Type': 'text/plain' }
    });
  }
};

2. Worker with AI Gateway Integration

export default {
  async fetch(request, env, ctx) {
    const response = await env.AI_GATEWAY.run('@cf/meta/llama-2-7b-chat-int8', {
      messages: [
        { role: 'user', content: 'What is Cloudflare?' }
      ]
    });

    return Response.json(response);
  }
};

3. Building an AI Agent with Agents SDK

import { Agent } from '@cloudflare/agents-sdk';

export default {
  async fetch(request, env, ctx) {
    const agent = new Agent({
      model: 'gpt-4',
      tools: [
        {
          name: 'get_weather',
          description: 'Get current weather',
          parameters: {
            type: 'object',
            properties: {
              location: { type: 'string' }
            }
          },
          execute: async ({ location }) => {
            // Implementation here
            return `Weather in ${location}: Sunny`;
          }
        }
      ]
    });

    const result = await agent.run(
      'What is the weather in San Francisco?'
    );

    return Response.json(result);
  }
};

4. MCP Server on Workers

import { McpAgent } from '@cloudflare/agents-sdk';

export default {
  async fetch(request, env, ctx) {
    const agent = new McpAgent({
      name: 'my-mcp-server',
      version: '1.0.0',
      tools: [
        {
          name: 'calculate',
          description: 'Perform calculations',
          inputSchema: {
            type: 'object',
            properties: {
              operation: { type: 'string' },
              a: { type: 'number' },
              b: { type: 'number' }
            }
          },
          handler: async ({ operation, a, b }) => {
            if (operation === 'add') return a + b;
            if (operation === 'multiply') return a * b;
            throw new Error('Unknown operation');
          }
        }
      ]
    });

    return agent.handleRequest(request);
  }
};

5. AI Search (RAG) with Workers Binding

export default {
  async fetch(request, env, ctx) {
    const query = 'How do I deploy a Worker?';

    const result = await env.AI_SEARCH.search({
      query,
      top_k: 5,
      include_metadata: true
    });

    return Response.json(result);
  }
};

6. R2 Storage Integration

export default {
  async fetch(request, env, ctx) {
    // Upload file to R2
    await env.MY_BUCKET.put('example.txt', 'Hello R2!');

    // Read file from R2
    const object = await env.MY_BUCKET.get('example.txt');
    const text = await object.text();

    return new Response(text);
  }
};

7. D1 Database Query

export default {
  async fetch(request, env, ctx) {
    const results = await env.DB.prepare(
      'SELECT * FROM users WHERE email = ?'
    ).bind('user@example.com').all();

    return Response.json(results);
  }
};

8. Durable Objects for Stateful Applications

export class Counter {
  constructor(state, env) {
    this.state = state;
  }

  async fetch(request) {
    let count = await this.state.storage.get('count') || 0;
    count++;
    await this.state.storage.put('count', count);

    return new Response(count.toString());
  }
}

export default {
  async fetch(request, env, ctx) {
    const id = env.COUNTER.idFromName('global');
    const obj = env.COUNTER.get(id);
    return obj.fetch(request);
  }
};

9. AI Gateway with Caching and Rate Limiting

const response = await fetch(
  'https://gateway.ai.cloudflare.com/v1/{account_id}/{gateway_id}/openai/chat/completions',
  {
    method: 'POST',
    headers: {
      'Authorization': `Bearer ${env.OPENAI_API_KEY}`,
      'Content-Type': 'application/json',
      'cf-aig-cache-ttl': '3600',
      'cf-aig-skip-cache': 'false'
    },
    body: JSON.stringify({
      model: 'gpt-4',
      messages: [{ role: 'user', content: 'Hello!' }]
    })
  }
);

10. Human-in-the-Loop Agent Pattern

import { Agent } from '@cloudflare/agents-sdk';

export default {
  async fetch(request, env, ctx) {
    const agent = new Agent({
      model: 'gpt-4',
      tools: [
        {
          name: 'send_email',
          description: 'Send email (requires approval)',
          requiresApproval: true,
          execute: async ({ to, subject, body }) => {
            // Wait for human approval
            const approved = await requestApproval({
              action: 'send_email',
              params: { to, subject, body }
            });

            if (!approved) {
              throw new Error('Action not approved');
            }

            // Send email
            return { status: 'sent' };
          }
        }
      ]
    });

    return agent.handleRequest(request);
  }
};

Reference Files

This skill includes comprehensive documentation in references/:

llms.md

Complete Cloudflare developer documentation covering:

• Agents: Build AI agents with tools, workflows, and MCP integration
• AI Search: Fully-managed RAG applications with retrieval-augmented generation
• AI Gateway: Control, monitor, and optimize AI API usage with caching, rate limiting, and observability
• Workers AI: Run AI models on Cloudflare's edge network
• Browser Rendering: Automate browsers at the edge
• Constellation: ML inference at the edge
• Containers: Deploy containerized applications

other.md

Additional platform documentation including:

• D1: Serverless SQL database
• Durable Objects: Stateful serverless compute
• R2: Object storage with S3-compatible API
• Pages: Static site hosting with edge functions
• Zero Trust: Security and access control
• CDN: Content delivery and optimization

Use the reference files when you need:

• Detailed API specifications
• Advanced configuration options
• Platform limits and pricing details
• Step-by-step tutorials
• Best practices and design patterns

Working with This Skill

For Beginners

Start with these core concepts:

1. Workers Basics: Understand the fetch handler pattern and edge computing
2. Bindings: Learn how to connect Workers to R2, D1, Durable Objects, and AI services
3. Getting Started Guides: Follow the tutorials in the reference docs for hands-on learning
4. Local Development: Use Wrangler CLI for local testing and deployment

Key resources:

• Getting started guides for Workers, Pages, and AI
• Simple examples above (#1, #6, #7)

For Building AI Applications

Focus on these areas:

1. Agents SDK: Build AI agents with tool calling and workflows
2. AI Gateway: Route and optimize LLM requests
3. AI Search (RAG): Implement retrieval-augmented generation for context-aware responses
4. MCP Integration: Create or consume Model Context Protocol servers
5. Workers AI: Run AI models directly on the edge

Key resources:

• Examples #2-5, #9-10
• Agent patterns and MCP documentation in llms.md

For Advanced Features

Explore these capabilities:

1. Durable Objects: Build stateful applications with strong consistency
2. Dynamic Routing: Implement intelligent request routing in AI Gateway
3. Guardrails & DLP: Add content filtering and data loss prevention
4. Human-in-the-Loop: Implement approval workflows for sensitive operations
5. Evaluation Frameworks: Test and improve AI agent performance
6. Custom Costs: Track and optimize AI API spending

Key resources:

• Advanced examples (#8, #10)
• Platform-specific documentation in both reference files

Key Concepts

Workers

Serverless JavaScript/TypeScript runtime running on Cloudflare's edge network. Workers handle HTTP requests using the fetch event handler.

Bindings

Connections between Workers and other Cloudflare services (R2, D1, AI Gateway, etc.). Configured in wrangler.toml and accessed via the env parameter.

Agents SDK

Framework for building AI agents with:

• Tools: Functions that agents can call
• Workflows: Multi-step agent orchestrations
• MCP Support: Standard protocol for AI tool integration

Model Context Protocol (MCP)

Open standard for connecting AI systems to data sources and tools. Cloudflare supports both MCP clients and servers.

AI Gateway

Unified interface for accessing multiple AI providers with:

• Request routing and fallbacks
• Caching for cost optimization
• Rate limiting and access control
• Observability and analytics
• DLP and guardrails

AI Search (RAG)

Fully-managed retrieval-augmented generation system that:

• Indexes documents from R2, websites, or APIs
• Performs semantic search using embeddings
• Generates context-aware responses
• Handles chunking, query rewriting, and caching

Durable Objects

Stateful serverless compute primitives that provide:

• Strong consistency guarantees
• Persistent storage per object
• Global uniqueness and routing
• WebSocket support

Edge Computing

Running code at Cloudflare's edge locations (300+ cities worldwide) for:

• Lower latency (closer to users)
• Higher performance
• Global distribution
• Cost efficiency

Best Practices

Workers

• Keep Workers lightweight (< 1MB after compression)
• Use async/await for all I/O operations
• Leverage caching with Cache API
• Handle errors gracefully with try/catch

AI Agents

• Define clear, concise tool descriptions
• Implement proper error handling in tools
• Use structured outputs for consistency
• Test agents thoroughly before production
• Implement rate limiting for external API calls

AI Gateway

• Enable caching for repeated queries
• Set appropriate rate limits
• Use custom metadata for tracking
• Monitor costs and usage patterns
• Implement fallback providers

Security

• Never hardcode API keys (use environment variables)
• Validate all user inputs
• Use AI Gateway authentication
• Implement guardrails for sensitive operations
• Enable DLP for PII protection

Resources

Official Documentation

• Cloudflare Developers
• Agents SDK Docs
• AI Gateway Docs
• Workers Docs

Developer Tools

• Wrangler CLI: Local development and deployment
• Dashboard: Web-based management console
• API: Programmatic resource management

Community

• Discord: Cloudflare Developers community
• GitHub: Official examples and templates
• Blog: Product updates and tutorials

Notes

• This skill was automatically generated from official documentation
• Reference files preserve the structure and examples from source docs
• Code examples include language detection for syntax highlighting
• All examples use modern JavaScript/TypeScript patterns
• Workers runtime is V8-based with Web Standards APIs

Updating

To refresh this skill with updated documentation:

1. Re-run the scraper with the same configuration
2. The skill will be rebuilt with the latest information
3. Check changelog for breaking changes or new features