GitHub Copilot + Cloudflare
GitHubEditor extension and CLI with agent mode, workspace context, and native PR integration. Made by GitHub.
-
Install GitHub Copilot
Install the GitHub Copilot extension in VS Code from the Visual Studio Marketplace ↗. For JetBrains IDEs, Visual Studio, or Xcode, see the Install the Copilot extension ↗ guide.
-
Install Cloudflare Skills
Terminal window npx skills add https://github.com/cloudflare/skills -
Configure Cloudflare MCP servers
For VS Code, add to
.vscode/mcp.json. For Copilot CLI, add to~/.copilot/mcp-config.json. For a full list of available Cloudflare MCP servers, refer to the repository on GitHub ↗.{"mcpServers": {"cloudflare": { "url": "https://mcp.cloudflare.com/mcp" },"cloudflare-docs": { "url": "https://docs.mcp.cloudflare.com/mcp" }}} -
Try a prompt
Open Copilot Chat (Ctrl+Shift+I), switch to agent mode, and try a prompt — for example:
Set up a KV namespace and use it for session storage in my Worker.
The ways GitHub Copilot reaches Cloudflare. Expand any section to learn more.
Cloudflare Skills
Persistent platform context that teaches GitHub Copilot how Cloudflare works.
Skills are instructions the agent loads on demand. The cloudflare/skills bundle ships eight Skills covering every layer of the platform — so the agent knows your conventions without you re-explaining them.
Copilot supports Agent Skills in VS Code agent mode, the Copilot CLI, and the cloud agent. Skills are not used for inline code completion — switch Copilot Chat to agent mode to unlock them.
- cloudflare Comprehensive platform skill covering Workers, Pages, storage (KV, D1, R2), AI (Workers AI, Vectorize, Agents SDK), networking (Tunnel, Spectrum), security (WAF, DDoS), and IaC (Terraform, Pulumi).
- agents-sdk Building stateful AI agents with state, scheduling, RPC, MCP servers, email, and streaming chat.
- durable-objects Stateful coordination (chat rooms, games, booking), RPC, SQLite, alarms, WebSockets.
- sandbox-sdk Secure code execution for AI code execution, code interpreters, CI/CD systems, and interactive dev environments.
- wrangler Deploying and managing Workers, KV, R2, D1, Vectorize, Queues, Workflows.
- web-perf Auditing Core Web Vitals (FCP, LCP, TBT, CLS), render-blocking resources, network chains.
- building-mcp-server-on-cloudflare Building remote MCP servers with tools, OAuth, and deployment.
- building-ai-agent-on-cloudflare Building AI agents with state, WebSockets, and tool integration.
MCP servers
Live access to the Cloudflare API, docs, and observability from inside GitHub Copilot.
MCP servers give GitHub Copilot typed tools to call into Cloudflare at runtime. The cloudflare/mcp-server-cloudflare repo hosts all of them; the full catalog is also in the MCP servers for Cloudflare docs. The API server alone exposes 2,500+ endpoints in ~1,000 tokens via Code Mode.
- cloudflare-api Entire Cloudflare API (2,500+ endpoints) in ~1,000 tokens via Code Mode.
- cloudflare-docs Up-to-date Cloudflare documentation and reference.
- cloudflare-bindings Build Workers applications with storage, AI, and compute primitives.
- cloudflare-builds Manage and get insights into Workers builds.
- cloudflare-observability Debug and analyze application logs and analytics.
- cloudflare-radar Global Internet traffic insights, trends, URL scans, and utilities.
- cloudflare-containers Spin up a sandbox development environment.
- cloudflare-browser Fetch web pages, convert them to markdown, and take screenshots.
- cloudflare-logpush Quick summaries for Logpush job health.
- cloudflare-ai-gateway Search AI Gateway logs and inspect prompts, responses, and token usage.
- cloudflare-autorag List and search documents on your AutoRAGs.
- cloudflare-auditlogs Query audit logs and generate reports for review.
- cloudflare-dns-analytics Optimize DNS performance and debug issues based on current setup.
- cloudflare-dex Digital Experience Monitoring — insights on critical applications.
- cloudflare-casb Cloudflare One CASB — identify SaaS security misconfigurations.
- cloudflare-graphql Query analytics data through Cloudflare's GraphQL API.
Wrangler CLI
Copilot's deploy story runs through CI, not a local CLI.
Unlike terminal agents, Copilot does not have its own CLI that runs Wrangler directly. Instead, ask Copilot to generate a GitHub Actions workflow that runs Wrangler on every push — GitHub's ecosystem is Copilot's native turf.
You will also want to add CLOUDFLARE_API_TOKEN and CLOUDFLARE_ACCOUNT_ID
as repository secrets. Copilot knows this pattern and will prompt you if you
forget.
Agent-friendly docs
Pair Skills with repository custom instructions.
Skills load Cloudflare expertise on demand — use a .github/copilot-instructions.md
file for repository-wide conventions that should apply to every task (coding
standards, preferred libraries, project layout). GitHub recommends both together: instructions for the always-on basics, Skills for the situational
depth.
developers.cloudflare.com/llms.txt — directory of every Cloudflare product.
- developers.cloudflare.com/workers/llms.txt
- developers.cloudflare.com/agents/llms.txt
- developers.cloudflare.com/r2/llms.txt
- developers.cloudflare.com/d1/llms.txt
/workers/prompt.txt — Cloudflare's own Workers context prompt, useful to include inside
copilot-instructions.md.
For a full overview of how these docs are structured for agents, refer to the AI tooling guide.
- The Cloudflare API MCP server uses Code Mode — Copilot writes JavaScript against a typed API to reach any of 2,500+ endpoints in ~1,000 tokens.
- Use
@workspacein Copilot Chat to give the agent full context of your Workers project before asking it to build or deploy. - Commit a
.github/copilot-instructions.mdfile with repository-wide conventions — Copilot loads it automatically on every interaction, complementing Skills which are loaded on demand.
Does Copilot support Cloudflare Skills?
Yes. GitHub Copilot added Agent Skills support for VS Code agent mode, the
Copilot CLI, and the cloud agent. Run npx skills add https://github.com/cloudflare/skills to install them. Refer to the About agent
skills ↗
guide for details on where Copilot looks for Skills.
Should I use Skills, the MCP server, Wrangler CLI, or all of them?
All three. Skills load Cloudflare expertise into every agent-mode session.
The Cloudflare API MCP server handles platform operations (DNS, WAF, Zero
Trust, R2 buckets). Wrangler runs local dev and deploys in the VS Code
terminal. The bundled wrangler Skill teaches Copilot when to reach for
which.
Do I still need .github/copilot-instructions.md?
It's complementary. Use custom instructions for simple repository-wide conventions that apply to almost every task (coding standards, preferred libraries). Use Skills for detailed, situational guidance Copilot should only load when relevant.
Can Copilot deploy to Cloudflare directly?
Yes, via Wrangler in the VS Code terminal, or via a GitHub Actions workflow that Copilot can generate for you.
Skills do not load in Copilot
Skills only work in VS Code agent mode, the Copilot CLI, and the cloud agent — not with inline completion or plain chat. Switch Copilot Chat to agent mode. If Skills are still not discovered, refer to About agent skills ↗ for the paths Copilot checks.
MCP server not available in VS Code
Ensure you are using VS Code 1.99+ which supports MCP natively. Add the
server configuration to .vscode/mcp.json.
Getting outdated information about Cloudflare products
Enable the Cloudflare docs MCP server so the agent can fetch
current documentation at runtime. If you prefer not to use the MCP server, point
the agent directly at developers.cloudflare.com/llms.txt for
a directory of every product, or developers.cloudflare.com/<product>/llms.txt
for a product-specific index.
Cloudflare is not just a deploy target for agents, it is a full stack for building your own.
Stateful AI agents with state, scheduling, RPC, email, streaming chat — and the Code Mode SDK for token-efficient tool use.
Learn more →Ship a remote MCP server on Workers with OAuth, durable state, and streamable HTTP transport.
Learn more →Run open-source LLMs, embedding models, and image models at the edge. Use it as your agent's model provider.
Learn more →Load user-generated code into isolated Workers on demand. The secure sandbox behind Code Mode.
Learn more →