Imagine harnessing advanced AI agent workflows directly inside Visual Studio Code without compromising on security or usability. With the arrival of full Model Context Protocol (MCP) support, developers can now integrate sophisticated AI tools, resources, and secure connections, all from their favorite code editor. This update positions VS Code as a central platform for building context-aware, automated development experiences.
Full Specification Support: What’s New?
VS Code now implements the entire MCP specification, expanding beyond basic tools and workspace awareness. Here’s what stands out:
- Authorization: Secure authentication is now straightforward, thanks to support for industry-leading identity providers like Okta, Stytch, and Descope. Forget the hassle of custom OAuth: connect seamlessly to remote MCP servers using trusted workflows.
- Prompts: Access dynamic, context-sensitive prompts with simple slash commands. These extendable workflows react to your current project, making automation and customization more accessible than ever.
- Resources: Work with semantic objects like images, logs, and more. These resources can be annotated, shared, or directly manipulated within your workspace, empowering real-time collaboration and productivity.
- Sampling: Let MCP servers tap into your existing language model subscriptions. This removes the overhead of managing separate AI SDKs, allowing for smoother reasoning and coordination in complex coding scenarios.
Security-First Approach: Next-Gen Authorization
The new authorization flow is a game changer. Developed with the help of major tech partners, it separates resource providers from authorization servers for greater security and flexibility. OAuth-based authentication is now built-in, making it easy to connect to platforms like GitHub securely. The GitHub MCP server, for instance, lets you manage repositories and issues using VS Code’s own authentication system, eliminating extra steps.
By supporting streamable HTTP transport, VS Code enables scalable, cloud-based MCP integrations that meet enterprise security standards. Developers can confidently build and extend remote MCP servers while keeping data protected.
Beyond Traditional Tools: Prompts, Resources, and Sampling
MCP’s capabilities go well beyond standard extensions. Prompts are now interactive workflows triggered from chat or the command palette, tailored to your current development context. This brings smart automation to everyday tasks, letting you launch and customize workflows on demand.
With resources, the boundary between static data and actionable insight disappears. For example, a screenshot from a Playwright MCP server becomes a live object, ready for code reviews or instant sharing. Debug logs can stream in real time, turning routine outputs into collaborative resources.
Sampling empowers MCP servers to make direct requests to your preferred language models, supporting advanced scenarios like multi-agent collaboration and complex problem solving. Users retain control over privacy, costs, and which models are utilized.
How to Get Started
The complete MCP suite is now open to all VS Code users. Connect to the GitHub MCP server for secure repository management, explore new servers offering rich resources, or develop your own to take full advantage of MCP’s features.
To begin, visit the official VS Code MCP documentation, check out the server repository, or review the MCP specification. This release firmly establishes VS Code as a foundational client in the AI-driven development ecosystem.
Conclusion
With MCP’s full integration, Visual Studio Code offers a robust, secure, and flexible platform for next-generation AI development. Developers can now create, automate, and collaborate with confidence, leveraging powerful tools and workflows right where they code. The future of AI-powered software creation starts here, how will you innovate?
Source: Visual Studio Code Blog
Unlocking Advanced AI Workflows: Full MCP Support Now in Visual Studio Code