Visual Studio Code 1.105: AI Enhances Developer Workflow

Visual Studio Code 1.105 Introduces Advanced AI for Developers

Microsoft has unveiled Visual Studio Code 1.105, the latest release of its widely used code editor, bringing a suite of new AI-driven functionalities aimed at streamlining the development process. This update, also known as the September 2025 release, introduces intelligent solutions for common coding hurdles, most notably AI-assisted merge conflict resolution and an integrated marketplace for Model Context Protocol (MCP) servers. Developers can now download this version for Windows, Mac, and Linux from the official Visual Studio Code website.

The integration of artificial intelligence into VS Code reflects a growing trend in software development to leverage advanced algorithms for improved productivity and problem-solving. These new features are designed to reduce manual effort, offer deeper insights into code generation, and foster a more efficient coding environment. The focus on AI assistance underscores Microsoft’s commitment to evolving the developer experience through innovative technologies.

This release builds upon previous updates, such as VS Code 1.104, which brought automatic model selection for chat and enhanced agent security. The continuous introduction of AI capabilities positions Visual Studio Code as a leading tool for modern developers, equipped to handle increasingly complex coding tasks with intelligent support. Developers are encouraged to explore these new features to optimize their workflows and embrace the future of AI-powered coding.

Intelligent Conflict Resolution and Enhanced Chat Features

One of the most significant additions in Visual Studio Code 1.105 is the introduction of AI assistance for resolving Git merge conflicts. When a developer opens a file containing merge conflict markers, a new action becomes available in the editor’s lower right-hand corner. Selecting this action initiates an agentic flow within the Chat view, providing the merge base and changes from each branch as context for the AI.

This AI-powered feature aims to significantly reduce the time and complexity associated with manually resolving merge conflicts, a common and often frustrating task for developers. By offering intelligent suggestions and insights, the AI helps identify potential solutions and integrate changes more smoothly, ultimately improving collaboration and code integrity. This marks a substantial step towards automating traditionally manual and error-prone development processes.

Beyond conflict resolution, VS Code 1.105 also enhances the chat experience with several improvements. It now supports resuming recent chat sessions, allowing developers to pick up conversations with AI agents exactly where they left off. This continuity is crucial for maintaining context during extended debugging or development tasks, preventing the need to re-explain previous issues or objectives to the AI.

Another notable chat enhancement is the ability to view “thinking tokens” as expandable sections within AI responses, especially with the introduction of models like GPT-5-Codex. This feature allows developers to examine the model’s chain of thought, offering transparency into how the AI arrived at its suggestions. This insight can be invaluable for debugging, validating AI-generated code, or simply understanding the underlying logic, thereby fostering trust and improving the utility of AI assistance.

For screen reader users, the new accessibility.verboseChatProgressUpdates setting enables more detailed announcements regarding chat activity. This commitment to accessibility ensures that AI-powered features are usable by a wider range of developers. Additionally, keyboard shortcuts have been added for navigating up and down through chat messages, further enhancing efficiency and accessibility within the chat interface.

The Model Context Protocol Marketplace and Tooling Updates

Visual Studio Code 1.105 introduces a built-in marketplace specifically for Model Context Protocol (MCP) servers, which is accessible directly from the Extensions view. This integrated marketplace leverages the GitHub MCP Registry, providing developers with a streamlined experience for discovering, browsing, and installing MCP servers within their editor. The MCP marketplace simplifies the process of integrating various AI models and tools, making it easier for developers to customize their AI-powered coding environment.

The Model Context Protocol (MCP) plays a crucial role in enabling diverse AI models to interact seamlessly with the VS Code environment. By offering a centralized marketplace, Microsoft aims to foster a richer ecosystem of AI tools and services that developers can easily adopt. This not only enhances the editor’s capabilities but also empowers developers to experiment with different AI models tailored to specific tasks or programming languages.

Accompanying the MCP marketplace are significant improvements to tool naming conventions. VS Code 1.105 now supports fully qualified tool names in prompt files and chat modes. This enhancement is designed to prevent naming conflicts between built-in tools and those provided by MCP servers or extensions. Tool names are now qualified by the MCP server, extension, or tool set they belong to, ensuring clarity and preventing ambiguities in complex AI-driven workflows.

While developers can still use previous notation, a convenient code action is available to help migrate existing projects to the new, more robust naming scheme. This forward-looking change ensures better organization and compatibility as the ecosystem of AI-powered tools within VS Code continues to expand. It reflects a proactive approach to managing complexity and maintaining a consistent, reliable development environment.

Furthermore, the set of edit tools for Bring Your Own Key (BYOK) custom models has been improved for better integration with VS Code’s built-in tools. This enhancement provides greater flexibility for developers who wish to use their proprietary or specialized AI models within the editor, ensuring a smooth and cohesive experience. The runTests tool also now reports test code coverage to the agent, enabling the AI to generate and verify tests that comprehensively cover the code, further automating and refining the quality assurance process.

Broader Enhancements and Platform Support

Beyond the core AI features, Visual Studio Code 1.105 includes several other significant enhancements that improve overall platform support and developer experience. One such update is the introduction of macOS native broker support for Microsoft Authentication. This feature allows developers using M-series macOS devices that are Intune-enrolled to sign in through a native experience, leveraging the Microsoft Authentication Library (MSAL).

This native integration simplifies the authentication process for macOS users, providing a more seamless and secure login experience. It addresses specific needs of enterprise environments where Intune enrollment and secure authentication are paramount, ensuring that developers can access their resources efficiently and safely. The focus on native platform integration underscores Microsoft’s commitment to optimizing the experience for all users across different operating systems.

Another important update involves the default behavior of the “Delegate to coding agent” button in the Chat view. When working in a repository that has the Copilot coding agent enabled, this button now appears by default. This subtle yet impactful change streamlines the interaction with Copilot, making it more readily available for assistance without requiring manual activation. This reflects a trend towards making AI assistance more pervasive and integrated into the everyday coding workflow.

These broader enhancements, alongside the AI-driven features, demonstrate a comprehensive approach to improving Visual Studio Code. From facilitating easier authentication on specific platforms to refining interactions with AI coding agents, Microsoft continues to iterate on its popular editor. The sum of these updates contributes to a more powerful, intuitive, and efficient development environment for programmers across various disciplines and platforms.