Microsoft MCP server gives AI assistants access to MSBuild logs

Microsoft has introduced the Microsoft Binlog MCP Server, which gives AI assistants like GitHub Copilot direct access to MSBuild (.binlog) files. The Model Context Protocol server enables AI-powered build investigation through natural language conversation, Microsoft said.

Introduced June 17 and currently in a preview stage, the Microsoft Binlog MCP Server parses .binlog files and exposes 15 specialized tools that enable AI-driven diagnosis, property tracing, performance analysis, and build comparison. Microsoft said that AI assistants gain the ability to do the following:

• Investigate build failures by querying errors, warnings, and full project/target/task context
• Trace property origins to understand where a property got its value
• Analyze performance bottlenecks by identifying the slowest projects, targets, and tasks
• Compare two builds to spot differences in packages and properties
• Read embedded source files captured during the build

Instead of manually scrolling through the MSBuild Structured Log Viewer, Microsoft said that developers can ask their AI assistant questions like “Why did my build fail?” or “What’s making my build slow?” MSBuild’s binary logs contain detailed information about a build including every property evaluation, target execution, task invocation, error, and warning. Thus navigating that data manually can be overwhelming, especially when debugging a complex multi-project solution. Tapping an AI coding assistant for these investigations can save significant time and effort.

Microsoft noted that the easiest way to get started with the Microsoft Binlog MCP server is through the .NET Agent Skills repository. The dotnet-msbuild plugin, which is available for Visual Studio, Visual Studio Code, and terminal-based AI assistants such as GitHub Copilot CLI and Claude Code, bundles the Microsoft Binlog MCP Server along with curated skills and agents for MSBuild build investigation and optimization.