Weekly GitHub Copilot Roundup: Agents, MCP, and AI Code Review
This week, GitHub Copilot received significant enhancements, reinforcing its drive toward streamlined automation, extensibility, and practical productivity for both enterprise and individual developers. Updates advanced code review automation, cross-IDE agent support, robust contextual intelligence with MCP, reporting capabilities, and operational transparency.
AI-Assisted Code Review and Workflow Automation
Copilot’s AI-powered code review, after internal and early previews, is now automating over 600,000 PRs monthly at Microsoft. It surfaces issues, summarizes PRs, and supports interactive queries, speeding up reviews by up to 20% and supporting compliance through customizable instructions. Guidance is consolidated in copilot-instructions.md, with new UI enhancements, keeping developer ownership over merges. These features build on Copilot code review’s rollout in GitHub Mobile and support for structured, scalable human-AI collaboration in code reviews.
- Enhancing Code Quality at Scale with AI-Powered Code Reviews
- Upcoming deprecations and changes to Copilot code review
- Code review in the age of AI - Why developers will always own the merge button
Coding Agent Capabilities Expand Across IDEs
Copilot coding agents are now generally available not only in VS Code but also JetBrains, Eclipse, and Xcode. Developers initiate and monitor autonomous code tasks within any major IDE, reducing context switching and bottlenecks. Enhanced security defaults restrict agent internet access, with detailed configuration options meeting enterprise policies. Broad agent availability and secure, cross-platform automation signal GitHub’s commitment to robust, organization-ready deployment.
- Start and track GitHub Copilot coding agent sessions from Visual Studio Code
- Agent mode for JetBrains, Eclipse, and Xcode is now generally available
- Configure internet access for Copilot coding agent
- From chaos to clarity - Using GitHub Copilot agents to improve developer workflows
- Finish your work without touching any code
Model Context Protocol (MCP) and Extensibility
Copilot’s Model Context Protocol (MCP) is now fully live in VS Code, enabling real-time integration with external tools and data sources. This brings contextually aware AI support, automating tasks using live project data, APIs, and databases, and supporting hands-on workflow automation. VS Code’s v1.102 release leverages MCP for cross-environment collaboration, with tutorials guiding effective team adoption.
- Model Context Protocol (MCP) support in VS Code is generally available
- VS Code Live - Exploring v1.102 Release Features—AI Chat, MCP, Coding Agent & More
- Model selection
- Automating Developer Tasks with GitHub Copilot Agent Mode and MCP Servers in VS Code
Productivity and Reporting Enhancements
Copilot adds support for interactive issue forms on github.com and enhances administrative oversight with improved activity reports, offering granular insights into usage and authentication for license management and targeted support. Copilot Chat supports early software planning, helping teams clarify requirements and reduce rework.
- Support for issue forms in chat and file uploads in spaces
- New GitHub Copilot activity report with enhanced authentication and usage insights
- Use GitHub Copilot Chat to Plan Your Software Before Coding
Availability, Incident Management, and Community Engagement
The June GitHub Availability Report reviews resilience efforts, transparency in incident management, and plans for improved monitoring. The “For the Love of Code” summer hackathon encourages creative use of Copilot and open source, sustaining community engagement and informing future product directions.