AI News Summary - May 26, 2026

Claude Code Launches Multi-Repository Context Feature

Anthropic released a major update to Claude Code that enables simultaneous context awareness across multiple linked repositories. The feature uses an enhanced MCP server architecture to maintain consistency when working on microservices and monorepo projects. Early adopters report 40% faster refactoring workflows when dealing with cross-repository dependencies.

GitHub Copilot Workspace Adds Autonomous Testing Capabilities

GitHub announced that Copilot Workspace can now automatically generate, run, and iterate on test suites without developer intervention. The system uses a new agent framework that analyzes code changes, identifies edge cases, and writes comprehensive test coverage in multiple testing frameworks. The feature is rolling out to GitHub Enterprise customers first, with general availability planned for June.

MCP 2.1 Specification Released with Streaming Protocol Support

The Model Context Protocol consortium published version 2.1, introducing native support for streaming context updates and bidirectional communication channels. This enables AI coding assistants to receive real-time file system changes, git events, and IDE state updates without polling. Major implementers including Cursor, Zed, and JetBrains have committed to adoption within 30 days.

OpenHands Agent Framework Reaches Production Readiness

The open-source OpenHands project announced its 1.0 release, providing a stable foundation for building autonomous development agents. The framework now includes pre-built templates for common workflows like dependency updates, bug triage, and documentation generation. Over 200 companies are already using OpenHands in production environments according to the project's telemetry data.