AI News Brief - May 3, 2026

Claude Code Launches Multi-Repository Context Feature Anthropic released an update to Claude Code enabling developers to work across up to 10 linked repositories simultaneously with unified context awareness. The feature uses an improved codebase indexing system that maintains semantic understanding across project boundaries, reducing the need for manual context switching. Early adopters report 40% faster implementation times for microservices architectures.

MCP 2.1 Specification Adds Streaming Data Sources The Model Context Protocol consortium published version 2.1, introducing native support for real-time streaming data sources including database change feeds and event streams. This enables AI agents to maintain always-current context from production systems without polling overhead. Major vendors including Microsoft, Google, and Anthropic have committed to implementation within 30 days.

Open-Source Agent Framework "Conductor" Reaches 1.0 The Conductor project, an agent orchestration framework built on Rust, hit stable 1.0 release after 18 months of development. It features built-in MCP support, sub-100ms inter-agent messaging, and declarative workflow definitions that compile to optimized execution graphs. The framework already powers agent systems at three Fortune 500 companies.

GitHub Copilot Workspace Adds Visual Debugging for AI Suggestions GitHub introduced an interactive debugging interface that visualizes how Copilot generates code suggestions, showing the retrieved context, reasoning steps, and confidence scores. Developers can now adjust context windows and see real-time updates to suggestions, addressing long-standing concerns about AI coding assistant transparency. The feature enters public beta next week.