AI News Summary - July 6, 2026

Claude Code Expands Multi-Repository Support

Anthropic's Claude Code now supports simultaneous context from up to 10 linked repositories, enabling developers to work across microservices architectures more efficiently. The update includes improved dependency tracking and automatic detection of cross-repo function calls. Early adopters report 40% faster navigation in complex codebases.

MCP 3.0 Specification Released with Streaming Context

The Model Context Protocol consortium unveiled MCP 3.0, introducing bidirectional streaming context updates and enhanced security primitives for enterprise deployments. The new specification enables real-time collaborative editing scenarios where multiple AI agents can share context without redundant data transfers. Over 200 tools have committed to supporting the updated protocol by Q3 2026.

Open-Source AgentFlow Framework Reaches Production Readiness

AgentFlow 1.0, an open-source framework for building stateful AI agent pipelines, launched with built-in observability and debugging tools. The framework features declarative workflow definitions, automatic checkpointing, and native integrations with popular vector databases. The project has already garnered 12K GitHub stars since its beta release in May.

GitHub Copilot Workspace Adds Natural Language Architecture Planning

GitHub introduced Architecture Mode in Copilot Workspace, allowing developers to describe system designs in natural language and receive structured implementation plans with file scaffolding. The feature leverages improved reasoning models to suggest design patterns, estimate complexity, and identify potential technical debt. The update rolls out to Enterprise customers this week.