AI News Summary - June 11, 2026

Claude Code Introduces Multi-Repository Context Spanning

Anthropic announced Claude Code now supports simultaneous context awareness across up to 10 linked repositories, enabling developers to refactor code that spans microservices architectures. The feature uses an enhanced MCP integration that maintains cross-repo dependency graphs in real-time. Early adopters report 40% faster completion times for multi-service refactoring tasks.

OpenAI Releases Agent Protocol 2.0 Specification

OpenAI published Agent Protocol 2.0, a standardized API specification for AI agent communication and task delegation. The protocol includes built-in security controls, resource management, and inter-agent authentication mechanisms. Major frameworks including LangGraph, AutoGPT, and CrewAI have committed to implementing the specification by Q3 2026.

Cursor Editor Adds "Shadow Mode" for Learning Developer Patterns

Cursor's latest update introduces Shadow Mode, which silently observes coding patterns without making suggestions, then adapts its recommendations to match individual developer styles. The feature uses on-device learning to personalize completions while maintaining privacy. Users can review what the AI learned and selectively approve pattern recognition.

MCP Gateway Hub Reaches 5,000 Published Connectors

The Model Context Protocol ecosystem crossed 5,000 published connectors on the community-driven MCP Gateway Hub. Notable additions this week include native integrations for Kubernetes clusters, Snowflake data warehouses, and Figma design files. The hub now processes over 2 million MCP requests daily across participating AI applications.