AI News — 2026-09-12
AI News - September 12, 2026
GitHub Copilot Workspace Reaches General Availability with Enhanced Multi-File Refactoring
GitHub announced that Copilot Workspace has exited beta and is now generally available to all Copilot Business and Enterprise subscribers. The update introduces cross-repository context awareness, allowing the AI to understand dependencies across up to 10 linked repositories simultaneously. Early adopters report 40% faster completion times for large-scale refactoring tasks compared to the previous beta version.
Anthropic's MCP 2.1 Specification Adds Bidirectional Streaming Support
The Model Context Protocol specification has been updated to version 2.1, introducing bidirectional streaming capabilities that enable real-time data exchange between AI models and external tools. Major implementations from Microsoft, JetBrains, and Replit are expected within the next two weeks. The update addresses one of the most requested features from the developer community, particularly for database query tools and live debugging scenarios.
Open-Source Agent Framework "Conductor" Hits 50K GitHub Stars in First Month
Conductor, a new lightweight agent orchestration framework written in Rust, has rapidly gained adoption for its minimal overhead and native support for local model execution. The framework emphasizes deterministic behavior and includes built-in observability tools that make agent decision-making transparent to developers. The project has already attracted contributions from engineers at Anthropic, Meta, and several YC-backed startups.
Cursor IDE Introduces "Context Contracts" for Fine-Grained Privacy Controls
Cursor released version 0.42 featuring Context Contracts, a new system that allows developers to explicitly define which files, directories, and data sources the AI can access during coding sessions. This addresses enterprise concerns around accidental exposure of sensitive code or credentials to AI models. The feature works across both cloud and local model deployments, with zero-knowledge proofs ensuring compliance verification.