GitLab and Anthropic rebuild Git for AI agents

GitLab is developing a new source-code engine designed to handle thousands of AI agents working in parallel. Current Git architecture struggles with the load when multiple agents clone repos simultaneously. Internal tests show the new engine uses 2x fewer tokens, runs 50x faster, and reduces network traffic by 1000x.
GitLab Orbit, a context graph for the development lifecycle, was also introduced. It ties together code, issues, pipelines, and production signals into a unified structure. Compare the Market found that using the graph boosted code comment accuracy to 70 percent from 58 percent with standard methods.
The new SCM and governance tools are in closed beta.
Up next
Claude Code extension lands for Visual Studio
A new extension brings Claude Code edits into Visual Studio’s native diff window. Users can review, accept, or reject changes right in VS. The plugin also pipes C# and C++ compilation errors straight into the model context.
Related stories
- OpenAI’s Pando cut monorepo push latency nearly 5x
- Claude expands MCP Apps support with app tools and more
- Anthropic hands Agent Skills to the foundation behind MCP
- Claude now picks its own agent team in Managed Agents
- Anthropic’s daily-brief agent treats a failed read as news
- Claude’s SDKs run the click loop, but the browser is yours
Comments
No comments yet. Be the first.
Join the conversation
Sign in to leave a comment. It appears after moderation.

