Introduces git-inspired primitives to enable truly asynchronous and non-interfering multi-agent software engineering collaboration.
March 24, 2026
Original Paper
Effective Strategies for Asynchronous Software Engineering Agents
arXiv · 2603.21489
The Takeaway
It addresses the core friction in multi-agent SWE—concurrent edit interference—by implementing centralized delegation and isolated workspaces. This allows agents to work on interdependent subtasks in parallel with a 26.7% improvement in accuracy over single-agent baselines.
From the abstract
AI agents have become increasingly capable at isolated software engineering (SWE) tasks such as resolving issues on Github. Yet long-horizon tasks involving multiple interdependent subtasks still pose challenges both with respect to accuracy, and with respect to timely completion. A natural approach to solving these long-horizon tasks in a timely manner is asynchronous multi-agent collaboration, where multiple agents work on different parts of the task at the same time. But effective application