Agensh
A decentralized multi-agent runtime framework open-sourced by Microsoft Research. It has no central orchestrator; agents self-organize and collaborate through a shared Git repository, chat channels and a discovery board. It supports horizontal scaling to thousands of agents.
Timeline (last 90 days)
Microsoft Research open-sourced Agensh, a decentralized multi-agent framework.
Agensh has no centralized orchestrator; agents rely on a shared Git repository, chat channels, and a discovery board for self-organized collaboration.
On ProgramBench, Agensh achieved scaling from 1 to 1,024 agents.
Agensh's run on ProgramBench was constrained to a 6-hour time budget.
The original report did not disclose Agensh's specific task pass rate on ProgramBench, its performance curves at different scales, or comparison data against centralized orchestration approaches.
Agensh pushes coordination responsibilities down to shared artifacts, and its effectiveness depends heavily on individual agent capabilities and the quality of the collaboration protocol design; in scenarios where task decomposition is difficult, self-organization may manifest as large amounts of inefficient exploration.
The original report did not elaborate on how Agensh handles merge conflicts when multiple agents simultaneously commit changes to the same Git repository, or how discovery board information avoids becoming outdated.
Agensh's approach of moving the coordination mechanism from central scheduling down to shared artifacts (such as a Git repository) can serve as a scalability design reference for multi-agent systems.
The information currently public about Agensh is limited, lacking benchmark details and direct comparisons with other frameworks, and it reads more like a directional statement.
All Facts (9)
The information currently public about Agensh is limited, lacking benchmark details and direct comparisons with other frameworks, and it reads more like a directional statement.
50%UnverifiedAgensh has no centralized orchestrator; agents rely on a shared Git repository, chat channels, and a discovery board for self-organized collaboration.
50%UnverifiedOn ProgramBench, Agensh achieved scaling from 1 to 1,024 agents.
50%UnverifiedAgensh's run on ProgramBench was constrained to a 6-hour time budget.
50%UnverifiedThe original report did not disclose Agensh's specific task pass rate on ProgramBench, its performance curves at different scales, or comparison data against centralized orchestration approaches.
50%UnverifiedAgensh pushes coordination responsibilities down to shared artifacts, and its effectiveness depends heavily on individual agent capabilities and the quality of the collaboration protocol design; in scenarios where task decomposition is difficult, self-organization may manifest as large amounts of inefficient exploration.
50%UnverifiedThe original report did not elaborate on how Agensh handles merge conflicts when multiple agents simultaneously commit changes to the same Git repository, or how discovery board information avoids becoming outdated.
50%UnverifiedAgensh's approach of moving the coordination mechanism from central scheduling down to shared artifacts (such as a Git repository) can serve as a scalability design reference for multi-agent systems.
50%UnverifiedMicrosoft Research open-sourced Agensh, a decentralized multi-agent framework.
50%