[DeepSeek Releases New Agent Architecture] Odaily Planet Daily reported on the evening of October 31st that the DeepSeek Harness team published a technical article titled "Exploring Super-agent Design: Towards a More Autonomous Agent System." DeepSeek stated that improved model capabilities have spurred new application forms, with personal agents recently emerging in large numbers. However, current mainstream agent designs struggle to support this model at the scale of multi-tasking and team collaboration; the agent's working mode still operates on a single session as the basic unit, with different sessions independent of each other. Under this design, the upper limit of a single task depends on the model's capabilities, while the upper limit of multi-tasking depends on human attention. Through a two-layer architecture, event-driven architecture, asynchronous multi-channel interaction, storage-computation separation, and external watchdogs, the DeepSeek team has overcome these limitations. The system takes over the cumbersome session management and context synchronization at the lower level, allowing humans to move from specific execution details to higher-level decision-making and acceptance. DeepSeek refers to this agent model and its system architecture as Super-agent. The article mentions that internal Super-agent instances built based on this approach are already running continuously, participating in daily issue reproduction and fixing, code review responses, and iterations of DeepSeek Harness itself. (DeepSeek Harness Team)