SuperAgent
An open-source agent research prototype from the DeepSeek Harness team that adopts a Router-Worker architecture separating scheduling from execution. It decouples message triggering, task acceptance decisions, work items, and execution, and is oriented toward long-running workflow management.
Timeline (last 90 days)
The DeepSeek Harness team has published a SuperAgent design article and GitHub repository, released under the MIT license and explicitly labeled as a research prototype.
SuperAgent adopts a Router-Worker architecture that separates message triggering, task-acceptance decisions, long-term work items and actual execution: the Router decides whether to accept a task and whom to assign it to, and the Worker completes the task in an independent session and directory and writes back the result.
The official deployment documentation states that Workers execute with operating-system user permissions, recommends isolating them in a dedicated account, and warns that the prototype must not be connected directly to a company's production repository.
In simulated tests, SuperAgent's median task completion time was 12.4 minutes and its 95th percentile was 15.7 minutes; under a neutral prompt, the 95th percentile for a single session plus a SubAgent was 29.5 minutes.
In the above benchmark tests, coding work used a stand-in tool that only consumes time, so the tests measure scheduling, concurrency and interruption management, not the actual speed, quality or success rate of the model in completing code.
The control Agent under the neutral prompt never proactively delegated subtasks in any of the 5 tests.
All Facts (6)
The DeepSeek Harness team has published a SuperAgent design article and GitHub repository, released under the MIT license and explicitly labeled as a research prototype.
60%UnverifiedSuperAgent adopts a Router-Worker architecture that separates message triggering, task-acceptance decisions, long-term work items and actual execution: the Router decides whether to accept a task and whom to assign it to, and the Worker completes the task in an independent session and directory and writes back the result.
50%UnverifiedThe official deployment documentation states that Workers execute with operating-system user permissions, recommends isolating them in a dedicated account, and warns that the prototype must not be connected directly to a company's production repository.
50%UnverifiedIn simulated tests, SuperAgent's median task completion time was 12.4 minutes and its 95th percentile was 15.7 minutes; under a neutral prompt, the 95th percentile for a single session plus a SubAgent was 29.5 minutes.
50%UnverifiedIn the above benchmark tests, coding work used a stand-in tool that only consumes time, so the tests measure scheduling, concurrency and interruption management, not the actual speed, quality or success rate of the model in completing code.
50%UnverifiedThe control Agent under the neutral prompt never proactively delegated subtasks in any of the 5 tests.
50%