DeepSeek previews open-source 'super agent' designed to run coding work with minimal human supervision

Chinese artificial intelligence start-up DeepSeek has previewed a “super agent” system designed to handle multiple tasks with minimal human intervention, as it competes with US rivals OpenAI and Anthropic in AI coding assistants, the South China Morning Post reported on Sunday.

According to the newspaper, the system is described in a technical article published on Saturday by the company’s DeepSeek Harness team. The article said the design aims to reduce the need for people to coordinate multiple AI agents and monitor their progress, which it described as a key bottleneck in managing increasingly complex workflows.

DeepSeek has also published a version of the code on GitHub under the MIT licence, called “dsh-superagent”. The repository describes it as “research prototype for reference only, not an official DeepSeek product”, and says it is not identical to the internal version or to everything described in the article. According to the documentation, the agent works on a single GitHub repository, listens to GitHub and, optionally, the Feishu workplace app, and when an authorised person asks it for work, opens a work item, carries it out in its own session and replies in the original conversation, for example with a pull request. It acts only when asked or when a rule written by the team allows it, the documentation says, summarising the approach as “take as little human attention as possible, and give as much value as possible”.

DeepSeek said advances in AI models had led to agents that can automate a growing range of tasks, from managing calendars and drafting emails to preparing work reports and writing code, the Post reported.