What is DeepSeek's Super-Agent?

DeepSeek has released the code for Super-Agent, an experimental AI agent that takes on coding work for a software team. The Chinese AI lab posted the project on GitHub under the name dsh-superagent, with an MIT license. The team behind its agent tools laid out the design in a WeChat post on Saturday.

DeepSeek is aiming the agent squarely at Anthropic's Claude Code, Debby Wu reported for Bloomberg. According to her account, the agent can watch developers' group chats for coding problems and suggest solutions without being summoned. DeepSeek itself describes the release as a research prototype and says it is “not an official DeepSeek product.”

A tool built to save human attention

The agent works on a single GitHub repository and can also listen to Feishu, the workplace chat app owned by ByteDance. When an approved person asks it for work, the agent opens a task and handles it in its own session. It then replies in the original thread, for example with a pull request. A web panel lists the items that still need a human.

The goal is to spare people the job of coordinating several AI agents and keeping track of what they are doing. Ben Jiang reported that goal for the South China Morning Post, citing the team's article. DeepSeek's manual states the principle in one line.

“Take as little human attention as possible, and give as much value as possible.”

Can Super-Agent act without being asked?

The company also warns that an agent acting for an entire team can waste that team's attention as easily as it can save it. For that reason, the agent acts only when asked, or when a team's own rules allow it, according to the project's philosophy document.

The public code is not identical to DeepSeek's internal version. The README says some ideas described in the article have not been built yet.

A crowded race to run many agents

DeepSeek set up the team responsible for this work earlier this year to compete with coding tools from US labs, Bloomberg reported. In August, the company released a preview of its own agent app, a rival to Claude Code. The open-source app, now in public preview, handles coding, documents and spreadsheets.

US companies are chasing the same idea. On Wednesday at GitHub Universe, GitHub showed a Copilot app that runs agent work across several repositories in parallel sessions.

Getting agents to share work, however, remains difficult. In January, researchers behind the CooperBench study found that coding agents performed worse when working as a team. Their success rates averaged 30% lower than when each agent completed the tasks on its own. The study did not test DeepSeek's agent.

The release also arrives as DeepSeek contends with limits in the United States. Civilian federal agencies there were told to remove DeepSeek last month. The company has raised about $15 billion ahead of a planned listing, Bloomberg reported.

For now, the code on GitHub is a prototype. DeepSeek tells users to review and adapt it before running it.

This article was written with the assistance of AI.
News Factory APP - agentic news to boost your SEO & AEO.