DeepSeek Open-Sources Super-Agent, a Coding Agent That Reads Team Chats

DeepSeek just dropped a coding agent on GitHub, and the code is free to take. The Chinese AI lab released Super-Agent - repository name dsh-superagent-prototype - under an MIT licence on Saturday, alongside a WeChat post from the team that builds its agent tools. The pitch: an agent that watches your team's chat for coding problems and gets to work without being summoned.

One precision that matters: DeepSeek itself calls this a research prototype and "not an official DeepSeek product". The public code differs from the internal version, and some ideas in the team's write-up are not built yet.

Built to save human attention

The agent works on one GitHub repository and can also listen to Feishu, ByteDance's workplace chat app. When an approved person asks it for work, it opens a task, handles it in its own session, and replies in the original thread - for example with a pull request. A web panel lists what still needs a human. The design rule is one line: "Take as little human attention as possible, and give as much value as possible." DeepSeek is blunt about the flip side: an agent acting for a whole team can waste the whole team's attention, so it acts only when asked or when a team's own rules allow it.

The race to run many agents

The target is Anthropic's Claude Code, Bloomberg reports. DeepSeek set up a dedicated agent-tools team this year and released a preview of its own agent app in August. US companies are chasing the same idea: at GitHub Universe on Wednesday, GitHub showed a Copilot app running agent work across several repositories in parallel sessions. Coordination remains the hard part - in the CooperBench study published in January, coding agents working as a team succeeded about 30% less often than agents working alone.

The code is on GitHub now, and DeepSeek tells users to review and adapt it before running it. Sound advice for any agent you invite into your team's chat.