DeepSeek Coding Agent Super-agent Takes Aim at Claude Code

DeepSeek has released Super-agent, an open-source coding agent prototype that watches a team’s GitHub repository and chats and steps in with fixes without being prompted for every task. The DeepSeek Harness team announced the DeepSeek coding agent in a WeChat post on Saturday, October 31, and published the code under the MIT license, a move Bloomberg framed as a challenge to Anthropic’s Claude Code.

The code sits in the dsh-superagent repository on GitHub, which passed 4,000 stars within a day of going public. DeepSeek labels it a research prototype for reference only, “not an official DeepSeek product,” and warns it is not production-ready and differs from the version the team uses internally.

What the DeepSeek coding agent does

Super-agent is built as a bundle for DeepSeek Harness, the company’s agent scaffolding. It listens to one GitHub repository and, optionally, Feishu (Lark) group chats. When an approved team member asks it for work, it opens a work item, handles it in its own session and working directory, and replies in the original conversation, for example with a pull request. A local web panel lists only the items that need a human decision.

The design splits the agent into two layers. A persistent “Router” talks to people and keeps the shared context, while separate “Worker” sessions run tasks in parallel, so the agent can take a new request before the last one finishes. A Worker that owns a pull request keeps following its comments directly, and a Watchdog process watches for failures and stalled loops.

The project’s stated rule is to “take as little human attention as possible, and give as much value as possible.” In practice, proactive work nobody asked for needs a rule the team wrote, and it starts in a mode that only records what the agent would have done.

Requirements and risks

Setup needs a dedicated GitHub user account for the agent, Node.js 22.19 or later, GitHub CLI 2.48 or later and a model provider configured in DeepSeek Harness. Only github.com is supported, not GitHub Enterprise Server. DeepSeek also notes that the agent’s sessions run with full access as the operator’s OS user and recommends a separate OS account for isolation.

That warning matters after recent supply-chain scares around AI coding tools, such as the tampered Claude Code extension that planted an SSH backdoor on developers’ machines.

Why it matters

Chinese and US labs are now competing on agents, not just chatbots. Anthropic has been tightening how Claude Code can be used, including its new rule against recreating existing games, while US labs have accused Chinese rivals of copying their models, as in the GPT-6 Astra distillation report naming DeepSeek.

DeepSeek created its Harness team earlier this year to build agents that can compete with US coding services, and it has raised about 100 billion yuan ($15 billion) ahead of a planned IPO, Bloomberg reported. For now, Super-agent is a blueprint developers can study and test, not a finished product.