DeepSeek open-sources router-worker coding-agent prototype
TL;DR
- DeepSeek's Harness team released dsh-superagent-prototype on October 31, 2026 under MIT license, framed as research code and not an official DeepSeek product.
- The design is a two-layer multi-agent split: one resident Router talks to people and dispatches work, isolated Workers run tasks in parallel.
- The stated goal is to minimize human attention each autonomous action costs, with much of the engineering reportedly performed by coding agents themselves.
DeepSeek's Harness team on October 31 open-sourced a research prototype called dsh-superagent, a proactive coding agent that listens to a GitHub repository and, optionally, Feishu. When an allowed user asks for work, the agent opens a work item, runs it in its own agent session and directory, and answers in the original conversation, for example with a pull request. The project sits on GitHub under an MIT license.
The shape is a two-layer multi-agent architecture. A resident Router talks to people and dispatches work; isolated Worker sessions run tasks concurrently with full tools and report back to the Router rather than to users. Writeups covering the release, including a Router–Worker breakdown on AgentsPulse and a DEV Community explainer, describe the design as one router dispatching many parallel workers rather than a single-loop coding tool. Two of the researchers we follow had the repo link in circulation within hours of the publish.
The team is unusually blunt about what this is not. The README calls it a "Research prototype for reference only, not an official DeepSeek product," and the authors say much of the engineering was performed by coding agents themselves, with code that "may still contain bugs, and its coding style and engineering practices may be inconsistent." The released code is also not identical to DeepSeek's internal version, and some designs in the write-up are not implemented.
The underlying pitch, repeated in the project's own framing, is to "take as little human attention as possible, and give as much value as possible."
Shared on Bluesky by 2 AI experts
-
DSH Superagent an agent built on top of DeepSeek Harness that’s meant to stay out of the way, do work only according to team rules (e.g. pickup new bugs) github.com/deepseek-ai/...
View on Bluesky →
Originally reported by github.com
Read the original article →Original headline: GitHub - deepseek-ai/dsh-superagent-prototype: Towards a More Autonomous Agent System