AI ディレクトリ : AI Agent, AI Code Assistant, AI Developer Tools
What is DeepSeek Super Agent?
DeepSeek Super Agent refers to dsh-superagent, a research prototype published by the DeepSeek Harness Team on GitHub under the title "Towards a More Autonomous Agent System". It is a proactive coding agent for one GitHub repository, packaged as a bundle for DeepSeek Harness. The README states plainly that it is for reference only and is not an official DeepSeek product.
How to Use DeepSeek Super Agent
Setup has many steps, so the repository asks you to let a coding agent do it with you: tell your agent to read SETUP.md and set up dsh-superagent. It asks one short question at a time, shows the full plan once, then installs and configures everything. A manual path is documented in the deployment guide.
Once running, it listens to GitHub and, optionally, Feishu. When an allowed person asks for work, it opens a work item, handles it in its own agent session and directory, and answers in the original conversation, for example with a pull request. A web panel lists what needs a person.
Key Features
- Proactive by rule It acts only when asked or when a rule the team wrote allows it, and speaks only when someone must act or can use a result.
- Isolated work items Each request gets its own agent session and working directory.
- Pull request replies Results come back in the thread where the work was requested.
- Review panel A web panel shows the items waiting on a human decision.
- Documented design Numbered docs cover philosophy, authorization, execution, prompts, deployment and architecture.
Why Choose DeepSeek Super Agent?
Most coding agents wait in a terminal for a developer. This prototype is built around a different principle: take as little human attention as possible and give as much value as possible, so a team can hand it requests where they already talk. It is MIT licensed and written mostly in TypeScript, so teams can study and adapt it. DeepSeek warns that much of the engineering was done by coding agents, so expect bugs and inconsistent style.
Use Cases and Applications
Engineering teams experimenting with autonomous agents can use it as a reference design for triaging GitHub requests, routing work from a Feishu chat and returning pull requests for review. Researchers can compare it with the paper "Exploring Super-agent Design"; the README notes the public code differs from DeepSeek's internal version, and some ideas in the paper are not implemented.
Frequently Asked Questions
Is DeepSeek Super Agent free?
Yes. The code is open source under the MIT license. You still pay for whatever model and hosting you run it on.
Is it an official DeepSeek product?
No. The README calls it a research prototype for reference only and advises reviewing and adapting it before use.
How popular is the repository?
At the time of writing the repository showed 16.8k stars and 1.8k forks.