DeepSeek's Harness team open-sources dsh-superagent, a research prototype coding agent driven by GitHub
DeepSeek's Harness team has open-sourced dsh-superagent, a research prototype that receives development requests through GitHub and optionally Feishu. It works on each request in a separate session and returns results such as pull requests to the original conversation, according to TechNode.
DeepSeek's Harness team has open-sourced dsh-superagent, a research prototype that receives development requests through GitHub and optionally Feishu, TechNode reported on November 2, 2026.
What the prototype does
According to TechNode, dsh-superagent works on development requests in separate sessions. It then returns results, such as pull requests, to the original conversation in which the request was made.
Requests arrive through GitHub, with Feishu available as an optional channel.
Configuration and packaging
The prototype is configured around a single GitHub repository. It runs as a bundle for DeepSeek Harness.
TechNode describes dsh-superagent as a research prototype, not a finished product.
What this means
The release shows DeepSeek's Harness team working on an agent that takes requests where developers already work, in GitHub and optionally Feishu, and sends finished output such as pull requests back to the same conversation. Its design around a single GitHub repository and its packaging as a bundle for DeepSeek Harness define its current scope. The research-prototype label signals that it is an early-stage project. This is a tooling release, not a new model.
Related Articles
DeepSeek open-sources Super-agent prototype: Router-Worker design hits 15.7-minute P95 in simulation
DeepSeek has published an engineering write-up and a GitHub prototype for Super-agent, a two-layer agent architecture built around a Router and parallel Workers. In a simulated environment of about 300 messages and 20 work requests, DeepSeek reports a median task completion time of 12.4 minutes and a 95th-percentile time of 15.7 minutes.
DeepSeek V5 identifiers appear in official Harness repository on GitHub
Three DeepSeek V5 model identifiers have appeared in the company's official Harness repository on GitHub, according to TechNode. The identifiers are deepseek-v5, deepseek-v5-open and deepseek-v5-next, and they appear in tests for the software's DeepSeek account integration.
DeepSeek website code references Cyber, a cybersecurity model with identity-gated API access, in limited testing
Developers found references to a cybersecurity model called Cyber in the front-end code of DeepSeek's website, with two model IDs, cyber-personal-v3 and cyber-enterprise-v4, and a key system gated by real-name verification. DeepSeek has not announced the model, and launch, access and pricing remain unclear.
Tencent Hunyuan publishes Hy-Nomas report, releases 280-case world model benchmark under Apache 2.0
Tencent Hunyuan posted the Hy-Nomas technical report to arXiv on October 27 and released Hy-Nomas-Bench, a 280-case benchmark for interactive video world models. Hy-Nomas is an 8B diffusion world model, released with code and weights on October 21 under Apache 2.0.
Comments
Loading...