Introduction

The core design philosophy of DeepSeek Harness (DSH) is “everything is a plugin”. Developers can extend the capabilities of Harness by loading different plugins. In model inference scenarios, it is sometimes necessary to avoid network requests or rely on a purely local environment. The dsh-cobalt plugin was created specifically to meet this need.

Plugin Introduction

dsh-cobalt, with the project codename “Cobalt Element”, is maintained by developer uckkk and is open-sourced under the MIT license.

Core Features

  • Pure Node implementation: The code logic is fully based on the Node.js runtime.
  • No network dependency: The plugin runtime does not depend on external network services.

Installation and Dependencies

To enable this plugin, please execute the following command in the Harness environment:

dsh plugin add github:uckkk/dsh-cobalt

The plugin depends on the following environment libraries:
* @deepseek-ai/cordis (^4.0.1)
* @deepseek-ai/dsh-tools (>=0.1.0-rc.6)

Applicable Scenarios and Notes

This plugin is suitable for development scenarios that require model inference in offline or strictly isolated network environments.

  • Permissions and Security: The plugin runs with the permissions of the current Harness process. Please ensure it is installed from a trusted source, and read the source code and license (MIT) before use.

Conclusion

dsh-cobalt adds a lightweight, network-independent model inference plugin component to the DSH ecosystem.