Introduction

DeepSeek Harness (DSH) adopts an “everything is a plugin” architecture. When developing agents or workflows, you may sometimes need to avoid network requests or rely on a purely local environment. dsh-copernicium is an inference plugin designed for such scenarios.

Plugin Overview

dsh-copernicium is a model inference plugin maintained by uckkk, also known as the “Copernicium element.” It is focused on running in local environments and does not generate network traffic.

Core Features

  • Pure Node implementation: The core logic is written entirely in Node.js and does not depend on other runtimes.
  • No network dependency: The plugin does not need to connect to external networks at runtime, making it suitable for offline or intranet environments.

Installation and Enabling

  1. Install it by running the following command in a terminal:
dsh plugin add github:uckkk/dsh-copernicium
  1. After installation, restart the DSH process to load the plugin.

Typical Usage

This plugin belongs to the model inference category and requires the following environment dependencies:
* @deepseek-ai/cordis: ^4.0.1
* @deepseek-ai/dsh-tools: >=0.1.0-rc.6

Make sure the above dependencies are installed in your DeepSeek Harness environment.

Applicable Scenarios and Notes

  • Use cases: Suitable for development scenarios that require model inference while operating in an offline environment, an intranet environment, or a setup with strict network-connection restrictions.
  • Security recommendation: This plugin runs with the permissions of the current DSH process. Before installation, it is recommended to review the source code and license (MIT).

Conclusion

dsh-copernicium provides DSH with network-independent inference capabilities. For more details, refer to the plugin’s directory or GitHub repository.