Introduction

The design philosophy of DeepSeek Harness (DSH) is “everything is a plugin.” When developing agents, integrating knowledge or tools for a specific domain often requires writing additional scripts. The dsh-body-type plugin directly provides a ready-made solution for traditional Chinese medicine (TCM) body-type identification, eliminating the need to build from scratch and lowering integration barriers.

Plugin Overview

dsh-body-type is a DSH plugin for TCM body-type identification, maintained by developer uckkk. It encapsulates the analysis of TCM health-preserving plans into callable tools. At runtime, it entirely relies on a local Node.js environment.

Core Features

  • TCM Body-Type Identification: Provides the body_type tool, supports parameter-based invocation, and returns the corresponding TCM health-preserving plan.
  • Local Execution: Implemented purely in Node.js, with no network dependencies and fast response times.

Installation and Enablement

Run the following command to install the plugin:

dsh plugin add github:uckkk/dsh-body-type

After installation, restart the DSH process to invoke the registered tools in a session.

Typical Usage

After installation, you can directly invoke the tool registered by the plugin in a session. The plugin returns the appropriate health-preserving plan based on the input parameters.

Applicable Scenarios and Notes

  • Dependency Requirements: The plugin depends on @deepseek-ai/cordis (^4.0.1) and @deepseek-ai/dsh-tools (>=0.1.0-rc.6). Please ensure your environment versions meet the requirements.
  • Permissions and Security: The plugin runs with the current DSH process permissions. It is recommended to install only plugins from trusted sources.
  • License: The plugin is open source under the MIT License.

Conclusion

dsh-body-type integrates TCM body-type identification capabilities into the DSH ecosystem and is suitable for application scenarios that require rapid integration of domain-specific knowledge. For more details, check the plugin directory or the source code on GitHub.