Preface

DeepSeek Harness (DSH) adopts a plugin-based architecture. When developing agents or LLM pipelines, processing raw web page HTML directly often leads to token waste and difficulties in structural parsing. xby-read-website is a networked tool plugin designed to convert web page content into clean Markdown format, preserving links and structure, suitable for reading documents, analyzing content, and information gathering.

Plugin Introduction

Name: xby-read-website

Purpose: DSH plugin that converts web pages into clean Markdown format.

Maintainer: xby-skill

License: MIT

Core Features

The plugin provides two main commands:

  1. set_xby_apikey
    Set the API key in the chat. Once set, it is automatically persisted and remains valid after the DSH process is restarted.
  2. read_website
    Provides fast, token-efficient web page content extraction. The output is in clean Markdown format while preserving links and structure.

Installation

Use the following command to add the plugin to DSH’s web configuration file:

dsh plugin --profile web add github:xby_skill/xby-read-website

Usage

To use this plugin, first obtain an API key, then invoke the corresponding command in the conversation.

  1. Obtain the key
    Visit the Xiao Ben Yang official site to register and obtain an API key.
  2. Set the key
    Use the set_xby_apikey command in the chat to set the key.
  3. Extract content
    Use the read_website command to extract the target web page content. The system will automatically convert it into Markdown format.

Notes

  • Runtime Permissions: The plugin runs with the permissions of the current DSH process. It is recommended to review the source code and license before installation.
  • Project Structure: The project type is module, and the main entry file is ./dist/index.js.

Conclusion

xby-read-website provides a standardized path for efficiently processing web page content within the DSH ecosystem. To view the directory or source code, you can access:
* Plugin Directory
* GitHub Repository