Introduction

During the development of DeepSeek Harness (DSH), adding visual understanding capabilities to agents is a common requirement. The plugin xby-general-recognition provides this capability. It can recognize labels for subject objects in an image and output category labels for the subject objects, currently covering more than 50,000 object categories.

What Is This

This is a model inference plugin maintained by xby-skill. Its core purpose is to directly invoke the image recognition service during a conversation, without manually relying on external tools, and obtain results by providing different image data formats.

Core Features

The plugin provides the following four main features:

  1. set_xby_apikey
    Set an API key in the chat. This setting is automatically persisted and remains valid after DSH is restarted.

  2. general_recognition
    Recognize labels for subject objects in an image. The input must be a link to an image file.

  3. general_recognition_for_data_base64
    Recognize labels for subject objects in an image. The input must be Base64-encoded image file data.

  4. general_recognition_for_data_file
    Recognize labels for subject objects in an image. The input must be a file path to an image file.

Installation and Configuration

Before installation, register and obtain an API key on the Xiaobenyang official website.

Install the plugin using the following command:

dsh plugin --profile web add github:xby_skill/xby-general-recognition

After installation, call the set_xby_apikey feature in the chat and enter the API key to start using it.

Notes

This plugin is released under the MIT License.

Before installation, it is recommended to check the source code and license to ensure they fit your use case. The plugin runs with the permissions of the current dsh process, so make sure to invoke it in a secure network environment.

To view the directory page or source code, visit:
- GitHub repository
- Plugin directory page