@deepseek-ai/dsh-qwen-mm
Qwen-MM-Plugins integration bundle for DeepSeek Harness: bundled multimodal skills, MCP server rows, and an image attachment bridge for text-only model routes
安装
npx -p @deepseek-ai/dsh dsh plugin --profile web add github:RRRosmontis/dsh-qwen-mm说明文档
阅读完整 README ↗dsh-qwen-mm
把 Qwen-MM-Plugins 集成进 DeepSeek Harness,让纯文本的 DeepSeek 模型也能"看懂"图片、视频、文档,并调用网络搜索、3D/CAD 等工具。
自主实现说明:本插件的架构调研、编码、测试、文档与真机验证,均由 DeepSeek Harness(
dsh)驱动deepseek-v4-flash模型自主完成,全程无需人工编写脚手架。
English
Make DeepSeek Harness multimodal-native by integrating Qwen-MM-Plugins.
Autonomous implementation. This plugin was designed and implemented end-to-end, without hand-written scaffolding, by DeepSeek Harness (
dsh) running thedeepseek-v4-flashmodel — architecture research, coding, tests, documentation, and live verification were all performed by the agent itself.
功能总览 / Capabilities
| 能力 | 实现方式 | 需要密钥 |
|---|---|---|
| 读图 / 读视频 / 读文档 / 读代码 / 读数据 | core MCP 工具(read_image、visualize、media_info、read_video …) | ❌ 无 |
| 云端视觉 / OCR / ASR / 物体定位 | api MCP 工具(vision_chat、ocr、omni_* …),基于 DashScope Qwen VL/Omni | ✅ DashScope |
| 网络搜索 + 以图搜图 | search MCP 工具(web_search、web_extractor、image_search) | ✅ Serper / Tavily / Exa |
| 长视频问答 / 视频剪辑 / Blender / FreeCAD / 数学视频 | video-memory、video-edit、blender、freecad、edu-agent | 按能力而定 |
| 拖图上传 → 纯文本模型读懂 | 下方"图片附件桥接" | ❌ 无(走 vision_chat) |