DeepSeek Harness Multimodal Plugin: Integrating Qwen Capabilities

`wuwangmao/dsh-qwen-multimodal` is a JavaScript plugin designed for DeepSeek Harness (DSH), intended to provide multimodal support for text-only models. By calling Alibaba Cloud Bailian's Qwen API, the plugin introduces three core capabilities to DSH: image understanding, speech transcription, and image generation. The plugin requires no build step and depends on the `subprocess` or `tools` service at runtime. Its core features include: using `qwen3-vl-flash` for image description and OCR; using `qwen3-asr-flash` for speech-to-text; and using `qwen-image-plus` for image generation. It supports specifying a language and a custom P