Add vision support to DeepSeek V4 in OpenCode

Hi, folks. Recently I just have built an OpenCode plugin to delegate visual tasks to the models in your existing AI subscriptions: Recognizing contents in an image with main model set to DeepSeek V4 Pro It just uses the vision-capable models in your existing AI subscriptions. Picking a vision-capable model in your existing AI subscriptions With this plugin, visual tasks with DeepSeek V4 set as the main model on OpenCode are unlocked. Installation: opencode plugin opencode-vision -g This plugin handles: imag