dsh-vision-bridge:在 DeepSeek Harness 聊天框粘贴图片,任意模型都能看懂——纯文本模型自动转成图片路径并调用 vision_chat 识图,支持图片的模型直接接收原图。粘贴即用,自动识别、自动回答。dsh-vision-bridge: paste any image into the DeepSeek Harness chat box — every model can see it: text-only models get an automatic vision_chat bridge, vision-capable models receive the image directly. Paste, send, done.
Per-session Node REPL terminal in the DSH web UI: a session-local floating terminal (ghostty-web) bridged over WebSocket to an in-process node:repl that has the live agent Cordis Context (agent.ctx) in scope.
DSH plugin: for text-only model routes, save uploaded images under .dsh_images and analyze them via modlens (external vision provider), replacing the image block with text evidence.