Image routing for text-only models in DeepSeek Harness: a global analyze_image tool (Kimi vision) plus automatic rewriting of pasted images into attachment references when the active model cannot see images.
시각 도구
$ npx -p @deepseek-ai/dsh dsh plugin --profile web add github:visail/dsh-vision-tool
dsh plugin: vision capability proxy. Routes image understanding for text-only main models to a small multimodal model, and backs off entirely when the active model declares multimodal support (inputModalities includes 'image').
시각 도구
$ npx -p @deepseek-ai/dsh dsh plugin --profile web add github:YJLTF/dsh-vision-tool
Wallhaven wallpapers for the DeepSeek Harness web GUI: search wallhaven.cc from a settings page, wear any result as the shell background (with opacity/blur/scrim controls), and download the original to disk. The host half fetches everything, so the browser never has to reach wallhaven — which matters behind a proxy or a poisoned DNS.
Agent 협업 · 시각 도구
$ npx -p @deepseek-ai/dsh dsh plugin --profile web add github:HaydenSmith1121/dsh-wallhaven-wallpaper
Wallpaper Engine integration for the dsh web GUI: scan the local Wallpaper Engine library (Steam workshop 431960 + local projects), use its wallpapers as the page background (image / video / web / scene-preview), and control them from a right-side '澹佺焊璁捐' panel (opacity, scope, fill, blur, vignette, fps, parallax, carousel, theme linkage). Agent tools wallpaper_scan / wallpaper_list / wallpaper_set / wallpaper_config.
Agent 협업 · 시각 도구 · 인터페이스 확장 · 서비스 연동
$ npx -p @deepseek-ai/dsh dsh plugin --profile web add github:codeMonkey-Pine/dsh-wallpaper
Wallpaper Engine bridge for the DeepSeek Harness web GUI: browse your local Wallpaper Engine library and use any dynamic wallpaper as the GUI background — video, web, image and scene previews, with scrim and translucency controls. Hot-pluggable via cordis.patch.yml + profile, no dsh source changes.
시각 도구 · 인터페이스 확장 · 서비스 연동
$ npx -p @deepseek-ai/dsh dsh plugin --profile web add github:pbadgpmeb22791-sketch/dsh-we-wallpaper
Zhipu (智谱) GLM-4V vision understanding for DSH: agent tool zhipu_vision (local image path or URL + prompt) calls the Zhipu vision API and returns the model answer; paste-to-identify UX — pasting an image into the chat composer saves it on the host and inserts the recognition instruction into the composer through the official conversation input face, so you type your own question after it and send once; the text-only chat model (DeepSeek) answers via the tool. A sidebar 视觉 panel with image+question combo send (sessions driver, same as the task board), a saved-image gallery, and a plugin config card in the GUI settings. Hot-pluggable — mounted via the web profile cordis.patch.yml + a profile node_modules copy, no dsh source changes.
Agent 협업 · 시각 도구 · 인터페이스 확장
$ npx -p @deepseek-ai/dsh dsh plugin --profile web add github:xingling80/dsh-zhipu-vision
A DeepSeek Harness tool plugin that lets a text-only agent 'see' local images: it reads an image file, auto-detects the real format, and returns (or writes to a Markdown file) a detailed text description produced by a separate OpenAI-compatible vision model.
Self-contained global DSH plugin registering the parse_docs model tool: read PDF/DOCX/PPTX/XLSX/TIFF/batch images via local MinerU as Markdown. Bundles its own parse.ps1; no skill package dependency