視覚ツール · DeepSeek Harness プラグイン — 37 ページ

bbs.dsher.cn 社区面板 — DSH Web 的正式 bundle 插件（npm 安装形态）：右下角悬浮入口 + 浏览/搜索/发帖/回帖/点赞/图片上传 + 一键登录

OpenRouter image, video, and speech generation as dsh tools, shipped as an out-of-tree profile bundle

Per-model vision (image input) toggle for DeepSeek Harness (dsh): list every configured model and flip a switch to enable/disable image support without hand-editing settings.yaml. 为 DeepSeek Harness 提供按模型的「支持图片」开关：无需手改 settings.yaml。

持久化「添加图片」按钮：在对话输入栏工具行左侧补充系统文件选择入口（accept=image/*，多选），选中的图片走官方 draft 附件管线（conversation.input.left 加法式座位）。

Persistent rembg GPU background-removal tool (remove_background) for DeepSeek Harness, in the dsh-vision-toolkit bundle form.

DeepSeek Harness plugin: content-routing LLM adapter. Text-only requests go to DeepSeek; requests containing images go to a user-configurable OpenAI-compatible vision provider. Read_image works natively for text-only models.

DSH plugin: display images inline in chat with lightbox zoom, download (save-as), and prev/next navigation

Automatic reasoning and image capability detection for custom DeepSeek Harness models

DeepSeek Harness-native integration for agent-vision-toolkit: image Q&A, OCR, grounding, UI restoration, pixel diff, Artifacts, and Web UI.

Give DeepSeek Harness eyes and hands: multi-provider media generation, vision routing, and an integrated Iris workbench.

在设置页编辑第三方模型的思考档位与图片输入声明,解决手写配置易错与档位拼写不可控

The see tool for the DeepSeek Harness: offline OCR with positions (macOS Vision) + ASCII layout art (PIL) + vision-model semantic description (mimo-v2.5 via the opencode-go gateway). Works with any model, including text-only ones.

Dynamically change DSH Web CSS from Settings: raw CSS, background image, and a user JS file.

DSH Web GUI 全屏 Markdown 阅读器插件：GFM 渲染、大纲目录与滚动定位、KaTeX 数学公式、Mermaid 图表管线（离线优雅降级）、相对图片与文档内跳转。工作区门禁的只读路由；零构建、零运行时依赖（宿主端纯 Node ESM，浏览器端单文件 bundle 复刻 __ModuleLoader__ 契约）。

First-party DeepSeek Harness plugins for the sai Android coding agent

Persistent dual-face Design Studio plugin for the DSH harness: the design_studio agent tool (design systems under temp_design_folder/, identity presets, OpenRouter vision review, and a real harness-subagent design agent that lists, reads and edits the design files itself), the /design-studio live-preview route + JSON API, a design-brief prompt section, and the built-in Design Studio web UI (conversation tab, agent chat, settings pages). Install once — the tab appears on every launch. Zero runtime dependencies (plain JS).

Dynamic multimodal-to-text projection and cross-model vision bridge for DeepSeek Harness (dsh), allowing seamless hot-switching between vision models (Gemini/Claude) and text-only models (DeepSeek).

DSH web plugin: a Vue 3 + Vite complete file workbench (explorer tree / code editor / markdown+image+html preview / file search / CRUD)

Windows key/mouse/screenshot/window automation (SendInput over a persistent helper) for the DeepSeek Harness

TokenLab provider bundle for DeepSeek Harness with native Responses, Messages and Chat routing plus full multimodal and async tools.

One-click MD3 image attachment for the DeepSeek Harness composer, with automatic per-session vision-model routing: attach an image and the session switches to a vision model, then restores your previous model.

Vision recognition plugin for DeepSeek Harness: paste images into the composer, recognize them via GLM-4V on the host side, and inject the result into the conversation.

Native macOS Computer Use (screenshot + mouse/keyboard control) tool for DeepSeek Harness

dsh 插件 · 放风筝引擎：agent 编程时屏幕上放一只动画风筝——token 越多事件越密风筝飞得越高，随风到处漂移摆动，一根线牵在屏幕底边；潍坊系框架卡组（沙燕/金鱼/蝴蝶/八卦/龙头……硬翅软翅板式立体），形状×图案×配色全是可替换的数据配置，还支持把用户上传的图片糊上风筝面，贴图随风筝侧正斜的姿态实时仿射变换。