시각 도구 · DeepSeek Harness 플러그인 — 36 페이지

DeepSeek Harness vision bridge: session screenshots route by model capability. Image-capable models receive attachments inline; text-only models see the standard placeholder, and the bridge reads the image through a vision-language model — on demand via the vision_bridge_read tool or automatically via agent/pre-step context injection. 会话截图按模型能力自动分流：多模态模型原样直读；纯文本模型经 VLM 代读（tool/auto 双模式）。

DSH profile bundle: auto-bridge images for text-only models (DeepSeek etc.) — instead of the native "model does not support images" rejection, the image is materialized into the workspace and the model is instructed to run the configured vision tool (vision_glance, fuzzy-selectable) and answer from its analysis. Works on WSL/Ubuntu and native Windows; sessions on excluded presets (default minimal) keep the original blocking behavior.

Lightweight DSH upload bridge: images and common documents become workspace paths so a text-only DeepSeek model can read them with Qwen-MM-Plugins vision tools.

辅助视觉模型：图片→文本描述，供文本模型、浏览器截图兜底与聊天贴图降级使用

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

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

External vision model plugin for image recognition when the active model lacks vision

DeepSeek Harness plugin that uses configured model providers for image analysis and context compaction.

DeepSeek Harness plugin that exposes MuApi's 100+ image, video, and audio generation models as an agent tool.

DSH 视觉增强方案：识图技能（图片理解/OCR/文档解析）+ 幂等宿主补丁，让纯文本模型也能收发图片。

DeepSeek Harness plugin: direct multimodal image transfer declarations + per-session image clearing (surface replace), with a settings page and composer controls.

Local image understanding for the DeepSeek Harness: OCR, table-layout detection, and semantic description of textless images via macOS Vision, fully on-device (images never leave your Mac). 本地化识图，图片不出本机。

DSH Web 插件：调用智谱 GLM 视觉模型分析图片（analyze_image 工具 + 插件设置卡片）。

格物审视面插件：gewu_prep（HTML 逐页截图 + subagent_vision 简报契约生成，宿主直起 Chrome）+ gewu_locate（真值核验：文字定位到页码/行号）。让无视觉能力的主脑经 subagent_vision 完成视觉检验。

Zero-config local vision and realtime audio routing for DeepSeek Harness

Zero-config multi-provider vision routing for DeepSeek Harness

DSH plugin exposing adb-based tools that let the model operate an Android phone (screenshot, UI dump, tap, swipe, input text, press keys, open apps). | 让模型通过 adb 操作安卓手机的 DSH 插件（截图、UI 树、点击、滑动、输入文本、按键、打开应用）。

Replaces the dsh Web client base background with a randomly switching public Unsplash image, revealed through translucent base surfaces.

dsh auto-vision bridge: paste an image into a text-only model's composer and a configured multimodal model transcribes it as text, automatically.

DeepSeek Harness plugin: generate LaTeX math typeset as font-free SVG/PNG images, and draw geometry/function figures from a declarative SVG spec, ready to embed in documents

Preview images / SVG (and prepare CSV) in the DeepSeek Harness chat, rendered as real browser <img> elements. Preview-first workflow: show the user the asset, wait for approval, then apply the real change.

DeepSeek Harness settings-pro plugin: IM bridge (Telegram/WeChat), DeepSeek usage, cross-restart memory, desktop pets, and a vision bridge

Shared client-side content-preview kernel for community file surfaces: the kind-union content pane (markdown / JSON / CSV / image / code, plus the Tier0/Tier1 sandboxed HTML render with its capability bridge), rendered-content search, and the per-file host-open chrome. Consumed at the SOURCE plane — each surface's tsdown client bundle inlines it — never mounted as a plugin row.

Git-agnostic local file browser in the right sidebar: a lazy file tree with structured HTML/Markdown/JSON/CSV/image previews over its own Typert Remote, defaulting to the session's workspace but not limited to it