woyeshishen/dsh-vision-plugin
Project Overview项目介绍
dsh-vision-plugin adds image understanding to DeepSeek Harness by wiring up an OpenAI-compatible external vision model. A text-only main model can call the describe_image tool to hand off an image and receive a plain-text description for screenshots, photos, charts, OCR, and UIs. Images flow only to the secondary model; the main model always works with text. The plugin auto-loads at DSH startup and ships with a GUI settings page for URL, API key, and model selection, with credentials stored encrypted. Use it when conversations need to reason over image content. Caveat: an external /chat/completions endpoint with image support is required.
为 DeepSeek Harness 引入图像理解能力:接入 OpenAI 兼容的外部视觉模型后,纯文本主模型(如 deepseek)即可调用 describe_image 工具,将截图、照片、图表等交给外部模型并获得纯文本描述。架构上图像仅送往辅助模型,主模型始终处理文本。安装后插件自动加载,提供 GUI 配置页填写 URL、API Key 和模型名,密钥加密保存不再回显。适用于需要识别 UI、OCR 或图表内容的对话场景。注意:需自备支持图像的 OpenAI 兼容 /chat/completions 端点。
请帮我安装这个 DSH 插件。安装前先完成【兼容性检查 + 安全性检查】,检查通过再动手。
插件:dsh-vision-plugin(woyeshishen/dsh-vision-plugin)
仓库:https://github.com/woyeshishen/dsh-vision-plugin
本站详情页:https://www.yhbd.top/plugins/woyeshishen-dsh-vision-plugin/
本站登记:类型 plugin · 归类 原生 DSH 插件 · 许可证 Apache-2.0 · ⭐ 2 · 最近提交 2026-08-15 · 主语言 TypeScript
按下面顺序执行,每步先把结论告诉我,再进入下一步:
【1 兼容性检查】
① 我这边:DSH 版本、Node 版本、操作系统、当前 profile(web / desktop)。
② 读它的 README、package.json、插件 manifest,列出它要求的 DSH 版本 / Node 版本 / 操作系统 / 外部依赖 / 需要另外先装的运行时。
③ 逐条比对,结论只写「满足 / 不满足 / 未知」三种;不满足的给出可行替代方案。
④ 检查是否和我已装的插件冲突:命令名重复、skill / tool 重名、端口占用、重复注册的 MCP server。
【2 安全性检查】
① 仓库可信度:和上面「本站登记」是否一致;star / fork 数、创建时间、最近提交,是否归档或长期停更。
② 安装脚本:逐行看 package.json 的 preinstall / install / postinstall,以及 install.sh、setup.ps1 之类脚本。出现 curl|bash、下载后直接执行、混淆代码、访问与插件功能无关的域名,立刻停下来告诉我,不要继续装。
③ 依赖:列出新增依赖,标出无人维护、或与知名包拼写近似的可疑包(typosquatting)。
④ 权限与副作用:它会读写哪些目录、访问哪些域名、需要哪些 DSH 权限(filesystem / network / shell / clipboard 等),以及怎么卸载和回滚。
⑤ 如果它要求 sudo / 管理员权限,或权限明显超出功能所需,先停下来问我。
【3 安装】
上面两步没有「不满足」和「高危项」时才执行;用官方推荐方式安装,不要自行提权。
【4 汇报】
用表格输出:检查项 / 结论 / 依据 / 是否需要我决策。拿不准的一律写「未知」并说明要我怎么确认——不要猜,也不要替我决定。
Send this message to DSH in your current session: it verifies compatibility and security first (answering met / not met / unknown item by item) and only installs once everything checks out — it will stop and ask you if it finds a high-risk item. The box scrolls; the copy is the full prompt. CLI install commands may not be accurate across systems, so DSH is the safer route.把上面这条消息直接发给当前会话里的 DSH:它会先核对兼容性与安全性(逐条给「满足 / 不满足 / 未知」),确认没问题再安装,有高危项会停下来问你。框内可滚动,复制到的是完整提示词;安装命令不一定准确,发给 DSH 更稳。
- Only 2 stars - very few users, little community feedback星标只有 2,几乎没人在用,遇到问题缺少社区反馈
DSH walks through these 9 checksDSH 会逐条核对这 9 项
Compatibility兼容性
- DSH, Node, OS and profile requirementsDSH 版本 / Node 版本 / 操作系统 / profile 是否满足要求
- External dependencies and runtimes (Electron / Python / Docker, ...)外部依赖与运行时(Electron / Python / Docker 等)是否齐备
- Conflicts with installed plugins: command names, skill / tool names, ports, duplicate MCP registration与已装插件是否冲突:命令名、skill / tool 重名、端口占用、重复 MCP 注册
Security安全性
- Repo matches the facts registered here; archived or abandoned?仓库是否与页面登记一致,是否归档或长期停更
- Safety of preinstall / install / postinstall and install.sh / setup.ps1preinstall / install / postinstall 与 install.sh、setup.ps1 是否安全
- curl|bash, download-then-execute, obfuscation, unrelated domains → stop immediatelycurl|bash、下载即执行、混淆代码、无关域名 → 立刻停止
- Typosquatting or unmaintained packages among the new dependencies新增依赖里有没有 typosquatting 或无人维护的包
- Requested permissions vs. what the feature actually needs申请了哪些权限、是否超出功能所需(filesystem / network / shell / clipboard)
- Any sudo / admin requirement, plus uninstall and rollback是否要求 sudo / 管理员权限,以及卸载与回滚方式
Anything uncertain must be marked unknown with a note on how to confirm it. This site's signal screen is a static snapshot, not a security audit.拿不准的必须标「未知」并说明要我怎么确认。本站的信号筛查是静态快照,不能替代安全审计。
Or use CLI install (for developers)或使用命令行安装(适合开发者)
CLI Install命令行安装
dsh plugin --profile web add @woyeshishen/dsh-vision-plugin
把 woyeshishen/dsh-vision-plugin 加入你的 DSH 配置(web profile)即可启用。
READMEREADME
dsh-vision-plugin
中文 | English
Adds image understanding to DeepSeek Harness (DSH): wire up an OpenAI-compatible external vision model, and a text-only main model (e.g. deepseek) can call the describe_image tool to hand an image to it and get a plain-text description — understanding screenshots, photos, charts, OCR, UIs, and more.
Design: images go only to the secondary model (external vision model); the main model always deals with text.
Features
| 🖼️ Image understanding | The main model calls describe_image and gets a plain-text description |
| ⚙️ GUI configuration | Fill in URL / API key / model on a settings page — no config files to edit |
| 🔒 Secure credentials | API key stored in the credential store, never echoed |
| 📦 Install once, keep working | Auto-loads at DSH startup, survives restarts |
Install
One-liner (recommended)
Windows (PowerShell)
irm https://raw.githubusercontent.com/woyeshishen/dsh-vision-plugin/main/scripts/install.ps1 | iex
macOS / Linux
bash <(curl -fsSL https://raw.githubusercontent.com/woyeshishen/dsh-vision-plugin/main/scripts/install.sh)
dsh plugin command
From npm
dsh plugin --profile web add @woyeshishen/dsh-vision-plugin
From GitHub
dsh plugin --profile web add github:woyeshishen/dsh-vision-plugin
After install, the plugin auto-mounts into the profile; restart DSH (or hot-reload) to activate.
Usage
Step 1: Configure the external vision model
Open Settings → Multimodal Vision:
| Field | Description |
|---|---|
| URL (Base URL) | OpenAI-compatible endpoint, e.g. https://api.example.com/v1 |
| API key | Secret for the external model (stored encrypted, never echoed) |
| Model | Click "Load models" to fetch and pick from the endpoint |
Click Save.
Step 2: Ask the main model to look at an image
In a conversation, say:
Take a look at
D:\path\to\image.pngand describe what's in it.
The main model calls describe_image, sends the image to the external vision model, and continues reasoning from the returned description.
Tool
describe_image
| Parameter | Required | Description |
|---|---|---|
path |
✅ | Image file path; supports png / jpg / jpeg / webp / gif |
prompt |
❌ | Specific question about the image; defaults to "describe the image in detail" |
Requirements
- DeepSeek Harness
- An OpenAI-compatible (
/chat/completions), image-capable external vision model
nexu-io/open-design
Devin-AXIS/iPolloWork
liustack/modlens
Alisa0808/vox-director
EthanYoQ/AI-Novel-Writer
Anionex/agent-vision-toolkit
ysr666/dsh-vision-router
tong-io/tongflow