DeepSeek Harness Plugin Hub

发布与管理完整 Harness Profiles,发现适合你的插件。

探索

插件目录环境预设文档中心动态

社区

发布插件联系我们报告问题

相关链接

Plugin Hub GitHubDeepSeek Harness 官方项目系统状态隐私说明
© 2026 DeepSeek Harness Plugin HubPowered byPaxTech

独立、非官方社区项目,与 DeepSeek 官方无隶属、授权或背书关系。

Show Image — DeepSeek Harness 插件(DSH Plugin)
DeepSeek Harness Plugin Hub
ProfilesPlugins分类动态文档登录管理 Profiles
ProfilesPlugins分类动态文档登录
← Plugins

@aetheri-ai/dsh-show-image

Show Image

DeepSeek Harness 插件:一个面向模型的 show_image 工具,可在对话中向人类查看者展示本地图像。

插件会安装到这里;不确定时保持 web。

npx -y @deepseek-ai/dsh plugin --profile web add @aetheri-ai/dsh-show-image@0.1.1
README兼容性版本

兼容性与来源证明

Show Image 以 @aetheri-ai/dsh-show-image 发布,当前版本为 0.1.1。Plugin Hub 会校验它的 manifest,并保存精确安装来源,便于复现安装结果。

DSH 兼容范围
*
运行环境
any
发布来源
npm
Registry 更新时间
2026/9/20

版本

0.1.1stable
2026/8/29
0.1.0stable
2026/8/29

相关插件

正在加载相关插件…

最新版
0.1.1
DSH
*
HMR
重启进程
Tree shaking
未声明可安全裁剪
解包体积
15.9 kB
文件数
6
Surface
any
许可证
MIT
发布源
npm
GitHub
★ 0
周下载
41
最近提交
2026/8/29
查看源码 ↗项目主页 ↗
README Badge

点击下方 Badge 复制 Markdown,粘贴到 README 即可。

这是你的 Plugin?认领权益 · 优先安全扫描

验证 package.json 声明的 GitHub 仓库,即可管理这个公开页面。认领后,Hub 会优先安排当前版本的安全扫描,并在通过后公开展示结果。

认领这个 Plugin →
报告问题

相关插件

继续浏览 vision-media 分类下经过校验的插件。

Tool Describe Image@linxin666/dsh-tool-describe-image面向模型的 describe_image 工具,用于 dsh Web GUI:通过在兼容 OpenAI 的端点调用视觉语言模型,为文本模型提供图像理解能力,以描述一张图像(本地路径、http(s) URL 或附件引用)。可热插拔 —Modlens@liustack/modlens面向仅支持文本的 LLM 的插件视觉能力,由免费的 Antigravity CLI 提供支持Vision Toolkit@anionex/dsh-vision-toolkit面向 Harness 原生集成的 DeepSeek 与 agent-vision-toolkit:图像问答、OCR、定位、UI 还原、像素差异、Artifacts 和 Web UI。Deepseek Ivideodeepseek-ivideoiPolloWork HyperFrames Video Studio,以及 27 个可编辑视频模板,以原生 DeepSeek Harness 对话视图呈现。

README

@aetheri-ai/dsh-show-image

English | 中文

A DeepSeek Harness plugin that adds the show_image tool: the model presents an existing local PNG, JPEG, WebP, or GIF file to the human viewer, rendered inline in the Web conversation transcript.

Install

Install the bundle on a profile whose base supplies the filesystem, tool runtime, and durable attachments (dsh-base covers all three). Choose the command that matches how dsh is installed:

# dsh installed globally
dsh plugin --profile web add @aetheri-ai/dsh-show-image

# Run dsh directly from npm
npx @deepseek-ai/dsh plugin --profile web add @aetheri-ai/dsh-show-image

# A deepseek-harness source checkout
pnpm dsh plugin --profile web add @aetheri-ai/dsh-show-image

Restart the Web profile, then start a new conversation. A recently published release may not have reached a configured registry mirror; append --registry=https://registry.npmjs.org to the install command when necessary.

Confirm that the bundle is composed:

dsh --profile web --dump-config

The output contains an @aetheri-ai/dsh-show-image section and a show-image row. Remove the bundle with:

dsh plugin --profile web remove @aetheri-ai/dsh-show-image

Use

In a new session, ask the model to display an existing supported image, for example Show docs/architecture.png in the conversation. The model calls show_image with:

{ "file_path": "docs/architecture.png" }

Relative paths resolve against the session workspace. The tool commits the picture to the durable attachment store and returns an image block that the Web client renders inline (with the existing load, retry, and lightbox behavior). The active model route still needs the normal dsh provider/API-key configuration to make a tool call.

Design

  • Presentation-only: the tool exists for the human viewer. It does not require a vision-capable model route — use read_image when a model must inspect pixels.
  • Policy-respecting: paths resolve through the filesystem provider, so sandbox confinement and fs policy apply unchanged.
  • Durable: images are stored as content-addressed attachments, never as temporary paths or inline base64 in the log.
  • Conditional: the tool registers only while a durable attachment store is mounted.

Compatibility

Peer dependencies track the dsh npm release line (^0.1.1-rc.2, cordis ^4.0.1). dsh is in developer preview; this plugin updates with each breaking release.

Development

pnpm install
pnpm run test        # unit + real-Loader composition tests
pnpm run build       # tsdown bundle into lib/
pnpm run typecheck

License

MIT