DeepSeek Harness plugin

dsh-codex-media-yujianji

Codex-powered image understanding & generation for DeepSeek Harness: paste images into the Web GUI (converted to sha256 text refs via an injected paste bridge) and let the ChatGPT-Plus-backed Codex

Jump to install

Source facts

Repository
yujianjian1013/dsh-codex-media
Latest update
Aug 21, 2026
Category
Tools & Capabilities
GitHub stars
0
Format
plugin
Catalog evidence
Upstream dsh.bundle evidence
Evidence path
package.json#dsh.bundle
Checked against
0.1.0-rc.8
Upstream check date
2026-08-20

This evidence comes from the upstream catalog. This site has not installed, run, or security-reviewed the plugin.

Install

Start with a prompt that asks an agent to review the GitHub repository and source. Switch to the command if you want to install it yourself.

Copy this prompt into DSH, Codex, or another agent and ask it to review the GitHub repository and source first.

Do not install or run any commands yet. Read this plugin's GitHub repository, README, and relevant source code. Then answer the questions below clearly and directly so I can decide whether it fits my needs:

1. What is this plugin, and what problem does it solve?
2. Who is it for, and what are its typical use cases?
3. How is it used after installation? Include one minimal example.
4. What known limitations or privacy, security, compatibility, or maintenance risks does it have?
5. Give a clear recommendation: recommend, conditionally recommend, or do not recommend, with reasons.

Distinguish statements documented by the repository, inferences from source code, and unknowns. If evidence is insufficient, say so explicitly. Do not guess or simply repeat the README.

GitHub: https://github.com/yujianjian1013/dsh-codex-media
Plugin: dsh-codex-media-yujianji
Author: yujianjian1013

Check the source files

Read the README and other files from this plugin directory before installing.

File explorer3 files
README.mdSource · read only

dsh-codex-media

用本机 Codex(ChatGPT Plus 额度)给 DeepSeek Harness 补上「生图 / 改图」,并可把图片写入剪贴板/委托 Codex 独立任务。

图片理解不再走本插件:DeepSeek 多模态模型(deepseek-v4-flash-vision-exp)原生支持图片输入,粘贴的图片会直接作为 image block 发给模型。本插件只负责 Codex 擅长的生成/编辑

一个包,一个角色 + 一个兼容壳:

1. MCP server(mcp/server.mjs,stdio)——四个工具,经 dsh-mcp-client 注册为 mcp__codexmedia__*:

工具作用
------
generate_image生图:Codex 托管 image_generation,输出到你传的路径
edit_image改图:带参考图重绘(参考图接受本地路径 / sha256 附件 id / clipboard)
clipboard_to_image剪贴板图片存盘(pwsh -STA 助手)
codex_task把一段独立任务交给本地 Codex 执行(读/改文件、跑命令),省 DeepSeek token

2. Host 插件(lib/host.js)——无操作占位,仅让包作为合法的 dsh profile bundle 存在;已移除原先把粘贴图片转成 sha256: 文本引用的粘贴桥。

依赖

  • DSH ≥ 0.1.0-rc.6,Node ≥ 22.19
  • 本机安装并登录 ChatGPT 的 Codex CLI(生图/改图走 Plus 额度)
  • Windows(剪贴板助手用 pwsh -STA;非 Windows 上 clipboard 输入不可用,其余功能正常)

安装

# 1) 把插件装进 profile 并注册为 bundle(参考 dsh-vision-proxy 的安装方式)
dsh plugin --profile web add github:yujianjian1013/dsh-codex-media
# 或手动: profiles/<name>/package.json 的 dependencies 加
#   "dsh-codex-media": "github:yujianjian1013/dsh-codex-media"
#   并在 dsh.profile.bundles 里加 "dsh-codex-media",然后 pnpm install

包内的 cordis.patch.yml 会自动插入 host 插件行(现在是 no-op)。MCP 行是机器相关的,加到你 profile 自己的 cordis.patch.yml:

- insert:
    - id: mcp-codexmedia
      name: '@deepseek-ai/dsh-mcp-client'
      config:
        serverName: codexmedia
        transport: stdio
        command: node
        args: !!js '(() => { const { createRequire } = process.getBuiltinModule("node:module"); const home = process.env.DSH_HOME || ((process.env.USERPROFILE || "") + "/.dsh"); return [createRequire(home + "/profiles/web/package.json").resolve("dsh-codex-media/mcp/server.mjs")]; })()'

重启 dsh web,刷新页面。

使用

  • 识图(原生):直接用 DeepSeek 多模态,粘贴图片即会作为图片发给模型,无需任何插件。
  • 生图:直接说"画一只……"→ 模型调 generate_image,output_path 按会话工作目录传入。
  • 改图:粘贴参考图 + "把它改成……"→ edit_image 带参考图重绘。
  • 剪贴板:说"把剪贴板图片存到 X"→ clipboard_to_image
  • Codex 子代理:codex_task(独立任务,非图处理)。

注意

  • 生图/改图消耗 ChatGPT 订阅的图像额度;原图留存在 ~\.codex\generated_images,可定期清理。
  • Codex 连接 chatgpt.com 时 WebSocket 偶发重置(10054)会自动降级 HTTPS 重试,属正常现象。
  • scripts/ 里保留了一个不依赖 DSH 的纯 PowerShell 版(codex-media.ps1,见 scripts/README.md)。

License

MIT