stoneface10/dsh-codex-connect-plus 预览 preview

stoneface10/dsh-codex-connect-plus

Plugin插件 Native原生 ⭐ 4 Apache-2.0 Vision & Media视觉与多媒体Models & Routing模型与路由

在你的DeepSeek Harness中使用ChatGPT订阅,即可访问Codex模型,并通过gpt-image-2生成或编辑图片——无需OpenAI平台API密钥。

Project Overview项目介绍

This is a native plugin built exclusively for DeepSeek Harness (DSH) that enables DSH users to access OpenAI Codex models and gpt-image-2 image generation using an existing personal ChatGPT/Codex subscription. It eliminates the requirement for a separate OpenAI Platform API key, so users do not need to set up pay-as-you-go API billing. All requests are authenticated via a one-time ChatGPT OAuth login, and usage draws directly from the signed-in user’s existing subscription quota. It integrates seamlessly with DSH’s native model picker, so no separate CLI workflow is required to use Codex models.

It is designed for DSH users who already hold an active ChatGPT/Codex subscription but prefer to use that subscription instead of a separate OpenAI API key. To install the plugin from the public npm beta, run dsh plugin --profile web add dsh-codex-connect-plus@beta from your DSH command line interface. After installation completes, navigate to the DSH settings menu: Settings → Plugins → Plugin Configuration → Codex Connect Plus. Once there, complete the ChatGPT OAuth login, adjust your optional capability and quota settings to your preferences, then save your profile changes, and you can start using Codex immediately.

The plugin requires DeepSeek Harness version 0.1.2-alpha.1 or newer, and Node.js version ^22.19.0 or >=24.0.0 to run properly. It follows strict security practices: OAuth credentials are stored only in DSH’s dedicated credential file, sensitive authentication data is redacted from errors, and image requests use fixed HTTPS endpoints that reject untrusted redirects. It is released under the open-source Apache-2.0 license, with proper attribution retained for all adapted third-party code. There are clear usage limits: each reference image is capped at 4MB, and you can include a maximum of 8 reference images per edit request.

这是一个专为DeepSeek Harness(DSH)开发的原生插件,支持用户在DSH中直接使用已有的ChatGPT/Codex订阅配额调用Codex大模型和gpt-image-2图像服务,无需额外申请OpenAI Platform API密钥。用户只需通过ChatGPT OAuth完成一次登录,模型请求和图像生成编辑都会使用已登录账号的订阅额度。它支持通过DSH原生的模型选择器选取Codex模型,无需额外的独立CLI工作流。

它面向拥有ChatGPT/Codex订阅但没有OpenAI平台API密钥,或者更愿意使用现有订阅配额的DSH用户。典型工作流为:通过DSH插件命令行工具完成安装后,进入设置→插件→插件配置页面,使用ChatGPT完成OAuth登录,按需调整可选功能开关和配额安全设置,保存配置后即可在对话中正常调用Codex模型和gpt-image-2图像生成编辑功能。

本插件要求DeepSeek Harness版本不低于0.1.2-alpha.1,Node.js版本需要^22.19.0或者>=24.0.0。它采用Apache-2.0开源许可,OAuth凭证仅保存在DSH专用凭证文件中,符合安全规范。图像生成请求不自动重试,单张参考图像不超过4MB,单次编辑最多支持8张图像,功能可用性受上游Codex服务政策和账号权限限制。

Pre-install check安装前体检Compatibility · Security兼容性 · 安全性 1 warning1 项注意
  • Only 4 stars - very few users, little community feedback星标只有 4,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 'github:stoneface10/dsh-codex-connect-plus#v0.1.0-beta.6'

把 stoneface10/dsh-codex-connect-plus 加入你的 DSH 配置(web profile)即可启用。

READMEREADME

Codex Connect Plus

Release License

English | 中文

Codex Connect Plus — ChatGPT subscription to Codex models and gpt-image-2 without an OpenAI Platform API key

Use your ChatGPT/Codex subscription inside DeepSeek Harness for Codex models and gpt-image-2 generation/editing—without an OpenAI Platform API key.

Sign in once with ChatGPT OAuth. Codex model requests use the signed-in account's subscription capacity; image generation/editing uses the same OAuth session and remains subject to account availability, region, limits, and upstream policy.

What you get

  • Codex models in the normal DSH model picker — no separate CLI workflow.
  • gpt-image-2 generation and editing — create images from prompts or edit 1–8 local references, with inline previews and local output files.
  • No OpenAI Platform API key or pay-as-you-go API billing setup — authorization comes from your user-approved ChatGPT/Codex OAuth session.
  • Quota-conscious defaults — automatic model retries default to 0, and image requests are never silently retried.

Real DeepSeek Harness UI showing Codex model selection and gpt-image-2 image generation

Real DSH UI: select a Codex model and generate an attached image with gpt-image-2 in the same conversation.

Community derivative. Not affiliated with or endorsed by OpenAI, DeepSeek, or the upstream Codex Connect maintainers.

Features

  • ChatGPT/Codex OAuth login and provider-native automatic refresh.
  • OpenAI Codex model catalog through Harness's normal LLM service.
  • Optional standalone Codex search provider.
  • Optional view_image tool for vision-capable models.
  • codex_image_generate for 1-4 gpt-image-2 outputs.
  • codex_image_edit for 1-8 local PNG/JPEG/WebP references and an optional mask.
  • Durable DSH attachments, session-authorized replay, inline preview, and local files under outputs/codex-image.
  • Fixed upstream origins, no redirects, bounded inputs/responses, cancellation, timeout, strict image signatures, and redacted provider errors.

Showing the opening section of the README — the full document lives in the repository以上为 README 开头摘要,完整文档在仓库内 · View the full README on GitHub →在 GitHub 查看完整 README →

← 上一个 Prev dsh-multi-model-orchestrator 下一个 Next vibemeter →