DSH Hub

chenbin-dev/dsh-auth-everying

dsh-auth-everying

UIWeb UI2 GitHub stars· updated 2026-08-21

dsh-auth-everying is a community DeepSeek Harness plugin. Read the repository README before installing.

Install

npx @deepseek-ai/dsh plugin --profile web add dsh-auth-everying

Restart `dsh web` after install. Bundle APIs can change during the developer preview.

README badge

dsh-auth-everying DSH Hub badge
[![DSH Hub](https://dshhub.dev/badge/dsh-auth-everying.svg)](https://dshhub.dev/plugins/dsh-auth-everying)

Paste this into your README. The star count updates with every catalog sync.

From the README

Excerpt from chenbin-dev/dsh-auth-everying, cleaned of badges and images.

dsh-auth-everying

dsh-auth-everying is a DeepSeek Harness plugin for importing local coding-platform login state and provider configuration. It lets you choose which imported models appear in the DSH model picker.

Features

  • Imports local Claude, Codex, Grok, Gemini, Copilot, OpenCode, and CC Switch configuration.
  • Supports OAuth login for compatible official providers.
  • Discovers models from OpenAI-compatible CC Switch gateways through /v1/models or /models.
  • Exposes the Codex app-compatible low, medium, high, xhigh, and max levels without changing their wire values. The unsupported minimal level is hidden for CC Switch Codex routes.
  • Uses medium as the concrete Codex default so DSH does not add its separate Default provider-default option.
  • Adds ultra only to models that declare or are known to support it. Non-reasoning and unsupported models do not show that selector.
  • Preserves the configured Codex model_reasoning_effort and sends low, medium, high, xhigh, max, and ultra with the same wire value.
  • Supports Windows installation and build workflows.

Requirements

  • A working DeepSeek Harness installation with the dsh command available in PowerShell or a terminal.
  • A web profile, because the settings panel is provided by the DSH web interface.
  • Node.js 22.19 or later only when building the project from source.

Install

After the package is published to npm, the shortest installation command is:

dsh plugin --profile web add dsh-auth-everying

Until the first npm release is published, install directly from the public GitHub repository:

Run these commands in PowerShell or a terminal:

dsh plugin --profile web add github:chenbin-dev/dsh-auth-everying
dsh web

The package name and the GitHub repository are different distribution channels. Publishing the repository does not by itself publish the npm package.

If DSH is already running, restart dsh web after installing or updating the plugin so the new bundle is loaded.

First-Time Setup

  1. Open the DSH web interface and go to Settings > dsh-auth-everying.
  2. In Sources, select the local login or provider configuration to import.
  3. Click the import action and wait for the scan to finish.
  4. In Imported, enable the models that should be available in DSH.
  5. Start a new conversation or refresh the model picker.

The plugin reads local configuration from common locations, including CC Switch, Codex, Claude, Grok, Gemini, and OpenCode. It does not require copying API keys into the project or into the README.

CLI

Show available sources and imported routes:

dsh plugin --profile web exec dsh-auth-everying status

Import sources by ID after checking status:

dsh plugin --profile web exec dsh-auth-everying import live:codex-auth live:grok-auth

Use the source IDs reported by status when your local installation exposes different IDs.

Troubleshooting

The plugin does not appear in Settings

Confirm that the install command completed successfully, stop the existing dsh web process, and start it again. Also verify that the command uses the web profile.

The imported route has no models

Open the plugin settings and run the source scan again. For a CC Switch OpenAI-compatible provider, confirm that its gateway is reachable and supports either /v1/models or /models. The configured default model is retained when the gateway does not provide a model list.

Related plugins