Codex Dream Skin logo

Codex Dream Skin

Community
xnydl
codex-dream-skin

Apply, launch, verify, repair, update, or restore a full decorative skin for the Windows Codex desktop app. Use when the user asks for a Codex theme that goes beyond official color settings, wants the pink-purple Dream/Fiona-style interface, needs the skin reapplied after a Codex update, or needs a safe rollback without modifying WindowsApps or app.asar.

Overview

Publisherxnydl
Repositorycodex-dream-skin
Skill namecodex-dream-skin
Stars
54
Forks
2
Bundled files
23
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 23 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by xnydl on GitHub. Read the source before you install it.

Installation

Install the Codex Dream Skin AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/xnydl/codex-dream-skin/tree/main/windows
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/xnydl/codex-dream-skin.git /tmp/codex-dream-skin
mkdir -p .claude/skills
cp -r /tmp/codex-dream-skin/windows .claude/skills/codex-dream-skin
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Codex Dream Skin in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Codex Dream Skin on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Codex Dream Skin is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Codex Dream Skin

Apply a reversible renderer skin through Chromium DevTools Protocol while launching the official Store-installed Codex executable. Never replace or take ownership of files under WindowsApps.

Workflow

  1. Install Node.js 22 or newer, close Codex, then run scripts/install-dream-skin.ps1 once to preserve the user's native appearance settings, seed the bundled 世事宜梦境 (ssyai-dream) theme, and create launch/restore/tray shortcuts.
  2. Run scripts/start-dream-skin.ps1. The shortcut asks before restarting an already-open Codex app; CLI callers must explicitly add -RestartExisting.
  3. Run scripts/verify-dream-skin.ps1 -ScreenshotPath <absolute-path> after launch. Treat a missing continuous wallpaper, home shell, native composer, sidebar layer, or injection marker as failure. The native suggestion count is responsive and may be two to four.
  4. Inspect the screenshot against references/qa-inventory.md. Verify both the home screen and a normal task before signing off.
  5. Run scripts/restore-dream-skin.ps1 to remove the live skin, close the saved CDP session, and reopen Codex normally. Add -RestoreBaseTheme to restore only saved appearance keys, -RecoverConfigBackup for explicit byte-for-byte recovery of a damaged config, or -Uninstall to delete shortcuts. A completed config restore archives that install's backup so a later install captures a fresh baseline.

Guardrails

  • Preserve the official executable, package signature, user threads, pets, plugins, and authentication state.
  • Theme images must be UI-free wallpapers. Never import a README screenshot, fake window, sidebar, card, composer, logo, or text baked into the bitmap.
  • Paint one continuous 16:9 wallpaper across the full Codex window. Let the sidebar, main area, header, and composer act as coordinated readability layers; keep the home route expressive and task routes quieter.
  • appearance: auto follows the native computed color-scheme first and the system appearance only as a fallback. Image brightness may tune color and composition, but must not flip the user's shell mode; explicit light and dark remain authoritative.
  • Attach the "选择项目" treatment to Codex's real project-selector toolbar and keep the current project button clickable; never draw a disconnected replacement.
  • Keep decorative layers pointer-events: none and keep real buttons, navigation, and composer above them.
  • On app updates, rerun install and launch; the scripts discover the current Appx package dynamically. Saved paths are never trusted for process control unless they still match a registered package identity.
  • The default launcher scans for a free port when 9335 is occupied. An explicitly requested occupied port fails closed.
  • Keep the injection daemon running for navigation/reload resilience. Its state and logs live under %LOCALAPPDATA%\CodexDreamSkin.
  • The watcher registers a generation-checked early payload for connected renderers so reload/navigation can paint the skin before the normal load-event fallback; unsupported CDP targets fall back safely.
  • The active theme, saved themes, imported images, pause marker, and tray controls live under %LOCALAPPDATA%\CodexDreamSkin. Reject empty or over-16 MB images before copying or encoding them.
  • Every managed-store write rejects junctions and other reparse points in every existing path component. Imports also use the bundled Node metadata parser before copying to reject dimensions above 16384px or 50MP.
  • CDP targets must use a same-port loopback WebSocket, belong to the current Store package, retain the launch-time Browser ID, and expose expected Codex renderer markers.
  • Loopback prevents LAN exposure, but Chromium CDP has no same-user authentication. Run only trusted local software while the skin is active, and use restore to close the debug session when it is no longer needed.
  • Preserve config.toml as strict UTF-8. Never use encoding-dependent whole-file PowerShell reads/writes, silently transcode UTF-16, or overwrite a file that changed after it was read. Ambiguous TOML shapes must fail before writing rather than receive a best-effort rewrite.
  • Keep install/start/restore/verify serialized with the per-user operation lock in common-windows.ps1.

Checks

powershell
powershell -NoProfile -File tests\run-tests.ps1
node --check scripts\injector.mjs
node --check assets\renderer-inject.js

Resources

  • scripts/injector.mjs: CDP connection, renderer injection, verification, screenshot, and removal.
  • scripts/common-windows.ps1: Store-package discovery, Node validation, port ownership, state, and process identity safety.
  • scripts/config-utf8.ps1: atomic UTF-8 configuration backup, selective restore, and explicit recovery.
  • assets/dream-skin.css: full visual layer.
  • assets/renderer-inject.js: idempotent DOM integration and cleanup.
  • assets/dream-reference.jpg: pure 2560 × 1440 世事宜梦境 abstract wallpaper (procedurally generated, no photos or third-party art) seeded as the default and as a saved theme; it contains no Codex UI.
  • assets/theme.json: shared adaptive theme contract for the seeded preset.
  • scripts/theme-windows.ps1: persistent active/saved theme store, safe image import, pause state, and preset seeding.
  • scripts/tray-dream-skin.ps1: Windows Forms tray for apply, pause, import, save, switch, and complete restore.
  • references/qa-inventory.md: required functional and visual signoff coverage.
  • references/runtime-notes.md: troubleshooting and update behavior.
  • tests/run-tests.ps1: configuration, state, recovery, payload, and CDP validation regression checks.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Codex Dream Skin AI skill do?

Apply, launch, verify, repair, update, or restore a full decorative skin for the Windows Codex desktop app. Use when the user asks for a Codex theme that goes beyond official color settings, wants the pink-purple Dream/Fiona-style interface, needs the skin reapplied after a Codex update, or needs a safe rollback without modifying WindowsApps or app.asar.

Why use Codex Dream Skin on TypingMind?

Because you install it once and use it with any model. Codex Dream Skin is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Codex Dream Skin in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/xnydl/codex-dream-skin/tree/main/windows. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Codex Dream Skin?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Codex Dream Skin?

As many as you like. As long as a model supports skills, you can use Codex Dream Skin with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Codex Dream Skin AI skill free?

Yes. It is published on GitHub by xnydl under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇