vane-gpt-image-skills README

The project's README from github.com/VaneAi888/vane-gpt-image-skills. As one markdown file: README.md.

🎨 Vane GPT Image Skills

GPT Image 2.6 好玩又能用的开源玩法合集,持续更新

𝕏 关注我的 X 主页 · @VaneAILab

这里收集可以直接上手的 AI 生图 Skill。每个玩法都有自己的效果展示和使用说明,从角色表情包开始,以后也会逐步收录封面、配图和其他创意玩法。

目前已收录 4 个 Skill。选择感兴趣的玩法,先看效果,再跟着教程开始制作。

📚 Skill 合集

🎭 角色与表情包

照片转动态表情包

上传人物照片、宠物照片或已有角色,选择喜欢的画风和聊天主题,让专属形象用表情和动作说话。

12 种画风 · 6 个主题 · 默认一套 16 张

12 种画风 · 12 组动态演示

下面展示 4 种画风,每组包含 4 个动作小样。点击动图可放大查看。

Q 版萌系 手绘涂鸦风
Q 版萌系四动作动态演示 手绘涂鸦风四动作动态演示
毛绒布偶风 复古漫画风
毛绒布偶风四动作动态演示 复古漫画风四动作动态演示

👉 查看全部 12 种画风、6 个主题与使用教程

🎬 电影海报与视觉设计

电影级海报设计

根据影片故事和素材,选择构图、设计文字层级,制作中文或英文电影海报方案、提示词或实际图片。

36 种版式 · 中英文各 4 组预览 · 默认竖版 2:3

英文海报预览

01—09 · 人物主导 10—18 · 字体主导
电影海报版式 01 至 09 人物主导 电影海报版式 10 至 18 字体主导
19—27 · 场景与空间 28—36 · 概念与图形
电影海报版式 19 至 27 场景与空间 电影海报版式 28 至 36 概念与图形

中文海报预览

01—09 · 人物主导 10—18 · 字体主导
中文电影海报版式 01 至 09 中文电影海报版式 10 至 18
19—27 · 场景与空间 28—36 · 概念与图形
中文电影海报版式 19 至 27 中文电影海报版式 28 至 36

中英文预览分别设计,并非逐字翻译。图片为构图示例,其中的人物姓名、片方及荣誉文字不代表真实影片信息或授权背书。

👉 查看 36 种版式介绍与完整使用教程

书页里的电影

从书中的人物、故事或意象出发,把阅读时的想象变成电影感海报。小说和非虚构作品都可以尝试。

36 本书籍效果示例 · 4 组完整预览 · 默认竖版 2:3

01—09 10—18
书页里的电影示例 01 至 09 书页里的电影示例 10 至 18
19—27 28—36
书页里的电影示例 19 至 27 书页里的电影示例 28 至 36

以上是书籍海报的创意展示,非官方改编或授权海报。

👉 查看全部书籍示例与使用教程

🧧 节日与生活创意

微信红包封面制作

用文字或参考图设计专属微信红包封面,先确认原图,再查看模拟效果并导出上传图片。

36 种风格参考 · 4 组完整预览 · 单张封面设计

01—09 · 传统艺术 10—18 · 趣味角色
红包封面风格 01 至 09 传统艺术 红包封面风格 10 至 18 趣味角色
19—27 · 现代设计 28—36 · 自然艺术
红包封面风格 19 至 27 现代设计 红包封面风格 28 至 36 自然艺术

这些是风格模拟预览,不是可直接上传的封面原图,也不是微信官方模板。

👉 查看 36 种风格与完整使用教程

🚀 开始使用

先克隆合集,再选择安装单个或全部 Skill:

git clone https://github.com/VaneAi888/vane-gpt-image-skills.git
cd vane-gpt-image-skills
python3 scripts/install_skills.py --all

安装器不会覆盖已存在的同名 Skill。如果只想安装其中一个,请查看统一安装教程。安装完成后开启新的 Codex 对话,用 $vane-portrait-gif-stickers、$vane-cinematic-poster、$vane-book-cinema 或 $wechat-red-packet-cover 调用。

👉 统一安装教程 · 效果与安装测试记录

使用前请确认环境支持读取参考图和图像生成;GIF 动作、图片文字和透明背景的效果取决于实际素材与可用工具。

✨ 持续更新

未来计划按用途整理更多生图玩法,例如封面与社交媒体、文章与知识配图、照片与生活创意,以及趣味视觉实验。新 Skill 完成并具备使用说明后,会加入对应分区。

👋 关于作者

我是 韦恩 Vane,一名关注 AI 创作与品牌视觉设计的设计师。我分享 AI 工具、视觉创意和实用工作流,也会把有趣的实践整理成开源 Skill。

欢迎关注我的 X 主页 @VaneAILab,一起交流 AI 工具与创作玩法。

📄 许可证

项目代码采用 MIT License。示例中涉及的第三方角色仅用于非官方风格演示;使用照片或角色素材时,请确认自己有相应使用权。

vane-gpt-image-skills (VaneAi888/vane-gpt-image-skills)

Searching Database…

What is this?A curated collection of ready-to-use AI image generation skills for Codex, built around GPT Image 2.6. Each skill is a fun, practical creative workflow with its own demo results and instructions.
Is it one tool, or a collection of separate skills?It's a collection of separate skills, not one app. You can install a single skill or all of them, and each one is then called by its $name inside Codex.
How do I get them running?Clone the repository, then run the installer to choose one skill or all of them. After installing, open a new Codex chat and call a skill by name, for example '$wechat-red-packet-cover'.
What do I need before I can use them?A Codex environment that can read reference images and generate images. Because the skills depend on image generation, you need that capability available in your setup.
Will the results always come out as shown in the demos?Not necessarily. GIF motion, text inside images and transparent backgrounds depend on your source material and the tools you have. The red packet outputs are style previews, not upload-ready WeChat templates.
Can I make a sticker set from my pet's photo?Yes. The portrait GIF sticker skill takes a photo of a person, pet or existing character, lets you choose one of 12 art styles and 6 chat themes, and produces a 16-sticker set by default.
Are the movie poster skills based on real films?No. The author says the sample images are composition examples, and any names or credits shown do not reflect real films or endorsements.
What does the author plan to add next?More skills organized by use case, such as covers and social media, article and knowledge illustrations, photo and lifestyle creativity, and visual experiments.
Creator
VaneAi888
Size
66 MB
Stars
428
Updated
CommitNov 17
Age
13d
Commits · 90 Days