Vane GPT Image Skills
Vane GPT Image Skills: two Codex skills for GIF sticker packs from a photo and pixel covers with illustrations for X articles
- 112 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- The skill reads uploaded photos and characters and sends them to an image generation model
- You need rights to the photos and characters you use
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/vane-gpt-image-skillsDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The skill reads uploaded photos and characters and sends them to an image generation model
- You need rights to the photos and characters you use
This is third-party code. Review the repository files before installing.
What it does
Vane GPT Image Skills collects skills for image generation with the GPT Image model. The collection now holds two skills. vane-portrait-gif-stickers turns an authorized photo of a person, a pet or an existing character into a recognizable hero and makes a series of animated GIF stickers by chat theme: 12 styles, 6 themes, 16 stickers by default with short Chinese phrases. vane-pixel-visuals prepares X article covers in 5:2, 16:9 body illustrations, process diagrams and comparisons in one purple pixel style. A script does the installing: after git clone you run python3 scripts/install_skills.py --all, and it does not overwrite existing skills with the same name. Then open a new Codex conversation and call $vane-portrait-gif-stickers or $vane-pixel-visuals. The repository holds previews, frame and pack export scripts and documents for installing and testing. The author warns that the environment must be able to read reference images and generate pictures, and that GIF motion, text in images and transparent backgrounds depend on the material and available tools.
Who it is for. Authors and designers who make stickers from their own photos and covers for X articles and are ready to work with GPT Image in Codex.
Good fit when
- You want a set of 16 animated stickers from your own photo or pet in anime, pixel or ink style
- An X article needs a 5:2 cover and 16:9 illustrations in one branded style
- A blog needs process diagrams and comparisons in a single pixel look
- You need both skills installed with one script and no overwriting of existing ones
- You want to browse previews of 12 styles and pick one
Not a fit when
- Using other people's photos or protected characters without rights: the author asks you to confirm you hold them
- You need a tool for something other than Codex: the ready calls are written as $vane-portrait-gif-stickers
- The environment cannot read reference images or generate pictures
- You want a cover in another style: vane-pixel-visuals keeps one purple pixel look
Example request
Make a set of 16 pixel-style GIF stickers on the work theme from this photo of my catLimitations
- You need Codex and image generation access with GPT Image and reference reading.
- GIF motion, text in pictures and transparency depend on the material and tools.
- Use photos and characters only with rights to them, and the README shows examples with third-party characters as an unofficial style demonstration.
- The collection holds two skills for now, and the author only plans the rest.
- Sticker captions default to Chinese.
How to disable. Delete the vane-portrait-gif-stickers and vane-pixel-visuals folders from the Codex skills directory where the installer put them.
FAQ
What is in Vane GPT Image Skills?
Two skills: vane-portrait-gif-stickers for a series of GIF stickers from a photo or character and vane-pixel-visuals for covers and illustrations for X articles in a purple pixel style.
How do I install the skills?
Clone the repository and run python3 scripts/install_skills.py --all. The script does not overwrite existing skills with the same name. Then start a new Codex conversation.
How many stickers do I get and in which styles?
16 per series by default, and the README lists 12 styles, from chibi and realistic to pixel, anime, plush, papercut and ink, and 6 themes such as work, AI and celebration.
What format are the covers?
An X article cover is made in 5:2 and body illustrations in 16:9, all in the author's single purple pixel brand style.
What do the skills need to work?
Codex and an environment with GPT Image generation access and reference image reading. The result for GIF motion, text and transparency depends on the material and available tools.
Discussion
No questions or reviews yet.
Related
SkillsA skill and CLI for generating and editing images with GPT Image 2 and 2.5 plus a library of ready prompts
Five image-generation skills: Qwen-Image prompts, logos with shared geometry, diary comics, transparent PNGs and process dioramas
GPT-Image2 Style Library
gpt-image-2-style-library
A skill picks a matching template and style from a library of GPT-Image2 cases and assembles it into a finished prompt
A Codex skill that turns one character image into nine transparent animated GIF stickers in 36 styles.
A skill that adds image generation to Claude Code and Cursor through the Codex CLI on a ChatGPT subscription, no API key.
Sprite Gen
sprite-gen
A skill for Codex and Claude and a Python CLI: builds 2D game sprites from one base image, removes the background and assembles animation atlases
README in short
The README presents the collection as a set of working image generation scenarios: it shows GIF sticker previews across 12 styles and an example cover. It also covers script installation, calling in Codex, plans for new skills, the author's details and the MIT license with a note on third-party content.
README badge
Original SKILL.md
--- name: vane-portrait-gif-stickers description: Turn an authorized photo or existing character into a consistent character and genuinely animated GIF sticker packs in selectable styles, from Q版、真人实拍、像素到水墨. Use for 照片或现有角色制作动态表情包、个人专属表情包、韦恩表情包 and sticker-pack previews. --- # 照片或角色变多画风动态表情包 用用户提供或授权的照片或现成角色，按选定画风制作可辨认的角色，再按聊天场景制作不同表情、动作和循环 GIF。一个系列默认 16 张、中文短句；用户可指定数量、语言、人物或角色名、画风和输出规格。保持脸部、发型、服装或角色造型等识别点，不把所有角色画成统一模板。 ## 角色参考 - 用户提供照片时，用图像生成工具按选定画风制作清晰的角色参考，展示给用户；后续动作引用照片和角色参考。用户提供已有角色时，默认保留其外观和画风，除非用户要求转换；动物、吉祥物等非人角色不转成人物。若没有照片或现成形象，先请用户提供。已有多张图片时遵循用户指定的身份参考。 - 用户明确要求“韦恩”时，若本机存在 `~/.codex/wayne-sticker-profile/wayne-photo.png` 和 `~/.codex/wayne-sticker-profile/wayne-character-reference.png`，将它们作为参考；否则请用户提供参考图。用户上传自己的照片时，使用其照片，不使用韦恩参考，也不要把人物命名为韦恩。 - 这些个人参考保存在 Skill 目录之外，不加入发布包或输出 ZIP。仅在用户要求时长期保存他人照片或形象；不要静默采集、上传到外部服务、发布或把个人形象放入开源文件。图像生成由用户的制作请求授权，遵循所用工具的正常处理方式。 ## 画风选择 画风是角色的视觉呈现，主题决定聊天场景和文案；两者可自由组合。用户已指定画风或提供现成角色时直接按要求执行；用照片制作但未指定画风时，简洁询问用户选择，也可使用自定义画风或参考图。用户没有偏好或要求直接开始时，照片默认使用 Q 版萌系；已有角色默认保持原画风。用户要求对比多种画风时，先做指定的少量角色样张，不自动生成多整套。 1. **Q版萌系**：头身比例可爱、表情清晰、轮廓简洁，保留人物识别特征。 2. **真人实拍风**：写实人物比例、自然皮肤与摄影质感，保持面貌和服饰；生成画面应明确为生成素材，不说成实际拍摄。 3. **像素风**：有限色板、清晰像素网格和可识别的表情；导出缩放优先最近邻，避免模糊像素边缘。 4. **日漫二次元**：日式动画感、干净线条与分层上色，保留角色辨识度。 5. **3D立体风**：突出雕塑般的立体造型、清楚的体积、材质与光影；如需卡通 3D，可明确要求可见的建模造型，避免结果难以与真人摄影区分；不默认做成毛绒材质。 6. **手绘涂鸦风**：自然不规则线条、手绘笔触和活泼配色，确保小尺寸仍清晰。 7. **扁平矢量风**：简化几何形状、平面色块和清楚轮廓；只描述为矢量视觉风格，除非实际交付可编辑矢量文件。 8. **极简线稿风**：少量简洁线条和留白，靠轮廓与表情传达动作；保证线条在小尺寸可见。 9. **毛绒布偶风**：柔软绒毛、布偶造型与触感，面部、四肢等主体部位也呈现织物或填充材质；保留角色独特脸部和配色，避免只给真人照片的衣服增加绒毛。 10. **拼贴剪纸风**：以可见的纸张边缘、纹理和重叠层次突出手工拼贴感，避免只生成普通扁平插画；人物各部件在动作中保持连贯。 11. **国风水墨风**：水墨笔触、留白与适度晕染，保留清楚的主体轮廓和人物特征。 12. **复古漫画风**：复古印刷配色、漫画线条与适度网点，表情和文字保持易读。 以上是可选方向，不限制用户自定义或混合画风。生成时锁定同一系列的脸、比例、配色、材质、线条、光影和识别性服饰；不得把用户选定的写实风、像素风等自动替换成 Q 版。像素风使用连续帧导出器时加 `--pixel-art` 避免平滑缩放；仍需检查生成图本身的像素网格是否清楚、一致。 ## 六个主题 根据用户选择制作指定系列；不要因为列有六个主题就自动生成全部系列。用户只说做表情包时，可从日常聊天或其明确场景开始。以下是选题方向，名字可替换为用户自己的名字，每个系列自行设计差异化动作、道具和文案： 1. **工作中**：工作群和协作，如“安排上了”“我看看”“马上处理”“辛苦啦”；可使用电脑、文件和咖啡。 2. **AI 搭子**：AI 工具与创作，如“让我问问 AI”“提示词来了”“灵感闪现”“又学会了”；可使用机器人、灯泡和屏幕。