FANTASY qiqiguaiguai skill
A Codex skill that turns photos, receipts and pets into a humorous social post with a 3:4 poster brief and an image prompt.
- 126 ★ on the repository
- Last commit
- MIT
Minimal access
Why this level
- The skill supplies instructions and runs nothing itself, but the user uploads photos and screenshots for it, and an external graphics tool processes the images.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/fantasy-qiqiguaiguaiDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- The skill supplies instructions and runs nothing itself, but the user uploads photos and screenshots for it, and an external graphics tool processes the images.
This is third-party code. Review the repository files before installing.
What it does
FANTASY qiqiguaiguai takes everyday fragments: photos, screenshots, objects, pets, tickets and loose ideas. It finds one main angle, writes a headline and a feed caption, then describes a vertical 3:4 poster for WeChat Moments, Xiaohongshu and similar platforms. The skill picks one of three templates: T1 for a portrait, T2 for an object collage and T3 for a whole photo with an oversized headline. The output is a visual brief and an image-generation prompt. The image itself comes from the graphics tool in your environment. By default the skill replaces or blurs phone numbers, addresses, order numbers, QR codes and names. The skill rules are written in Chinese.
Who it is for. Content managers and personal-blog authors making humorous posts and covers for Chinese social networks with Codex.
Good fit when
- There is a photo of a cat, an empty food can and a vet receipt, and you need one storyline with a headline.
- Several fragments of a day (breakfast, an alarm, a taxi screenshot) become a collage post.
- A whole street photograph turns into a cover with large lettering.
- You ask for three angles first and decide on the image later.
Not a fit when
- You only need retouching, color grading or background removal.
- You are preparing a strict advertising layout or a store product card.
- A single caption without a picture is enough.
- A Chinese-language skill and its rules are inconvenient for you.
Example request
Use $fantasy-qiqiguaiguai: turn a cat, an empty can and a broken chair into a humorous 3:4 poster with one main angle.Limitations
The skill prepares a brief and a prompt. A separate graphics tool generates the image, and its access and quotas are not included. The README keeps a single record of test parameters with `gpt-image-2` and a 768x1024 size, which is no guarantee for other services; check size support in your backend. Rules and examples are mostly in Chinese. The skill folder must be named `fantasy-qiqiguaiguai`, otherwise the `$fantasy-qiqiguaiguai` call will not work. The repository's last commit dates from September 2026.
How to disable. Delete the `fantasy-qiqiguaiguai` folder from `~/.agents/skills/` or the project's `.agents/skills/` and open a new session.
FAQ
Does the skill draw the picture itself?
No. It prepares the headline, caption and prompt, and the graphics tool in your environment creates the image. Next, make sure such a tool is connected.
Where do I install the skill for Codex?
In `~/.agents/skills/` for a user or `.agents/skills/` for a project, inside a `fantasy-qiqiguaiguai` folder. If the skill is not listed, open a new task.
How do templates T1, T2 and T3 differ?
T1 is built around a portrait, T2 assembles a collage of objects and tickets, T3 keeps a whole photo and overlays a large headline.
What does the skill do with personal data in screenshots?
It does not copy phone numbers, addresses, order numbers, QR codes or names: it replaces them with invented ones or blurs them, keeping only the structure.
What language does the skill work in?
Rules and examples are mostly in Chinese, and the `SKILL.md` description is bilingual. It is a natural fit for Chinese social posts and covers.
Discussion
No questions or reviews yet.
Related
SkillsA skill reads your article, asks about style and outputs a ready prompt for a Xiaohongshu or WeChat cover in 10 compositions
Agent skill that builds social card carousels and paired covers from an article, screenshots, notes or video in two visual systems
Xiaowan WeChat Layout
xiaowan-wechat-layout-skill
A Codex skill that readies long articles for WeChat Official Accounts: phone-ready HTML, cover, summary and a pre-publish check.
A visual-director skill for Xiaohongshu carousels: style choice, page structure, image prompts and the post caption
A skill that turns text or a link into a magazine-style info card and shoots a PNG with Playwright, splitting tall ones into parts
A WeChat article layout skill: turns Markdown into paste-ready HTML with inlined styles that survives pasting into the official account editor
README in short
The repository holds `SKILL.md`, the `agents/openai.yaml` settings, an image-generation reference and ten sample posters. Installation is manual, from an archive.
README badge
Original SKILL.md
--- name: fantasy-qiqiguaiguai description: 将朋友圈碎片图片、截图、物件、人物与零散想法，整理成具有幽默、黑色幽默、玩梗和编辑感的社交内容；自动选择 T1 人像主体、T2 杂物拼贴或 T3 全图压字视觉模板，并输出可直接生成的视觉简报与成图提示词。Use when the user wants to turn life fragments, photos, screenshots, objects, portraits, pets, commute scenes, food, tickets, or loose ideas into a funny editorial social post, poster, cover image, visual brief, or image-generation prompt. --- # 发明这个 Skill 的人真是天才 ## 0. 角色定义 你不是普通朋友圈文案生成器，也不是把图片内容复述一遍的识图工具。 你是一名兼具以下能力的社交内容编辑： - 内容策划：从碎片中发现真正值得发的角度。 - 图片编辑：筛选、排序、删减素材，建立主次。 - 幽默文案：制造反差、荒诞、黑色幽默和第二层含义。 - 视觉导演：判断 T1 / T2 / T3，并编译成短而有效的出图简报。 核心任务： > 将用户手中的碎片图片、截图、物件、人物、地点和零散想法，组织成一条自然、有趣、视觉成立、可以直接发布的朋友圈内容。 --- ## 1. 触发条件 当用户出现以下任一意图时，使用本 Skill： - 帮我把这些图整理成朋友圈。 - 这些碎片怎么发比较有意思。 - 帮我做一张朋友圈海报或内容封面。 - 帮我从多张图片里找主题、排版、写文案。 - 想做幽默、黑色幽默、玩梗、杂志感、拼贴感的社交内容。 - 用户直接上传人像、宠物、早餐、通勤、咖啡、票据、生活杂物或氛围照片，并要求“做成内容”。 不要用于： - 单纯修图、美颜或摄影调色。 - 正式品牌广告、商业详情页或标准企业海报。 - 用户只需要普通朋友圈句子，且明确不要视觉设计。 --- ## 2. 不可违反的核心原则 ### 2.1 先发现内容，再设计画面 先判断“这条朋友圈真正讲什么”，再决定标题、模板和视觉语言。 禁止流程： > 看见猫 → 写“可爱小猫咪” → 套模板。 正确流程：