baoyu-xhs-images Skill: Turn Content Into Xiaohongshu-Style Image Cards

Breaks an article or idea into 1 to 10 cartoon-style image cards for Xiaohongshu and similar social posts.

Installs: 31.6K; on skills.sh
Stars: 27.2K; on GitHub
Files: 28; besides SKILL.md
Updated: Nov 2026; last repo commit

baoyu-xhs-images makes your agent turn content into a series of image cards. It analyzes the content, recommends a strategy, style, layout and palette from 12 styles, 8 layouts and 3 palettes, asks you to confirm, then generates the cover first and uses it as the reference for the other cards so they match. It needs an image backend: Codex's imagegen, Cursor's GenerateImage, or the repo's baoyu-image-gen skill with a provider key or a Codex login.

Good for

Creators posting to Xiaohongshu and similar image-card feeds
Marketers who turn an article into a swipeable card series
Teachers who want checklists and study notes as cards
Claude Code, Codex and Cursor users who already run skills

Install baoyu-xhs-images

Claude Code

Copy this prompt.

Install the baoyu-xhs-images skill in Claude Code. Run this command:

npx skills add jimliu/baoyu-skills --skill baoyu-xhs-images --skill baoyu-image-gen -a claude-code

Then check that it loaded and tell me how to use it.

2
Paste it into Claude Code. It shows the change and asks before running it.

Requires Node.js.

baoyu-image-gen is the repo's own image backend. It needs a key for one image provider (listed under environment variables), or a logged-in Codex CLI. Codex and Cursor have their own image tools, which baoyu-xhs-images prefers when it runs there. The README asks you to install only the skills you need, because its plain command installs 20+.

From the vendor's docs. Source.

What's in the skill

When it loads

Generates infographic image card series with 12 visual styles, 8 layouts, and 3 color palettes. Breaks content into 1-10 cartoon-style image cards optimized for social media engagement. Use when user mentions "小红书图片", "小红书种草", "小绿书", "微信图文", "微信贴图", "image cards", "图片卡片", baoyu-xhs-images, or wants social media infographic series.

Sections in SKILL.md

User Input Tools
Image Generation Tools
Batch Generation Policy
Confirmation Policy
Language
Options
Dimensions
Styles (12)
Layouts (8)
Palettes (optional override)
Presets (style + layout shortcuts)
Auto-Selection
Style × Layout Matrix
Outline Strategies
Reference Images
File Layout

Key facts

Author: jimliu, also known as Baoyu (repo owner)
License: MIT
Installs on skills.sh: 31,557 (Nov 2, 2026)
GitHub stars: 27,159 (Nov 2, 2026, whole repo)
Last commit: Nov 1, 2026
What installs: 29 files including SKILL.md: reference files for styles, palettes, layouts and prompts; no scripts
Needs: An image backend; SKILL.md bundles no scripts
Options: 12 styles, 8 layouts, 3 palettes and 26 presets
Output: image-cards/<topic>/ with analysis.md, outline.md, prompts/ and numbered PNGs
skills.sh audits: Gen Agent Trust Hub: pass, Socket: pass, Snyk: pass

What it includes

Reads EXTEND.md, or runs first-time setup if there is none
Blocking when you run it interactively. With --yes it skips setup and uses defaults.
Write · Load preferences
Writes analysis.md and recommends a plan
Content type, audience, hook, a recommended image count of 2 to 10, and a strategy, style, layout and palette.
Write · Analyze content
Shows the plan; you accept it, customize it, or compare three outlines
Path A confirms the plan, Path B asks five questions, Path C writes three outline variants (A story-driven, B information-dense, C visual-first) to choose from.
Smart confirm (required)
Renders the cover, then the other cards using the cover as a reference
Saves every prompt as a file first. Runs batches of 4 by default (--batch-size 1 to 8) and retries a failed card once.
Write · Generate images
Lists the files and the choices used
Topic, mode, strategy, style, palette, layout and each card.
Completion report
Picks the look and structure
12 styles including cute (default), notion and chalkboard; 8 layouts from sparse (default) to quadrant; palettes macaron, warm and neon.
--style, --layout, --palette
Shortcut for a style plus layout
26 presets such as knowledge-card, checklist, tutorial and poster.
--preset
Supplies reference images
A reference can be passed directly, or mined for style or palette traits.
Read · --ref
Skips every confirmation
For scheduled or unattended runs. It states the assumed choices before generating.
--yes
Adds a watermark to each card
Set watermark.enabled and watermark.content in EXTEND.md; it is added to the generation prompt.
Watermark
Edit, add or delete cards
SKILL.md says to update the prompt file first, then regenerate. It does not repair text on a finished image with code.
Write · Image modification

From the official docs

Permissions

Medium risk

It sends your content to an image provider and writes files into your project. Try one short series first and keep API keys out of git.

What it can do

Read your data
Create and change things
Delete things
Read-only mode

Watch out for

Your content goes to an image provider
Each card prompt is built from your source and sent to the backend you pick. Do not feed it text you cannot share with that provider. github.com
The image backend runs scripts
baoyu-image-gen, installed with it, runs scripts/main.ts with bun. Its SKILL.md says to use npx -y bun if bun is missing, or to install bun with brew. Install bun yourself if you want to control that. github.com
--yes skips every confirmation
With --yes it generates straight away using defaults or your EXTEND.md. Use it only for runs you have already tried by hand. github.com
It writes files in your project
Analysis, outline, prompts and PNGs go in image-cards/<topic>/. Existing files are renamed with a backup suffix. The first run also saves EXTEND.md. github.com
Image generation can cost money
It renders one image per card, 4 at a time by default (--batch-size 1 to 8). Provider billing is yours. Try a small series first. github.com

Technical details

Talks to	Your image provider receives each card prompt, built from your content. npx -y bun downloads bun if it is not installed (baoyu-image-gen)
Sign-in and keys	Provider keys for baoyu-image-gen live in a .env file: ~/.baoyu-skills/.env or <project>/.baoyu-skills/.env. The README tells you not to commit project-level .env files.
Scopes	Your project folder: it creates image-cards/<topic>/ and renames existing files to -backup- before overwriting An EXTEND.md preferences file A .env file with your image provider key

Questions

What does baoyu-xhs-images do?

It breaks content into a series of 1 to 10 cartoon-style image cards, built for Xiaohongshu and similar social feeds. You pick or accept a style, layout and palette, and it renders the cards through an image backend.

How do I install baoyu-xhs-images in Claude Code?

Use the install box above. It adds baoyu-xhs-images and baoyu-image-gen together, because Claude Code has no built-in image tool. Then set one provider key for baoyu-image-gen, listed under environment variables.

Does it need an API key?

Usually. The skill needs some image backend. Codex and Cursor have their own, and elsewhere the repo's baoyu-image-gen takes a provider key or a logged-in Codex CLI, which needs no OPENAI_API_KEY per its SKILL.md.

Does baoyu-xhs-images work in the Claude app?

Not documented. The skill writes files and calls an image backend from a terminal agent, and the README does not mention the Claude app. Use Claude Code, Codex or Cursor.

Is baoyu-xhs-images safe?

It bundles no scripts, but the baoyu-image-gen skill installed with it runs scripts through bun (or npx -y bun). It sends your content to the image provider you choose and writes a folder of files in your project. The first interactive run also saves an EXTEND.md preferences file. Keep API keys in a .env file you do not commit.

Can it write text in Chinese?

Yes. The skill is built for Xiaohongshu, responds in your language and writes each prompt in your preferred language. How well the text renders depends on your image backend, and the skill regenerates rather than repainting text.

How is it different from baoyu-slide-deck?

baoyu-xhs-images makes 1 to 10 cartoon-style cards for social posts. baoyu-slide-deck builds an ordered slide deck with an outline and merges it into PPTX and PDF. Both live in the same repo.

Information

updated Nov 2, 2026

Repository: jimliu/baoyu-skills
Folder: skills/baoyu-xhs-images

jimliu/baoyu-skills skills.sh

Sources

Docs, issue trackers and discussions this page is built from.

Show all 5

baoyu-skills README · github.com
skills CLI commands · github.com
baoyu-xhs-images SKILL.md · github.com
baoyu-image-gen SKILL.md · github.com
codex-imagegen.md · github.com