baoyu-comic Skill for Claude: Turn a Topic or Article into a Knowledge Comic

Makes your assistant turn a topic or article into an illustrated knowledge comic: storyboard, page images and a PDF.

Installs: 30.1K; on skills.sh
Stars: 27.2K; on GitHub
Files: 38; besides SKILL.md
Updated: Nov 2026; last repo commit

baoyu-comic has your assistant plan and draw an original educational comic. It analyzes your source, asks you to confirm art style, tone and layout, writes a storyboard and character sheet, saves one prompt file per page, generates the page images and merges them into a PDF. You choose from 6 art styles, 7 tones, 7 layouts and 5 presets, and can stop after the storyboard or the prompts.

It triggers on requests such as knowledge comics, educational comics, biography comics, tutorial comics or Logicomix-style comics. It needs an image generator. In Claude Code that means the separate baoyu-image-gen skill with a provider API key. The folder holds 39 files including SKILL.md and one script, merge-to-pdf.ts.

Good for

Teachers and trainers who explain ideas as comics
Writers turning a biography or article into a visual story
Creators who want manga, ink-brush or ligne-claire pages from one prompt

Install baoyu-comic

Claude Code

Copy this prompt.

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

npx skills add https://github.com/jimliu/baoyu-skills --skill baoyu-comic --skill baoyu-image-gen

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-comic needs an image generator. In Claude Code, install baoyu-image-gen with it and set a provider API key; the README lists OpenAI, Google, OpenRouter, DashScope, MiniMax, Replicate and others. SKILL.md uses Codex's built-in imagegen in Codex and Cursor's GenerateImage in Cursor, when they exist. The README also says the scripts need Node.js and the ability to run npx bun. The plugin adds all 21 skills from the repo, including posting skills; the manifest lists no hooks. The README tip is to install only the skills you need.

Built from the package data we hold. Not yet tested by us.

What's in the skill

When it loads

Knowledge comic creator supporting multiple art styles and tones. Creates original educational comics with detailed panel layouts and batch-capable image generation. Use when user asks to create "知识漫画", "教育漫画", "biography comic", "tutorial comic", or "Logicomix-style comic".

Sections in SKILL.md

User Input Tools
Image Generation Tools
Batch Generation Policy
Reference Images
Options
Visual Dimensions
Partial Workflow Options
Art, Tone & Preset Catalogue
Script Directory
File Structure
Language Handling
Workflow
Progress Checklist
Flow
Step Summary
Step 7: Image Generation

Key facts

Author: Jim Liu (jimliu/baoyu-skills)
Repo: jimliu/baoyu-skills
License: MIT (repo LICENSE file)
Installs on skills.sh: 30,103 (Nov 2, 2026)
GitHub stars (whole repo): 27,159 (Nov 2, 2026)
Last repo commit: Nov 1, 2026
Skill folder: 39 files including SKILL.md: 37 reference notes and 1 script (scripts/merge-to-pdf.ts)
Skill version in SKILL.md: 1.117.4
Needs: bun or npx to run the script; an image generator (baoyu-image-gen, Codex imagegen or Cursor GenerateImage)
skills.sh audits: Gen Agent Trust Hub warn, Socket pass, Snyk pass

What it includes

Reads EXTEND.md; if missing, first-time setup asks for your defaults
Reads EXTEND.md; if missing, first-time setup asks watermark, art, tone, language and save place
Files · Step 1.1: load preferences
Saves your source, writes analysis.md with audience, themes, figures and a page count
Files · Step 1.2: analyze the source
If comic/<slug> exists, asks whether to redo storyboard, images, back up or exit
If comic/<slug> exists, asks: regenerate storyboard, regenerate images, back up and redo, or exit
Files · Step 1.3: check existing comic
Asks style, narrative focus, audience, and whether to review the outline and prompts
Step 2: confirm style and options
Writes storyboard.md and characters/characters.md in your language
Files · Step 3: storyboard and characters
Optional pause for you to approve the storyboard and characters
Step 4: review outline
Saves one prompt file per page under prompts/ before any image is made
Files · Step 5: write prompts
Optional pause for you to approve the prompts
Step 6: review prompts
Generates characters/characters.png for multi-page comics with recurring characters
Files · Step 7.1: character sheet
Makes each page image, in batches of 4 by default, and retries a failure once
Files · Step 7.2: generate pages
Runs scripts/merge-to-pdf.ts through bun or npx -y bun to build the PDF
Runs code · Step 8: merge to PDF
Summarizes what was made
Step 9: completion report
Your request, saved preference, then Codex imagegen, codex exec, Cursor, others
Your request, then saved preference, then Codex imagegen, codex exec, Cursor GenerateImage, others
Network · Image backend order
Never swaps in SVG or HTML art; fixes bad text by regenerating the page
Never substitutes SVG or HTML for pictures; fixes bad text by regenerating, not painting over
No code-drawn images or text patches
--ref files used as direct references, style traits or a palette
Files · Reference images
--art, --tone, --layout, --aspect, --lang, --ref, --batch-size
Options
ohmsha, wuxia, shoujo, concept-story, four-panel add rules beyond art plus tone
Presets
--storyboard-only, --prompts-only, --images-only and --regenerate N
Partial workflows
Edit the prompt file first; adds and deletes renumber files and update the storyboard
Edit the prompt file first; adding or deleting pages renumbers files and updates the storyboard
Delete · Page changes
Uses --lang, then EXTEND.md, then your language, then the source language
Language handling
Edit EXTEND.md, or delete it to rerun setup
Edit EXTEND.md, or delete it to rerun setup; keys include backend, batch size and watermark
Files · Changing preferences

From the official docs

Permissions

Medium risk

This skill sends prompts and reference images to an image provider and writes a comic folder in your project. Use an API key with a spend limit and check the output folder.

What it can do

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

Watch out for

Your prompts go to an image provider
Each page is an image request to the backend you pick, with your reference images. baoyu-image-gen takes provider API keys (README lists OpenAI, Google, OpenRouter and others), so use a key you are happy to spend and check your provider's billing. github.com
It writes files and runs a script
It creates comic/<slug>/ with prompts, images and a PDF, and EXTEND.md under .baoyu-skills. Existing files are renamed with a -backup suffix before regenerating. The PDF script runs through bun or npx -y bun. github.com
It may run extra commands and other tools
workflow.md says to compress the character sheet with sips (macOS) or pngquant before using it as a reference. If neither Codex imagegen nor Cursor is present, SKILL.md can route image requests through the codex CLI using your codex login (references/codex-imagegen.md). github.com
Sensitive public figures
SKILL.md tells the assistant to use stylized alternatives for sensitive public figures. github.com

Technical details

Talks to	The image provider you choose, for every page image and reference image. bun download via npx -y bun if bun is missing
Scopes	Your working folder (comic/<slug>/) Your home folder for saved preferences (.baoyu-skills)

Questions

What does baoyu-comic do?

It turns a topic, article or story into an original educational comic. It writes a storyboard, makes the page images and merges them into a PDF. You pick art style, tone and layout.

How do I install baoyu-comic?

Use the install box at the top of this page. In Claude Code the box installs baoyu-comic with baoyu-image-gen, the image generator it needs there. The README warns that the repo has 20+ skills and says to install only the ones you use.

Do I need an API key?

In Claude Code, yes, for the image generator. baoyu-image-gen reads provider keys such as OPENAI_API_KEY or GOOGLE_API_KEY. In Codex, SKILL.md uses Codex's built-in imagegen, and in Cursor its GenerateImage tool.

How many pages does a comic have?

The workflow suggests 5 to 8 pages for a short story, 9 to 15 for medium complexity and 16 to 25 for a full biography. SKILL.md gives no time estimate; each page is a separate image request.

Is baoyu-comic safe to install?

It bundles one script that merges images into a PDF, and it sends prompts to the image provider you pick. It writes files in your project and under .baoyu-skills. Use a spend-limited API key.

Does it work in the Claude app?

The README documents no Claude app route for it. It needs a shell to run bun and an image generator, and the install box shows the Claude Code and Codex routes.

Can I make a single page or only a storyboard?

Yes. The options include --storyboard-only, --prompts-only, --images-only and --regenerate N for chosen pages.

Information

updated Nov 2, 2026

Repository: jimliu/baoyu-skills
Folder: skills/baoyu-comic

jimliu/baoyu-skills skills.sh

Sources

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

Show all 8

baoyu-skills README · github.com
baoyu-comic SKILL.md · github.com
baoyu-comic on skills.sh · skills.sh
baoyu-comic workflow.md · github.com
baoyu-comic first-time-setup.md · github.com
baoyu-image-gen SKILL.md · github.com
baoyu-comic merge-to-pdf.ts · github.com
baoyu-comic codex-imagegen.md · github.com