Jim Liu's baoyu-skills Turns Claude Code Into a Full Content Pipeline
A single-command install pulls 20+ Claude Code skills that turn markdown files into infographics, slide decks, comics, and cross-platform social posts.
Takeaways
- baoyu-skills bundles 20+ Claude Code and Codex skills for content generation and publishing.
- Installs in one command:
npx skills add jimliu/baoyu-skills, or via Claude plugin marketplace. - Infographic skill offers 21 layouts x 21 visual styles for publication-ready output.
- Diagram skill hand-writes SVG with auto dark-mode, no image model calls needed.
- Image-gen wraps OpenAI, Google, Azure, OpenRouter, Replicate, MiniMax, DashScope, Jimeng, Seedream.
- Publishing skills target X, Weibo, WeChat Official Accounts with API and browser fallbacks.
A collection of Claude Code skills called baoyu-skills has quietly become one of the most starred agent-skill repos on GitHub, sitting near 19.7k stars. Maintained by Jim Liu (Baoyu), it bundles more than 20 skills that turn a Claude Code or Codex agent into a full content production pipeline: infographics, cover images, slide decks, comics, diagrams, and one-click publishing to X, Weibo, and WeChat.
What actually ships
The repo is organized as a plugin marketplace with a single bundled plugin, so each skill is registered exactly once. Skills fall into three buckets: content generation, AI generation backends, and utilities. A few highlights from the catalog:
- baoyu-infographic generates professional infographics with 21 layout types and 21 visual styles, from fishbone diagrams to Eisenhower quadrants, rendered in styles like technical-schematic, claymation, or corporate-memphis.
- baoyu-diagram writes real SVG code by hand rather than calling an image model. Claude computes layout math itself so every diagram honors the design system, and output includes a
@media (prefers-color-scheme: dark)block so the same file renders in both modes.