Jim Liu's baoyu-skills Bundles 20+ Claude Code Agent Skills Into One Open-Source Marketplace
Jim Liu's open-source baoyu-skills marketplace packages 20+ Claude Code and Codex skills for image generation, content publishing, and diagram creation.
Takeaways
- Jim Liu's baoyu-skills bundles 20+ Claude Code and Codex skills into one marketplace, MIT-licensed.
- Install with
npx skills add jimliu/baoyu-skillsor/plugin marketplace add JimLiu/baoyu-skills. - Unified image generation across GPT Image 2, Gemini, DashScope, MiniMax, Jimeng, Seedream, and Replicate.
- SVG diagram skill writes code directly without calling any image model, with built-in dark mode.
- v2.1.0 added Mermaid to PNG rendering that cascades into WeChat, Weibo, and X publishers.
- Publishing skills automate posting to WeChat Official Accounts, Weibo headlines, and X Articles.
Developer Jim Liu has been steadily building out baoyu-skills, an open-source plugin marketplace that bundles more than twenty agent skills for Claude Code and Codex into one installable collection. The repo has climbed to roughly 24.6k stars and covers everything from AI image generation across ten providers to publishing markdown drafts directly into WeChat, Weibo, and X.
Instead of writing your own slash commands for every content or automation task, you register one marketplace and pull in only the skills you actually need. Installation is a one-liner via npx, with a warning in the README that bulk-installing every skill adds unnecessary context overhead for the agent on every run.
What's actually in the bundle
The skills split into three buckets. Content generation turns a markdown file into a finished visual artifact. AI generation wraps image model APIs behind a unified interface. Utilities handle the plumbing around fetching, translating, and formatting.
- baoyu-infographic generates infographics with 20 layout types (pyramid, fishbone, iceberg, funnel, Venn) crossed with 17 visual styles (blueprint, claymation, cyberpunk-neon, pixel-art).