baoyu-xhs-images

baoyu-xhs-images is one Skill inside Baoyu Skills, the collection by Baoyu (GitHub user Jim Liu), and it makes image cards for Xiaohongshu (Rednote). The site already lists the whole collection; this entry covers only the Xiaohongshu image part. Give it text, local files or reference images. It analyzes the material, writes an outline and a prompt for each image, then generates 1 to 10 cards. You can pick from 12 visual styles, 8 layouts and 3 color schemes.

It keeps every intermediate file. The source, analysis, outline, prompts and final PNGs all land in an image-cards folder, so if one card looks wrong you can go back to that step and fix it. Images can come from the agent's built-in tools, Codex image generation, or an image API you connect yourself. It only makes images. It doesn't log in to Xiaohongshu or post anything.

Who it's for:People who post image notes on Xiaohongshu and want to turn long text into a set of info cards.

Before you start: Install with npx skills add jimliu/baoyu-skills, which installs the whole collection; Codex users can link just this subfolder into .agents/skills. You need Node.js plus npx or bun. The README names Claude Code and Codex and also mentions Cursor, Hermes and OpenClaw. It reads the text, local files and reference images you provide, plus an EXTEND.md preferences file. Prompts and reference images go to whichever image service you choose; with baoyu-image-gen, API keys are read from environment variables or a .baoyu-skills/.env file. It doesn't log in to any account, doesn't install hooks or scheduled jobs, and isn't installed with curl piped to bash. Image costs depend on your provider. Check the platform's rules before you post. MIT license.

Based on the project's public documentation. Not tested by AI Inspiration World.