This automates scraping Xiaohongshu posts for any keyword you throw at it, bypassing the platform's aggressive anti-bot defenses by spinning up a headed Playwright browser that looks like a real user. It grabs the top N posts with their text, images, and comments, downloads everything locally, then hands you a raw markdown file. The real value is in phase two: Claude reads that file, actually views the images with vision capabilities, and synthesizes everything into a proper analytical report instead of just dumping post lists at you. You'll need playwright-cli and Python with requests installed. Useful when you're doing market research or competitive analysis and need to pull together what Chinese users are actually saying and showing about a topic on Little Red Book.
npx skills add https://github.com/piekill/xiaohongshu-summarizer-skill --skill xiaohongshu-search-summarizer