Xiaohongshu's recent AI labeling rules, requiring synthetic content to be labeled, have been bypassed by an open-source project named guizang-social-card-skill. This tool generates Xiaohongshu and WeChat posts using HTML and CSS, avoiding AI models for image generation. Instead, it uses real-photo libraries like Unsplash, rendering layouts as browser screenshots, thus evading detection by Xiaohongshu's audio-visual recognition model. The tool, developed by Guizang, offers 28 layout skeletons and uses AI for layout logic, not illustration. It targets specific formats like Xiaohongshu 3:4 graphics and WeChat covers, with resolutions of 1080×1440, 1080×1080, and 2100×900. The approach leverages HTML rendering to produce images that resemble natural photos, potentially avoiding classification as AI-generated content. However, the effectiveness of this method depends on Xiaohongshu's evolving definition of "AI-generated synthetic content."