I just came across an article analyzing Xiaohongshu's RedSkill, and I strongly agree with one point.

In the past, various AI skills were mostly buried in technical communities, barely accessible to ordinary users. Even if they stumbled upon them, they couldn't understand or use them—the barrier was extremely high.
Now, Xiaohongshu directly attaches RedSkill below posts. Users can first see the blogger's actual operation and usage effects, and if they find it suitable, they can claim it. This greatly lowers the cost of getting started.
The platform's recent update pace has also been intense. On July 3, it fully opened up the Skill upload format, no longer limited to simple text; various code and composite tools can now be published normally. This Wednesday, it will also launch the vibecoding embedded interactive feature within the site, at which point there will be no need to jump to an external Agent—half-screen debugging and full-screen interaction can be done directly within the post component.
It's clear that Xiaohongshu is not just testing features casually—it is solidly building an AI skill distribution ecosystem for ordinary users. Compared to the "passive search" in pure tech communities, this "grass-planting" model (see the effect first, then try it) is likely to become the mainstream way for the public to access AI tools in the future.
#rednote #skill #redskill