SoloMCN update lets Claude Code distill past work into reusable skills

On October 16, 2026, a V2EX user posted two updates about SoloMCN, an open-source project described as a "one-person MCN." Rather than building a fixed workflow, the project lets Claude Code handle the process end to end, from finding trending topics to publishing on Douyin. The author said the past week focused on one goal: making Claude retain what it has already done so it can repeat the work next time and improve over time.

The demonstration was an animated short in which an AI narrates how it learned to speak, moving through cave paintings, Egyptian murals, ink-wash scrolls, Van Gogh, pop comics, and Ghibli styles. Research, storyboarding, voiceover, and final output were all handled by Claude in the background. According to the author, just days earlier the project could not make animations at all. The visual style came from huashu-art-motion, an open-source Claude Code skill released by Huashu, which the author called a strong skill.

The post contrasts the usual way of using someone else's skill: you install it yourself, ask Claude in a conversation to make a video with it, and once the video is done, the know-how of connecting it to your own voiceover, subtitles, series style, and publishing pipeline stays scattered across chat logs and temporary scripts. The next attempt starts from scratch, and as more skills pile up, no one maintains them. In SoloMCN, the author said, adding such a skill to the control console turns it into a reusable practice: Claude distills the research, storyboarding, voiceover, and publishing steps it has performed into a fixed approach for next time. The project is at https://github.com/ymybxx/SoloMCN. The second post noted that the first had been mistakenly placed in a different V2EX section and was reposted. Both reports describe the same update, and no further plans were mentioned.