花叔

You can generate WeChat and Xiaohongshu cover images, write Douyin and X platform posts, proofread Chinese articles to reduce AI-detection, convert Markdown to PDF/DOCX/HTML, build PPTX decks from outlines, analyze Excel data into reports, and run structured web research saved to Markdown files.

Chinese-language content creators, self-media operators, video scriptwriters, and knowledge workers who produce articles, slides, and social posts. Founders and analysts also benefit from the 14 perspective skills that apply mental models from figures like Karpathy, Naval, Munger, and Feynman to decisions.

Each perspective skill encodes a public figure's mental models, decision heuristics, and expression style distilled from primary sources. Invoke one when you want a problem analyzed through that lens—for example, Feynman's cargo-cult detection and demonstration-first reasoning—rather than for generic explanation requests.

Several skills call the Gemini API for image generation and competitor video analysis, upload illustrations to ImgBB for permanent Markdown links, and read Weread shelves or personal archives of 1800+ posts. The agent-swarm skill coordinates parallel Claude agents via git worktrees and lock files.

Skills ship as runnable SKILL.md directories, which huashu-nuwa can distill from fuzzy requests. The huashu-skill-updater skill checks installed huashu-family skills against their remote GitHub repositories and reports available updates, keeping local installations synchronized.