ziguishian/xhs-visual-director-skill

★ 17Audience · writerComplexity · 1/5Setup · easy

Mindmap

mindmap
  root((xhs-visual-director-skill))
    What it does
      Analyze topic
      Pick visual style
      Plan 6-8 page post
    Outputs
      Style judgment report
      Image prompts per page
      Visual standards
      Publishing copy
    Style categories
      Dark tech magazine
      Minimalist business
      Notion card layouts
      Phone screenshot redesign
    Workflow
      Provide topic
      Review style report
      Pick direction
      Generate page prompts

mindmap root((xhs-visual-director-skill)) What it does Analyze topic Pick visual style Plan 6-8 page post Outputs Style judgment report Image prompts per page Visual standards Publishing copy Style categories Dark tech magazine Minimalist business Notion card layouts Phone screenshot redesign Workflow Provide topic Review style report Pick direction Generate page prompts

Click or tap to explore — scroll the page freely

Things people build with this

USE CASE 1

Turn a draft topic into a planned 6-8 page Xiaohongshu post with one image generation prompt per page.

USE CASE 2

Get a style recommendation from 24 preset visual styles before designing any graphics for a post.

USE CASE 3

Generate a complete publishing package including title, body text, hashtags, and pinned comment for a Xiaohongshu post.

Getting it running

Difficulty · easy Time to first run · 5min

In plain English

XHS Visual Director is an agent skill for Codex that helps plan visual content for Xiaohongshu, the Chinese social platform also known as RedNote or Little Red Book. Rather than writing captions or generic copy, it acts as a visual planner: it analyzes your topic, decides which visual style fits the content, and produces a full plan for a multi-page image post from cover to final card. Given a topic or draft, the skill outputs a style judgment report explaining which of its 24 built-in visual styles are appropriate and why, followed by three style options to choose from. Once a direction is chosen, it produces a 6 to 8 page structure with detailed visual specifications for each page, image generation prompts ready to paste into an AI image tool, unified visual standards covering color palette and typography, and then the publishing text including title, body, hashtags, and a pinned comment suggestion. The 24 preset styles include categories like dark tech magazine, minimalist business proposal, phone screenshot redesign, and Notion-style card layouts. Each style definition specifies which content types it suits, what it avoids, color palette, typography hierarchy, layout approach, and both positive and negative image generation prompts. The README describes a recommended workflow: provide a topic, review the style judgment report first before generating any pages, pick one of the three style options, generate page prompts, create images from those prompts, then review for readability and visual quality before generating publishing copy. The skill is written in Chinese and is designed for Xiaohongshu content creators, AI and tech content writers, personal brand builders, and people who want to convert drafts, screenshots, product images, or slide decks into polished social media posts.

Copy-paste prompts

Prompt 1

Use the XHS Visual Director skill to analyze this topic and give me a style judgment report with three visual direction options: [paste your topic]

Prompt 2

I picked the dark tech magazine style. Now generate detailed visual specifications and image generation prompts for all 8 pages of my Xiaohongshu post.

Prompt 3

Generate the publishing copy for my finished Xiaohongshu post: title, body text, five hashtags, and a pinned comment based on this topic: [paste topic]

Prompt 4

I want to convert my slide deck into a polished Xiaohongshu post. Use the visual director skill to pick the best style and plan the page structure.

Open on GitHub → Explain another repo

← ziguishian on gitmyhub — every repo by this author, as a profile.

Verify against the repo before relying on details.