ziguishian

这是一个用于规划小红书图文的 Agent Skill。它不是普通文案助手,而是一个“视觉导演”:先判断内容任务,再选择适合的视觉风格,最后输出完整图文结构、逐页视觉方案、图像生成提示词、发布文案和自检清单。

17
4
85% credibility
Found May 31, 2026 at 17 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
AI Summary

This is a creative assistant skill that helps everyday content creators design professional-looking posts for Xiaohongshu, a popular Chinese social media platform. Think of it as having a personal art director who first understands what you want to say, then suggests the perfect visual style, plans out your entire carousel (6-8 slides), creates detailed image generation prompts, and even writes your captions and hashtags. The skill specifically aims to avoid the 'cheap AI look' and instead produces content that feels premium and native to the platform. It's designed for creators who have ideas but lack visual design skills, helping them transform rough concepts into polished, engaging posts that stand out in the feed.

How It Works

1
💡 You have an idea for a post

You come up with a topic you want to share on Xiaohongshu but feel stuck on how to make it look professional.

2
🤖 You share your idea with your AI assistant

You tell your AI helper about your topic, like 'I want to explain why everyone should learn Vibe Coding' and mention you want it to look高级 (high-end) and modern.

3
🎨 Your assistant becomes a visual director

Instead of just writing text, your helper analyzes your topic and suggests three different visual directions you could take, explaining why each works for your specific content.

4
You choose your visual direction
🖥️
Tech-forward style

Dark backgrounds with glowing accents for AI and tech topics

💼
Business professional

Clean, premium look for commercial and industry content

📱
Phone-native style

Screenshots and dashboard layouts for tutorials and tools

5
📋 You get a complete content plan

Your helper creates a detailed 6-8 page structure with specific instructions for each slide, including where to place text, what colors to use, and what the overall layout should feel like.

6
🖼️ You generate professional images

For each page, you receive ready-to-use prompts that you can paste into image generators to create cohesive, high-quality visuals without the cheap AI look.

🚀 You publish your viral-ready post

Your assistant also gives you the perfect title, body text, hashtags, and even a pinned comment to drive engagement. Your Xiaohongshu carousel is ready to impress.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 17 to 17 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is xhs-visual-director-skill?

This is an Agent Skill for Codex that transforms topics into polished Xiaohongshu carousel posts. Instead of dumping out generic copy, it acts as a visual director: first analyzing your content goal, then selecting an appropriate visual style from 24 built-in options, and finally outputting complete page structures, per-page image prompts, unified visual specs, and a self-review checklist. The workflow produces 6-8 slide carousels optimized for the 3:4 mobile format with titles, body text, hashtags, and comment prompts ready to publish.

Why is it gaining traction?

The real pain this solves is the "cheap AI look" problem. Most creators using AI image generators end up with purple gradients, liquid glass effects, and illegible text that tanks engagement. This skill explicitly avoids those anti-patterns and teaches the Agent to prefer dark tech magazine aesthetics, high-contrast typography, and phone-readable layouts. The three-step workflow (style judgment, page planning, image prompts) gives creators a repeatable process rather than one-off outputs. For developers building agent skills or exploring Claude/GitHub Copilot agent workflows, this demonstrates how to structure multi-step reasoning with clear decision gates.

Who should use this?

Content creators posting tech, AI, or business topics on Xiaohongshu who are tired of generic carousel posts. Product managers and designers who need to turn raw notes or screenshots into shareable visual content. Developers exploring agent skills on GitHub who want a concrete example of a production-ready skill with style libraries and review checklists.

Verdict

A niche but well-scoped tool for a specific platform workflow. The 0.85% credibility score and 17 stars reflect early-stage status with limited community validation. Documentation is thorough and the style system is extensible, but there's no test coverage or active maintenance signals. Worth trying if you create Xiaohongshu content and want consistent visual quality, but don't bet production workflows on it yet.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.