/platform-algorithm
Use when understanding Xiaohongshu's recommendation algorithm, optimizing content reach, or troubleshooting low views
$ npx -y skills add vivy-yi/xiaohongshu-skills --skill platform-algorithm --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
- Fires itselfAuto-invocation. Claude auto-loads it when your prompt matches the work.Auto-invocation is when the right skill fires by itself at the right moment, driven by a FLOW.md router and a hook, instead of you invoking it by name. It is the difference between a skill being installed and a skill actually getting used.Read the full definition →
- You can call itInvoke it directly when you want it.
- Slash command
/platform-algorithm
Context preview
The summary Claude sees to decide when to auto-load this skill.
Use when understanding Xiaohongshu's recommendation algorithm, optimizing content reach, or troubleshooting low views
SKILL.md
platform-algorithm.SKILL.mdname: platform-algorithm
description: Use when understanding Xiaohongshu's recommendation algorithm, optimizing content reach, or troubleshooting low views
Platform Algorithm (平台算法深度)
Overview
Xiaohongshu's recommendation algorithm determines which posts appear on users' explore pages, who sees your content, and how widely your posts spread. Understanding the algorithm isn't about gaming the system—it's about creating content that serves the platform's goal: showing users content they'll find valuable and engaging. The core principle: the algorithm rewards content that generates engagement, keeps users on the platform, and builds community. It evaluates posts based on hundreds of signals: engagement metrics (likes, comments, saves, shares), content quality (originality, completeness, aesthetics), user relationships (who you interact with), and timeliness (recency, trending topics). The algorithm constantly evolves, but fundamental principles remain consistent: value to users, authenticity, engagement, and consistency. This guide provides deep understanding of how the algorithm works, practical strategies to optimize your content for algorithm favor, and troubleshooting for when your reach drops. While Xiaohongshu doesn't publicly disclose algorithm specifics, successful creators reverse-engineer what works through testing and observation. This guide compiles those learnings into actionable strategies.
**Key insight**: Top creators don't just "be authentic"—they optimize strategically for the algorithm while maintaining authenticity. They post when their audience is most active (not random times), use formats that generate saves (carousels over single images), encourage engagement (asking questions, ending with CTA), and build relationships (engaging with their community). These aren't manipulations—they're genuine strategies that help content reach the people who will value it most. Understanding the algorithm removes randomness from growth. Instead of posting and hoping for reach, you can create content with algorithmic principles in mind, dramatically increasing your chances of being featured on the explore page and reaching new audiences. The algorithm is your friend, not your enemy: it amplifies content that serves Xiaohongshu users. Focus on serving users, and the algorithm will serve you.
When to Use
**Use when**:
- Experiencing unexplained reach drops
- Launching new content and want to optimize for algorithm favor
- Troubleshooting why specific posts underperformed
- Analyzing competitors who consistently get explore page features
- Planning content strategy with algorithm in mind
- Wanting to understand platform changes and updates
**Do NOT use when**:
- Trying to "hack" or manipulate the system (backfires)
- Focusing on algorithm over audience value (prioritize audience first)
- Analyzing constantly instead of creating (analysis paralysis)
Core Pattern
**Before** (algorithm-ignorant): ❌ "Post randomly, wonder why some posts succeed and others fail" ❌ "No understanding of reach patterns (appears random)" ❌ "Can't recover from reach drops (don't know what went wrong)" ❌ "Growth feels random, unpredictable"
**After** (algorithm-aware): ✅ "Create content with algorithmic principles (engagement, saves, consistency)" ✅ "Understand reach patterns (know why posts perform as they do)" ✅ "Troubleshoot effectively (diagnose and fix reach issues)" ✅ "Growth feels systematic, predictable"
**Algorithm Fundamentals**:
| Factor | Weight | What It Measures | Optimization | |---------|--------|-----------------|----------------| | **Engagement** | High | Likes, comments, saves, shares, time spent | Create engaging content, ask questions, encourage saves | | **Content quality** | High | Originality, completeness, aesthetics | Invest in production value, unique insights, good visuals | | **User relationships** | Medium | Who views, interacts, shares your content | Engage with community, respond to comments, build connections | | **Timeliness** | Medium | Recency, trending topics, posting frequency | Post consistently, jump on trends, stay relevant | | **Account authority** | Medium | Niche expertise, follower trust | Demonstrate expertise, build credibility over time | | **Session quality** | Low-Medium | User engagement patterns (who they engage with after your post) | Create content that keeps users on platform |
Quick Reference
**Key Algorithm Signals**:
| Signal | What Works | What Doesn't Work | |--------|------------|-------------------| | **Initial engagement** | First 30-60 minutes critical | Posting when audience asleep | | **Saves** | Reference-worthy content (tutorials, guides) | Pure entertainment (fun but not save-worthy) | | **Comments** | Conversation starters, questions | Generic content, no engagement hook | | **Shares** | Relatable, valuable content | Self-promotional, boring | | **Time spent** | Long-form content (carousels, long videos) | Short, superficial content | | **Completion rate** | Content watched fully | Click-bait, disappointing content | | **Return viewers** | Consistent posting schedule | Sporadic posting, long gaps |
**Algorithm-Friendly Content Types**:
| Format | Algorithm Appeal | Save Rate | Best For | |--------|------------------|----------|----------| | **Carousel** | High | 10-15% | Tutorials, guides, reference content | | **Long-form video** | Medium-High | 8-12% | Deep dives, educational | | **Single image** | Medium | 3-5% | Quick tips, visual inspiration | | **Short video** | Low-Medium | 2-4% | Entertainment, trends | | **Live stream** | Variable | 5-10% | Real-time engagement, Q&A |
**Engagement Rate Benchmarks**:
| Follower Range | Excellent ER | Good ER | Needs Improvement | |---------------|-------------|----------|-------------------| | 0-10K | 10%+ | 5-10% | <5% | | 10K-50K | 8%+ | 4-8% | <4% | | 50K-200K | 6%+ | 3-6% | <3% | | 200K+ | 5%+ | 2-5% | <2% |
Implementation
Step 1: Understand Core Algo
Read more
name: platform-algorithm description: Use when understanding Xiaohongshu's recommendation algorithm, optimizing content reach, or troubleshooting low views
Platform Algorithm (平台算法深度)
Overview
Xiaohongshu's recommendation algorithm determines which posts appear on users' explore pages, who sees your content, and how widely your posts spread. Understanding the algorithm isn't about gaming the system—it's about creating content that serves the platform's goal: showing users content they'll find valuable and engaging. The core principle: the algorithm rewards content that generates engagement, keeps users on the platform, and builds community. It evaluates posts based on hundreds of signals: engagement metrics (likes, comments, saves, shares), content quality (originality, completeness, aesthetics), user relationships (who you interact with), and timeliness (recency, trending topics). The algorithm constantly evolves, but fundamental principles remain consistent: value to users, authenticity, engagement, and consistency. This guide provides deep understanding of how the algorithm works, practical strategies to optimize your content for algorithm favor, and troubleshooting for when your reach drops. While Xiaohongshu doesn't publicly disclose algorithm specifics, successful creators reverse-engineer what works through testing and observation. This guide compiles those learnings into actionable strategies.
**Key insight**: Top creators don't just "be authentic"—they optimize strategically for the algorithm while maintaining authenticity. They post when their audience is most active (not random times), use formats that generate saves (carousels over single images), encourage engagement (asking questions, ending with CTA), and build relationships (engaging with their community). These aren't manipulations—they're genuine strategies that help content reach the people who will value it most. Understanding the algorithm removes randomness from growth. Instead of posting and hoping for reach, you can create content with algorithmic principles in mind, dramatically increasing your chances of being featured on the explore page and reaching new audiences. The algorithm is your friend, not your enemy: it amplifies content that serves Xiaohongshu users. Focus on serving users, and the algorithm will serve you.
When to Use
**Use when**:
- Experiencing unexplained reach drops
- Launching new content and want to optimize for algorithm favor
- Troubleshooting why specific posts underperformed
- Analyzing competitors who consistently get explore page features
- Planning content strategy with algorithm in mind
- Wanting to understand platform changes and updates
**Do NOT use when**:
- Trying to "hack" or manipulate the system (backfires)
- Focusing on algorithm over audience value (prioritize audience first)
- Analyzing constantly instead of creating (analysis paralysis)
Core Pattern
**Before** (algorithm-ignorant): ❌ "Post randomly, wonder why some posts succeed and others fail" ❌ "No understanding of reach patterns (appears random)" ❌ "Can't recover from reach drops (don't know what went wrong)" ❌ "Growth feels random, unpredictable"
**After** (algorithm-aware): ✅ "Create content with algorithmic principles (engagement, saves, consistency)" ✅ "Understand reach patterns (know why posts perform as they do)" ✅ "Troubleshoot effectively (diagnose and fix reach issues)" ✅ "Growth feels systematic, predictable"
**Algorithm Fundamentals**:
| Factor | Weight | What It Measures | Optimization | |---------|--------|-----------------|----------------| | **Engagement** | High | Likes, comments, saves, shares, time spent | Create engaging content, ask questions, encourage saves | | **Content quality** | High | Originality, completeness, aesthetics | Invest in production value, unique insights, good visuals | | **User relationships** | Medium | Who views, interacts, shares your content | Engage with community, respond to comments, build connections | | **Timeliness** | Medium | Recency, trending topics, posting frequency | Post consistently, jump on trends, stay relevant | | **Account authority** | Medium | Niche expertise, follower trust | Demonstrate expertise, build credibility over time | | **Session quality** | Low-Medium | User engagement patterns (who they engage with after your post) | Create content that keeps users on platform |
Quick Reference
**Key Algorithm Signals**:
| Signal | What Works | What Doesn't Work | |--------|------------|-------------------| | **Initial engagement** | First 30-60 minutes critical | Posting when audience asleep | | **Saves** | Reference-worthy content (tutorials, guides) | Pure entertainment (fun but not save-worthy) | | **Comments** | Conversation starters, questions | Generic content, no engagement hook | | **Shares** | Relatable, valuable content | Self-promotional, boring | | **Time spent** | Long-form content (carousels, long videos) | Short, superficial content | | **Completion rate** | Content watched fully | Click-bait, disappointing content | | **Return viewers** | Consistent posting schedule | Sporadic posting, long gaps |
**Algorithm-Friendly Content Types**:
| Format | Algorithm Appeal | Save Rate | Best For | |--------|------------------|----------|----------| | **Carousel** | High | 10-15% | Tutorials, guides, reference content | | **Long-form video** | Medium-High | 8-12% | Deep dives, educational | | **Single image** | Medium | 3-5% | Quick tips, visual inspiration | | **Short video** | Low-Medium | 2-4% | Entertainment, trends | | **Live stream** | Variable | 5-10% | Real-time engagement, Q&A |
**Engagement Rate Benchmarks**:
| Follower Range | Excellent ER | Good ER | Needs Improvement | |---------------|-------------|----------|-------------------| | 0-10K | 10%+ | 5-10% | <5% | | 10K-50K | 8%+ | 4-8% | <4% | | 50K-200K | 6%+ | 3-6% | <3% | | 200K+ | 5%+ | 2-5% | <2% |
Implementation
Step 1: Understand Core Algo
版本: v3.0 Complete Edition 更新: 2025-01-22 状态: ✅ 完整 (139个技能)
Other skills on xiaohongshu-skills.
- /audio-processing
Use when processing audio for Xiaohongshu content, editing voiceovers, improving sound quality, creating podcasts, or producing audio-based posts
Open skill - /content-layout
Use when designing visual layout for Xiaohongshu carousel content, organizing information on images, creating easy-to-read text overlays, or structuring multi-slide content for maximum engagement and readability
Open skill - /content-planning
Use when planning Xiaohongshu content calendar, running out of content ideas, needing systematic approach to content creation, or wanting to align content with account goals
Open skill - /content-portfolio
内容作品集管理 - 系统化整理、展示和优化你的内容资产
Open skill - /content-repurposing
Use when repurposing Xiaohongshu content, recycling existing posts, adapting content for different formats, maximizing content value, or creating content variations from core material
Open skill - /content-scaling
内容规模化生产 - 从单打独斗到系统化内容工厂的高效方法论
Open skill

