linkedin-viral-pro
Compare original and translation side by side
🇺🇸
Original
English🇨🇳
Translation
ChineseLinkedIn Viral Pro — Orchestrator
LinkedIn爆款专业工具 — 编排器
This skill engineers virality end-to-end. Every output includes: the post, the visual plan, NotebookLM instructions, Gemini image prompts, Canva guidance, and strategy.
该技能可端到端打造内容传播力,每份输出都包含:帖子内容、视觉方案、NotebookLM使用说明、Gemini图片提示词、Canva操作指引以及运营策略。
The Pipeline
工作流程
CLAUDE (you are here)
↓ writes post + decides visual format + generates all prompts
NOTEBOOKLM
↓ user pastes Claude's content → researches deeper → builds structure
GEMINI
↓ user pastes image prompts → generates supporting visuals
CANVA
↓ user builds final visual using structure + images
LINKEDIN
↓ post + visual published together for maximum reachCLAUDE (you are here)
↓ writes post + decides visual format + generates all prompts
NOTEBOOKLM
↓ user pastes Claude's content → researches deeper → builds structure
GEMINI
↓ user pastes image prompts → generates supporting visuals
CANVA
↓ user builds final visual using structure + images
LINKEDIN
↓ post + visual published together for maximum reachFirst-Run Protocol
首次运行协议
- Check if exists in the PROJECT ROOT
config/user-profile.md - If NOT: read from this skill folder, create
templates/user-profile.mdin project root from it, tell the user to fill it out, and STOP. Do not proceed until profile is filled.config/user-profile.md - If YES: read it and load user context for all subsequent operations.
- Also check if exists — if yes, read it for learned preferences that should influence all decisions.
config/strategy.md
- 检查项目根目录下是否存在
config/user-profile.md - 如果不存在:读取本技能文件夹下的,以此为模板在项目根目录创建
templates/user-profile.md,告知用户填写该文件并终止运行,用户完成资料填写前不要继续操作。config/user-profile.md - 如果存在:读取该文件并加载用户上下文,用于后续所有操作。
- 同时检查是否存在,如果存在则读取其中记录的用户偏好,用于指导所有决策。
config/strategy.md
Command Routing
指令路由
"Give me a post" / "Write a post about X" / "Post idea"
"给我生成一篇帖子" / "写一篇关于X的帖子" / "帖子创意"
- Read +
config/user-profile.md(if exists)config/strategy.md - Read from this skill folder — follow its FULL instructions
agents/post-writer.md - After writing the post, ask visual mode (Step 4.5 in post-writer.md): A / B / C or any combination — save to as
config/strategy.mdvisual_mode - Create folder in project root
content/YYYY-MM/DD-topic-slug/ - Output files using templates: post.md, prompts.md, strategy-brief.md, performance.md
- Present everything in one response — output only the sections matching the selected visual mode
- 读取+
config/user-profile.md(如果存在)config/strategy.md - 读取本技能文件夹下的,严格遵循其全部说明操作
agents/post-writer.md - 写完帖子后,询问用户视觉模式(见post-writer.md第4.5步):A / B / C或任意组合,将选择结果作为保存到
visual_modeconfig/strategy.md - 在项目根目录创建文件夹
content/YYYY-MM/DD-topic-slug/ - 基于模板输出以下文件:post.md、prompts.md、strategy-brief.md、performance.md
- 在单次响应中呈现所有内容,仅输出与所选视觉模式匹配的章节
"Create calendar" / "Plan my month" / "Content calendar"
"创建日历" / "规划我的月度内容" / "内容日历"
- Read +
config/user-profile.md(if exists)config/strategy.md - Read — follow its full instructions
agents/content-calendar.md - Create in project root
content/calendars/YYYY-MM-calendar.md - Offer to scaffold individual post folders
- 读取+
config/user-profile.md(如果存在)config/strategy.md - 读取,严格遵循其全部说明操作
agents/content-calendar.md - 在项目根目录创建
content/calendars/YYYY-MM-calendar.md - 支持为单篇帖子创建对应文件夹框架
"Analyze this post" / "Score this" / "Rate my post"
"分析这篇帖子" / "给这个帖子打分" / "评估我的帖子"
- Read — follow its full instructions
agents/post-analyzer.md - Score the post, output fixes if below 40/50
- 读取,严格遵循其全部说明操作
agents/post-analyzer.md - 给帖子打分,如果分数低于40/50则输出优化方案
"Analyze performance" / "How am I doing" / "What's working"
"分析表现" / "我的运营情况怎么样" / "哪些内容效果好"
- Read — follow its full instructions
agents/performance-analyzer.md - Scan all files in project root
content/**/performance.md - Analyze any attached/pasted images or PDFs
- Update in project root
config/strategy.md
- 读取,严格遵循其全部说明操作
agents/performance-analyzer.md - 扫描项目根目录下所有文件
content/**/performance.md - 分析所有上传/粘贴的图片或PDF
- 更新项目根目录下的
config/strategy.md
"Generate visuals" / "Give me image prompts" / "Make an infographic"
"生成视觉素材" / "给我图片提示词" / "制作信息图"
- Read — follow its full instructions
agents/visual-generator.md
- 读取,严格遵循其全部说明操作
agents/visual-generator.md
"What's trending?" / "Find me topics" / "Latest news in [niche]"
"有什么热门趋势?" / "给我找一些话题" / "[细分领域]的最新消息"
- Read +
config/user-profile.mdconfig/strategy.md - Read — follow its full instructions
agents/trend-scout.md - Use WebSearch to find trending topics, viral posts, and breaking news in user's niche
- Present structured trend report with post angles
- Save top findings to (append, don't overwrite)
content/ideas/ideas.md
- 读取+
config/user-profile.mdconfig/strategy.md - 读取,严格遵循其全部说明操作
agents/trend-scout.md - 使用WebSearch查找用户所在细分领域的热门话题、爆款帖子和突发新闻
- 输出结构化的趋势报告,附带帖子创作角度
- 将核心发现保存到(追加写入,不要覆盖原有内容)
content/ideas/ideas.md
"Content ideas" / "What should I post"
"内容创意" / "我该发什么内容"
- Read +
config/user-profile.mdconfig/strategy.md - Read — run a quick trend search first
agents/trend-scout.md - Read +
references/viral-frameworks.mdreferences/hook-library.md - Generate 5-10 content ideas with: topic, angle, framework, visual format
- Save to (append, don't overwrite)
content/ideas/ideas.md
- 读取+
config/user-profile.mdconfig/strategy.md - 读取,先快速执行一次趋势搜索
agents/trend-scout.md - 读取+
references/viral-frameworks.mdreferences/hook-library.md - 生成5-10个内容创意,包含:主题、创作角度、使用框架、视觉格式
- 保存到(追加写入,不要覆盖原有内容)
content/ideas/ideas.md
Global Rules (Apply to ALL outputs)
全局规则(适用于所有输出内容)
- Never start with "I am excited to share..."
- Never use corporate jargon (synergy, leverage, circle back)
- Never write paragraphs longer than 3 lines
- Never end with "Follow me for more"
- Never use more than 5 hashtags
- Never put external links in the post body (first comment only)
- Never generate a visual for a story/hot take post (kills reach)
- Never give Gemini a vague prompt — always include style, mood, colors, subject
- All file paths for references/agents/templates are RELATIVE to this skill folder
- 永远不要以"我很高兴分享..."开头
- 永远不要使用企业行话(协同、复用、同步等)
- 永远不要撰写超过3行的段落
- 永远不要以"关注我获取更多内容"结尾
- 永远不要使用超过5个话题标签
- 永远不要在帖子正文中放外部链接(仅可放在第一条评论中)
- 永远不要为观点/热点类帖子生成视觉素材(会降低曝光量)
- 永远不要给Gemini模糊的提示词,必须包含风格、氛围、颜色、主体信息
- 所有参考/Agent/模板文件的路径都是相对于本技能文件夹的相对路径
Reference Files
参考文件
| File | Purpose |
|---|---|
| 17 proven post formulas with examples |
| 90+ hooks sorted by type and niche |
| LinkedIn algo mechanics 2024-2026 |
| 40+ CTAs that drive real engagement |
| NotebookLM paste content + prompts |
| Gemini image prompt formulas + styles |
| Canva templates, dimensions, layouts |
| 文件 | 用途 |
|---|---|
| 17种经过验证的帖子创作公式及示例 |
| 90+个按类型和细分领域分类的钩子素材 |
| 2024-2026年LinkedIn算法运行机制 |
| 40+个可带来真实互动的CTA素材 |
| NotebookLM粘贴内容及提示词模板 |
| Gemini图片提示词公式及风格参考 |
| Canva模板、尺寸、布局参考 |
Agent Files
Agent文件
| File | Purpose |
|---|---|
| Full post creation pipeline |
| Monthly calendar generation |
| Score posts (50-point rubric) |
| Gemini/Canva/NotebookLM prompts |
| Analyze results, update strategy |
| Search trending topics & news |
| 文件 | 用途 |
|---|---|
| 完整的帖子创作流程 |
| 月度内容日历生成 |
| 帖子打分(50分评分标准) |
| Gemini/Canva/NotebookLM提示词生成 |
| 发布结果分析、策略更新 |
| 热门话题及新闻搜索 |
Template Files
模板文件
| File | Purpose |
|---|---|
| Post file structure |
| Visual production prompts structure |
| Performance tracking (user fills) |
| Per-post strategy |
| Monthly calendar structure |
| User context (first-run setup) |
| 文件 | 用途 |
|---|---|
| 帖子文件结构 |
| 视觉素材制作提示词结构 |
| 表现跟踪表(用户填写) |
| 单篇帖子运营策略 |
| 月度内容日历结构 |
| 用户上下文资料(首次运行设置) |