founder-content
Complete content creation and multiplication system for founders and indie hackers. Use for writing social posts, repurposing content, creating threads, build-in-public updates, or content planning. Triggers on "write a post", "create content", "repurpose this", "thread", "build
What it does
Founder Content System
Everything for creating and multiplying content as a solo founder.
Master Content Creation Workflow
Core Principle: Research → Extract → Adapt → Write
Every piece of content must go through this workflow.
Step 1: Research Hot Content (REQUIRED)
Before writing ANY content, research what's working:
1. Search for viral/high-engagement posts on target platform
2. Find 3-5 top-performing posts on similar topic
3. Note: hook structure, format, engagement type, tone
4. Identify what makes them work (specifics, emotion, contrarian angle)
Search patterns:
[platform] [topic] viralsite:[platform].com [topic] lessons learned[topic] founder thread high engagement
Step 2: Extract Winning Patterns
| What to Extract | Why |
|---|---|
| Hook formula | First line determines if people read |
| Number usage | Specifics add credibility ($400→$180) |
| Emotion triggers | What makes people react (cringe, saved, wasted) |
| Story arc | How tension and payoff are structured |
| CTA design | What drives comments vs likes |
Step 3: Adapt with Your Voice
Core Voice Principles:
- Authentic — real stories, not theory
- Sharp — specific numbers, direct claims
- Self-deprecating — own failures openly
- No fluff — substance over motivation
Adaptation Rules:
- Keep the winning hook structure
- Replace content with YOUR real stories
- Add specific numbers ($3,000 wasted, saved $1,000+)
- Include genuine emotion (still cringe, learned the hard way)
- Avoid: vague claims, motivational fluff, humblebragging
Step 4: Platform-Specific Polish
| Platform | Key Adaptation |
|---|---|
| Twitter/X | Punchy, <280 chars, threads for depth |
| Longer, professional vulnerability, spaced lines | |
| Xiaohongshu | Conversational Chinese, emotional words, 2K images |
Build-in-Public Workflow
Step 1: Gather Context
From GitHub (auto mode):
- Recent commits since last post
- PR titles and descriptions
- Release notes if tagged
From user input (manual mode):
- What shipped (feature/fix/improvement)
- Who it helps
- Why now
- One metric (optional)
- One lesson learned
Step 2: Extract the Story
Every post answers 5 questions:
- What changed? (the ship)
- Who benefits? (the user)
- Why it matters now? (the context)
- One proof (metric, example, before/after)
- One takeaway (lesson or insight)
Step 3: Render for Each Platform
Twitter/X: Under 280 chars, concise, slightly spicy, one insight + one proof
LinkedIn: 8-20 lines with spacing, narrative + framework + takeaway
Xiaohongshu: Chinese-first, structure: 背景→步骤→结果→踩坑→总结
Step 4: Quality Check
- No identical cross-posts
- Each post has a takeaway
- No banned patterns
- Metrics/proof included where possible
Repurposing Framework
Core Principle: One Excellent Piece → 7-10 Platform-Native Derivatives
Step 1: Evaluate Source
High-Value (prioritize): Evergreen topics, top performers, content with data/frameworks, long-form (>1000 words)
Skip: Trend-based, low performers, thin content
Step 2: Extract Atomic Units
| Element | What to Extract |
|---|---|
| Hook | Opening line, attention-grabber |
| Stats | Numbers, percentages, metrics |
| Frameworks | Step processes, models |
| Quotes | Memorable phrases |
| Stories | Anecdotes, case studies |
| Takeaways | Key lessons, actionable tips |
Step 3: Apply STEPPS (from Contagious)
Every derivative needs at least one:
- Social Currency — Makes sharer look smart
- Triggers — Connected to daily habits
- Emotion — Evokes awe, surprise, anger
- Public — Visible behavior
- Practical Value — Useful, saves time/money
- Stories — Narrative that carries message
Step 4: Distribution Schedule
Day 0: Original published
Day 1-2: Tease/announcement
Day 3-7: First wave derivatives
Week 2-3: Second wave
Week 4+: Evergreen rotation
Content Pillars
Good pillars for a founder/builder:
- Your Tech/Product — What you're building, how it works
- Building in Public — Process > results, real learnings, metrics
- Founder Perspective — Unique angle (background, market, journey)
- Systems Thinking — Workflows, optimization, productivity
Weekly mix: 2-3 posts from pillars 1-2, 1 post from pillars 3-4
Voice Rules
Always:
- Include one takeaway per post
- Adapt content per platform
- Use metrics when available
Never:
- Motivational fluff ("believe in yourself")
- Humblebragging / name-dropping
- Vague claims ("game-changing", "revolutionary")
- Thought-leader cringe
- Dunking on competitors by name
Voice Summary:
- Authentic (not performative)
- Direct, earned confidence
- Self-deprecating humor
- Specific details
- No inspirational soup
Thread Formula
Tweet 1 (Hook): Surprising stat or contrarian take
Tweet 2-5: One key point per tweet with proof
Tweet 6: Common mistake / "what most get wrong"
Tweet 7: The solution/framework
Tweet 8: CTA + callback to original
Write 10-15 versions of hook before publishing.
Example Transformation
Input (GitHub commit):
feat: connect MCP to content scheduler
- Added automatic content storage
- Triggers on GitHub push
- Posts to 3 platforms
Output:
Twitter/X:
Just wired up Claude → DB → auto-posting pipeline. GitHub push now triggers content across 3 platforms. Surprising part: 80% of the work was tone adaptation, not infra.
LinkedIn:
Shipped: automated "build in public" pipeline
What it does: GitHub commits → AI-adapted posts → 3 platforms
What I learned: The hard part isn't automation—it's maintaining authentic voice at scale.
Xiaohongshu:
标题:用 Claude + MCP 搭了个自动发帖系统
背景:每次提交都想分享,但手动发三个平台太累
做法:Claude 读 commit → 生成三版本 → 自动发
踩坑:以为难点是技术,其实是语气适配
总结:自动化不是复制粘贴,是让机器学会"变脸"
Platform Defaults
| Platform | Language | Cadence | Format |
|---|---|---|---|
| Twitter/X | English | 3-5/week | <280 chars, threads rare |
| English | 1-2/week | 8-20 lines, spaced | |
| Xiaohongshu | Chinese + EN tools | 2/week | 干货 + 踩坑 mix |
Capabilities
Install
Quality
deterministic score 0.46 from registry signals: · indexed on github topic:agent-skills · 13 github stars · SKILL.md body (6,350 chars)