Social media content research and marketing skills for Claude Code and Cowork
npx skills add https://github.com/bradautomates/head-of-content --skill tiktok-research使用 CLI 安装这个技能,并在你的工作区中直接复用对应的 SKILL.md 工作流。
Your AI-powered content research coworker. Stop guessing what content works—let Claude research, analyze, and surface winning patterns across X/Twitter, Instagram, YouTube, and TikTok.
Check out this YouTube video where we walk through setting up and using Head of Content.
A Claude Cowork skill set that automates social media content research—identifying high-performing content, analyzing viral patterns with AI, and generating actionable content strategies.
| Service | What It Powers | Sign Up |
|---|---|---|
| Apify | X/Twitter, Instagram, TikTok scraping | Sign up for Apify |
| TubeLab | YouTube outlier detection | Sign up for TubeLab |
git clone https://github.com/bradautomates/head-of-content.git
cd head-of-content
Copy the example environment file and add your API keys (see API Keys Setup below):
cp .env.example .env
Then edit .env with your API keys.
head-of-content folderThe repository includes 6 .skill files that need to be added to Claude Desktop:
.skill file from your head-of-content folder:
x-research.skillinstagram-research.skillyoutube-research.skilltiktok-research.skillvideo-content-analyzer.skillcontent-planner.skillRepeat the "Add Skill" process for each file.
This repository requires 3 API keys.
Apify powers the social media scraping for X, Instagram, and TikTok.
Actors used:
Setup:
Cost: Free tier includes $5/month. Research runs typically cost $0.10-$0.50 depending on volume.
TubeLab provides YouTube outlier detection and video intelligence.
APIs used:
Setup:
Cost: Free tier available. Paid plans start at ~$29/month for higher limits.
Gemini AI analyzes video content to extract hooks, patterns, and structure.
Cost: Free tier includes 60 requests/minute. More than enough for research.
Copy the example environment file:
cp .env.example .env
Open .env in your editor and add your API keys:
# YouTube Research - TubeLab API
TUBELAB_API_KEY=your-tubelab-api-key-here
# X/Twitter, Instagram, TikTok Research - Apify API
APIFY_TOKEN=your-apify-token-here
# Video Analysis - Gemini API
GEMINI_API_KEY=your-gemini-api-key-here
Test that your environment is configured correctly:
# Load environment variables
export $(cat .env | grep -v '^#' | xargs)
# Verify they're set
echo "Apify: ${APIFY_TOKEN:0:10}..."
echo "TubeLab: ${TUBELAB_API_KEY:0:10}..."
echo "Gemini: ${GEMINI_API_KEY:0:10}..."
Before running research, configure which accounts to track in .claude/context/:
Edit .claude/context/x-accounts.md:
# X/Twitter Accounts to Track
| Username | Niche | Notes |
|----------|-------|-------|
| @levelsio | indie hacking | bootstrapped founder, viral tweets |
| @naval | philosophy/startups | high engagement threads |
| @sahaborhani | AI/tech | technical content |
Edit .claude/context/instagram-accounts.md:
# Instagram Accounts to Track
| Username | Niche | Notes |
|----------|-------|-------|
| @garyvee | marketing | high volume, reels focus |
| @hormozi | business | educational content |
Edit .claude/context/tiktok-accounts.md:
# TikTok Accounts to Track
| Username | Niche | Notes |
|----------|-------|-------|
| @garyvee | marketing | cross-platform content |
| @tiktokcreators | meta/tips | platform trends |
Edit .claude/context/youtube-channel.md with your channel info:
# YouTube Channel Context
## Channel Information
- **Channel Name:** Your Channel Name
- **Channel Handle:** @yourhandle
- **Channel ID:** UCxxxxxxxxxxxxxxxxxx
## Niche & Positioning
- **Primary Niche:** Your niche
- **Content Style:** Your style
- **Target Audience:** Your audience
## Content Strategy
- **Typical Video Length:** 10-20mins
- **Upload Frequency:** Weekly
- **Best Performing Topics:** Your best topics
## Goals
- **Growth Targets:** Your subscriber goal
- **Content Goals:** Your content goals
Once connected, these skills are available in Claude Desktop:
| Command | Description |
|---|---|
/x-research |
Research X/Twitter content |
/instagram-research |
Research Instagram reels and posts |
/youtube-research |
Find YouTube outliers in your niche |
/tiktok-research |
Research TikTok videos |
/content-planner |
Run all research + create content plan |
Type / in the chat to see all available skills.
In Claude Desktop, simply invoke the skill:
Run x-research
This will:
x-research/YYYY-MM-DD_HHMMSS/For comprehensive cross-platform research:
Run content-planner
This will:
content-plans/YYYY-MM-DD_HHMMSS/You can also just ask Claude:
Each research run creates a timestamped folder:
x-research/2024-01-15_143022/
├── raw.json # Raw API data
├── outliers.json # Identified outliers with metrics
├── video-analysis.json # AI-extracted patterns (if videos)
└── report.md # Human-readable summary
content-plans/2024-01-15_143022/
├── content-ideas.md # Cross-platform content opportunities
├── x-playbook.md # X/Twitter specific intelligence
├── instagram-playbook.md # Instagram specific intelligence
├── youtube-playbook.md # YouTube specific intelligence
└── tiktok-playbook.md # TikTok specific intelligence
Each platform uses optimized engagement weights:
| Signal | Weight | Why |
|---|---|---|
| Bookmarks | 4x | Strongest intent signal |
| Replies | 3x | Conversation starter |
| Retweets | 2x | Amplification |
| Quotes | 2x | Commentary |
| Likes | 1x | Passive approval |
engagement = likes + (3 × comments) + (0.1 × views)
engagement = likes + (3 × comments) + (2 × shares) + (2 × saves) + (0.05 × views)
score = zScore × recency_boost (5% daily decay)
Outlier threshold: Content scoring above mean + (2.0 × std_dev) is flagged as an outlier.
# Make sure .env exists and has your keys
cat .env
# Load environment variables before running
export $(cat .env | grep -v '^#' | xargs)
Make sure you've added accounts to track in .claude/context/:
x-accounts.mdinstagram-accounts.mdtiktok-accounts.mdThe free Apify tier has usage limits. If you hit them:
MIT License - Feel free to modify and distribute.
Contributions welcome! To add new skills or improve existing ones:
/skill-creator to scaffold new skillsPractical AI implementation for real business results
Learn how to implement AI tools and workflows that actually move the needle for your business on my YouTube channel.
Follow @corpcowboy_ on X