YouTube Research Skill

A Claude Code skill that combines TubeLab (outlier video database with 4M+ videos) and ytstudio (YouTube Studio CLI) to find data-backed video ideas.

What It Does

• Searches TubeLab’s database for outlier videos in any niche
• Pulls your own channel data via YouTube Studio CLI
• Cross-references external outliers with your channel’s top performers
• Generates professional PDF reports with actionable video ideas
• Ranks opportunities by averageViewsRatio (how much a video overperformed its channel)

Prerequisites

• ytstudio CLI - YouTube Studio command-line tool (pip install ytstudio-cli)
• TubeLab API key - Get one at tubelab.net
• Python 3 with fpdf2 package (pip install fpdf2)

Installation

Option 1: Copy to Claude Code skills directory

git clone https://github.com/Samin12/youtube-research-skill.git
cp -r youtube-research-skill ~/.claude/skills/youtube-research

Option 2: Manual

1. Create ~/.claude/skills/youtube-research/
2. Copy SKILL.md and scripts/ directory into it

Setup

Set your TubeLab API key as an environment variable:

export TUBELAB_API_KEY="your-api-key-here"

Make sure ytstudio is authenticated:

ytstudio login
ytstudio status  # verify it works

Usage

Once installed, Claude Code will automatically activate this skill when you ask things like:

• “Research outlier videos about Claude Code”
• “Find video ideas in the AI agents niche”
• “Analyze my channel performance”
• “What videos should I make next?”
• “Do TubeLab research on MCP servers”
• “Generate a video ideas report”

File Structure

youtube-research/
├── SKILL.md                    # Skill definition + full tool reference
└── scripts/
    └── generate_report.py      # PDF report generator

Report Generator

The PDF generator accepts a JSON data file:

python3 scripts/generate_report.py --input data.json --output report.pdf

See SKILL.md for the full JSON schema.

License

MIT