agentleFS
aya-talebaya-taleb/video-boost-skills/skills/youtube-research/SKILL.md

youtube-research

Skill1 starsChanged 22 days agoMITView on GitHub

Researches what is working in long YouTube videos (over 3 minutes) for one topic, market and language - top recent videos and channels, title and thumbnail patterns, hooks and retention, structure, end screens, mistakes, and YouTube's own rules (monetization and inauthentic content, AI disclosure, copyright) - using only sources from the last 12 months (6 with recent). Use when the user asks what works on YouTube for a topic, why long videos fail, or as part of viral-checklist with the youtube platform.

  • Reads credentials

16 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use instagram-research and x-research.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • bradautomatesSkill

    instagram-research

    Research high-performing Instagram content (posts and reels) from tracked accounts using Apify's Instagram Scraper. Identifies outlier content, analyzes top 5 videos with AI, and generates reports with actionable hook formulas.

    Also in 3 projects that use this

  • bradautomatesSkill

    x-research

    Research high-performing X/Twitter content from tracked accounts using Apify's Tweet Scraper V2. Identifies outlier tweets, trending topics, and content patterns to inform content strategy.

    Also in 4 projects that use this

The file

Files
SKILL.md

YouTube long-form research

Same method, inputs and output format as shorts-research (read it first), for videos longer than 3 minutes. Output: notes-youtube.json.

Where to look, in this order

  1. YouTube's own rules and tips (official): YouTube Help pages (monetization policies and the inauthentic content clarification, altered or synthetic content, end screens, A/B testing of titles and thumbnails, the Intros metric), the YouTube Official Blog, Creator Insider interviews with YouTube staff. Use the latest version of each page and its date.
  2. Top recent videos and channels in the topic (data, example): with YOUTUBE_API_KEY, run shorts-research/scripts/top_shorts.py with --long to get videos over 3 minutes. Otherwise search the web. Channel stats pages that show live numbers (for example OutlierKit or vidIQ channel pages) count as data with the date you read them; note which numbers are estimates.
  3. Studies and analysis: retention benchmarks by length, hook studies, title and thumbnail studies, with sample size and dates. Say clearly when a benchmark comes from one company's own channel.
  4. Failures: templated or mass-produced videos, misleading titles, slow intros, copyright risk for the topic (for book summaries: summarize ideas in your own words, keep quotes short, credit the author, do not read the book aloud).

What to pull out

Title and thumbnail patterns; the first 30 seconds (what top videos do, and the Intros metric); length that fits the topic; structure (chapters, numbered takeaways, payoff); pace (how often the visual changes); narration and voice; on-screen text; examples and stories; calls to action and where they sit; end screens (last 5 to 20 seconds, video must be 25 seconds or longer); series and playlists; posting rhythm; monetization and policy rules; topic-specific risks.

Then run source-check.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.