Traction scoreGitHub stars can be faked, so popularity alone can be misleading. Traction Score looks for broader signs of recent attention, adoption, and active maintenance.
Treats the title and thumbnail as one promise, and uses retention evidence to improve the video itself. A shared voice profile guides drafts; you still handle the upload.
Compatible AgentsThe repository documents support for these agents. The skills may also work with other agents that can load SKILL.md files, but they may need some setup or small changes.
Prepare your voice profile: copy templates/voice.md from the cloned repository to ~/.claude/youtube/voice.md and fill it in. Alternatively, give Claude three of your own videos and ask it to write your voice.md from them.
Prepare your voice profile using templates/voice.md: copy it to ~/.claude/youtube/voice.md and fill it in, or give Claude three of your own videos and ask it to write your voice.md from them.
A completed ~/.claude/youtube/voice.md profile, or three of your own videos to establish your voice
Python 3 available for hookscore.py
Use yt-script to develop the opening for my video idea and proposed title. Generate five hooks using the documented formulas, score them with hookscore.py, and show me the best two with their scored panels and why each fits the idea. Use my voice profile; if it is missing, ask me for three of my own videos first. Keep this first pass focused on the hooks.
Your voice profile or examples of your wording for chapter titles
Python 3 available for chapters.py
Use yt-chapters with my attached SRT transcript to draft approximately eight chapters. Validate that the first starts at 00:00, there are at least three chapters, and each lasts at least ten seconds. Rewrite the draft topic-word titles as clear section promises of three to five words in my voice, and return the chapter block for review.
Audience-retention CSV downloaded from YouTube Studio: video → Analytics → Engagement → audience-retention chart → download icon → Audience retention
A timestamped SRT transcript for the same video
Python 3 available for retention.py
Use yt-retention to analyze my attached YouTube Studio audience-retention CSV alongside the timestamped transcript. Identify the single biggest viewer leak, show what was being said at that point, and recommend one change. Keep the first result focused on that one fix; if retention is healthy and the issue is packaging, say so.