v0.1.0 — Advanced YouTube SEO Skill Suite

First public release of the YouTube SEO Skills suite for Claude Code.

What's included

8 user-invokable skills:

Skill	Purpose
/youtube-seo	Orchestrator — routes intent, holds the advanced 6-tier ranking model and niche benchmarks
/youtube-seo-audit	Full channel audit with parallel delegation, topical authority scoring, traffic-source analysis
/youtube-seo-video	Single-video deep dive with retention-curve diagnosis, entity coverage, Key Moments schema, audio loudness
/youtube-seo-optimize	Paste-ready Browse + Search title variants, entity-rich descriptions, translated metadata, 15s hook script, VideoObject schema
/youtube-seo-channel	Identity, topical authority, session-chain playlists, trailer script, translations, community strategy
/youtube-seo-keywords	Intent classification, entity mapping, refined opportunity formula, seasonality, topic clusters
/youtube-seo-thumbnail	CV analysis (face/emotion), CLIP similarity vs SERP, Gestalt composition, native Test & Compare plan
/youtube-seo-competitor	Pattern extraction from transcripts, hook analysis, format-market-fit, entity gap matrix, moats

4 Python helper scripts under scripts/:

• fetch_video.py — YouTube Data API → yt-dlp fallback, unified JSON, optional transcript + thumbnail download
• fetch_channel.py — channel details, recent uploads, top-by-views, batched video stats
• analyze_thumbnail.py — OpenCV + optional CLIP for SERP similarity scoring
• audio_loudness.py — yt-dlp + FFmpeg loudnorm (LUFS, true peak, LRA)

Install

```bash
curl -fsSL https://raw.githubusercontent.com/deeployCO/youtube-seo-skills/main/install.sh | bash
```

Pin to this version:

```bash
curl -fsSL https://raw.githubusercontent.com/deeployCO/youtube-seo-skills/main/install.sh | bash -s -- --ref v0.1.0
```

Key differentiator

Unlike generic YouTube SEO advice, these skills model YouTube's actual session-watch-time recommender (Reinforce / persona matching), with surface-specific strategies (Browse ≠ Search ≠ Suggested ≠ Shorts feed) and numeric niche benchmarks. Skills refuse to fabricate Tier 1 metrics (retention, real CTR) — they ask for a YouTube Studio CSV export when needed instead of guessing.

Requirements

• Claude Code
• (Optional) YOUTUBE_API_KEY env var for full tag/stats access
• (Optional) `ffmpeg`, `yt-dlp`, and Python deps from `scripts/requirements.txt` for the helper scripts

MIT licensed. Built by Deeploy.