Trend Research Skill

Purpose

Ground topic selection in real YouTube performance data, not guesses.

Process

1. Call YouTube Data API search.list for the channel's niche, sorted by
   relevance/date, to get currently trending video titles.
2. Call videos.list on the top results to pull view count, like count,
   and publish date — this tells you actual performance, not just
   existence.
3. Note recurring title patterns, formats, and hooks across top performers
   — not to copy them, but to understand what's currently pulling views in
   this niche.
4. Cross-reference with deep-research's practitioner-pass findings.

Output

A short ranked list: topic candidates with the real numbers behind them
(views, recency, why it's trending), handed to the user alongside
deep-research findings before any script is written.

Notes

• Respect YouTube API quota — this uses search.list (100 units) and
  videos.list (1 unit per video) calls; batch video IDs into a single
  videos.list call rather than one call per video.
• Never fabricate numbers if the API call fails — report the failure.