trend-discovery

v2026.09.25

Discover rising category conversations, formats, sounds, questions, and creator patterns across social platforms, then separate durable demand signals from short-lived noise.

GitHub
Install command
npx skhub add gooseworks-ai/trend-discovery
Markdown
SKILL.md

Trend Discovery

Find trends a brand can use without producing a generic list of popular hashtags.

Inputs

  • Brand, category, audience, market, and platforms.
  • Decision horizon: react this week, plan next month, or shape a quarter.
  • Optional competitor and creator seed list.

Workflow

  1. Build search seeds from category language, problems, desired outcomes, products, competitors, creators, and adjacent interests.
  2. Use scrapecreators-api to collect relevant TikTok trend feeds, songs, song-linked videos, hashtags and creators; Instagram trending Reels, audio and search results; YouTube trending Shorts and search results; and Reddit or Pinterest discovery signals where the category warrants them.
  3. Compare the recent window with a baseline. A trend must show acceleration, cross-account repetition, or migration across platforms—not merely high lifetime views.
  4. Use outlier-post-finder and transcript-intelligence to understand the content, not just the metric.
  5. Keep topic, sound, format, creator, and search-demand signals distinct. Classify each as emerging, accelerating, established, fading, or seasonal. Score brand fit, audience fit, regional relevance, shelf life, production effort, and reputational risk.
  6. Turn the best signals into a response: participate, adapt the format, answer the question, create an evergreen variant, or ignore.

Output

  • Coverage and baseline.
  • Ranked trend table with evidence, source links, stage, confidence, shelf life, and brand fit.
  • Format/audio/topic patterns.
  • Five recommended content tests with a clear reason and timing.
  • Watchlist and false positives.

Do not claim a trend from one viral post. Label directional evidence and missing data.

Discovery
Tags

No tags published for this skill.

Version
Latest version metadata

Version

v2026.09.25

Published

Sep 25, 2026

Category

Uncategorized

License

MIT

Source path

skills/social/composites/trend-discovery

Default branch

main

Latest commit

a9f4676

Tree SHA

99d4730