---
name: trending-scanner
description: "Scan live trends across YouTube trending, Reddit hot, Hacker News front page, X/Twitter trends, and Google Trends for a given niche, then rank trend ideas by hotness with a content angle for each. Use when the user (or Mia) wants to find what's hot before it peaks for a niche or platform."
---

# Trending Scanner

Find what's hot before it peaks. Scan live sources for a niche, rank what's rising, and turn each trend into a concrete content angle.

## Step 1 — Define the focus

Pin down the niche, the target audience, and the platform focus before scanning. Ask the user if any of these is missing — a scan without a niche returns noise instead of signal.

## Step 2 — Scan each source

Use WebSearch/WebFetch to scan each of these, gathering specific, current items (not generalities):

- **YouTube trending** — search for trending videos and rising topics in the niche; note view velocity and recency.
- **Reddit hot** — check the relevant subreddits' hot/rising tabs for surging threads and recurring questions.
- **Hacker News front page** — scan the front page for stories adjacent to the niche (especially for tech/business audiences).
- **X/Twitter trends** — look for trending hashtags, topics, and viral posts relevant to the niche.
- **Google Trends** — check for breakout and rising queries in the niche; favor terms with upward momentum.

## Step 3 — Score each trend

Rate every trend's **hotness** as High / Medium / Low based on momentum, freshness, and relevance to the niche. Tag each as **rising** or **near-peak** — the goal is to catch waves on the way up, not on the way down.

## Step 4 — Output

Present trend ideas **grouped by source**, ranked by hotness within each group. For every trend, give:

- The trend (what it is)
- Why it's hot
- Hotness rating (High / Medium / Low)
- Rising vs. peaking
- A suggested content angle tailored to the niche

## Rules

- Only report trends actually found in research — never fabricate trends, view counts, or numbers.
- Prefer rising trends over already-peaked ones.
- If a source returns nothing relevant, say so rather than inventing filler.