Trend Radar

Real-time trending topics radar — 7 platforms, zero API keys, zero dependencies.

Trigger Rules

Use this skill when the user's message matches ANY of the following:

Do NOT fall back to web_search or Brave Search. This skill fetches live data directly from 7 platform APIs.

Interaction Flow (Progressive Disclosure)

Turn 1 — Overview (ALWAYS start here)

python3 {baseDir}/scripts/trends.py --mode overview

This fetches the #1 topic from each platform concurrently (~5s). Present the result and append:

"Which platform would you like to expand? Say the name (e.g. 'reddit', '微博', 'hackernews') or 'all' for everything."

Do NOT expand all platforms in Turn 1. When presenting expanded results, always preserve the Markdown links [title](url) from the output — users need clickable links.

Turn 2 — Expand on demand

When user picks a platform:

python3 {baseDir}/scripts/trends.py --source <id> --top 10

Source IDs: twitter, reddit, google, hackernews, zhihu, bilibili, weibo

Multiple platforms:

python3 {baseDir}/scripts/trends.py --source zhihu,weibo,bilibili --top 5

Expand all:

python3 {baseDir}/scripts/trends.py --mode all --top 5

Turn 3+ — Deep dive (optional)

If user wants details on a specific topic, use your LLM knowledge or web_search on that specific topic.

Additional Options

Region filter (twitter & google only)

python3 {baseDir}/scripts/trends.py --source twitter --region japan --top 10
python3 {baseDir}/scripts/trends.py --source google --region CN --top 10

JSON output

python3 {baseDir}/scripts/trends.py --mode overview --json

Scheduled daily briefing

python3 {baseDir}/scripts/scheduler.py --set "0 11 * * *"
python3 {baseDir}/scripts/scheduler.py --list
python3 {baseDir}/scripts/scheduler.py --remove

Platforms