Twitter Trend Radar Skill

Use this skill when the user asks to discover emerging products, tools, indie games, browser games, AI apps, SEO/GEO opportunities, or “what is starting to trend on X/Twitter”.

This skill is inspired by the logic of qiayue/Twitter-Trend-Radar: scan X for launch-moment tweets that contain links and engagement, extract domains, reverse-check domain age and traffic-like signals, then rank likely early opportunities.

This implementation uses the local bird CLI instead of paid Twitter/X API or third-party Twitter search APIs.

Core logic

Build a pool of launch-signal search queries.
Run bird search "<query>" --json for each query.
Parse tweets from bird JSON output.
Extract outbound links and domains.
Filter noisy domains such as x.com, youtube.com, github.com, producthunt.com, etc.
Check domain age through RDAP when available.
Score each domain/tweet cluster by:
- engagement
- recency
- link presence
- domain freshness
- launch-signal phrase strength
- repeated mentions
Output a report with suggested landing pages and next actions.

Safety and risk policy

Use this skill read-only by default. Do not post, reply, DM, like, or follow unless the user explicitly asks and accepts platform/account risk.

bird can use X/Twitter GraphQL with browser cookies. This is not the official public X API. X can rate-limit, block, or change private endpoints. Keep searches low frequency, cache results, avoid concurrency, and do not run this as a public multi-user SaaS using one account.

Requirements

Install bird and confirm it can read/search X:

bird --version
bird whoami
bird search "just launched filter:links" -n 5 --json

Recommended bird setup:

bun add -g @steipete/bird

If browser cookie extraction fails, configure bird according to your environment. Common options:

bird --chrome-profile "Default" whoami
bird --firefox-profile "default-release" whoami

Main command

Run the bundled script:

python scripts/twitter_trend_radar.py --topic "browser game" --days 30 --min-likes 20 --limit 30 --format markdown

For JSON:

python scripts/twitter_trend_radar.py --topic "AI agent" --days 14 --min-likes 50 --format json

Save a report:

python scripts/twitter_trend_radar.py --topic "indie game" --days 30 --min-likes 20 --output reports/indie-game-radar.md

Useful topics

For AI/product opportunities:

python scripts/twitter_trend_radar.py --topic "AI app" --days 14 --min-likes 50
python scripts/twitter_trend_radar.py --topic "AI agent" --days 14 --min-likes 50
python scripts/twitter_trend_radar.py --topic "SaaS" --days 30 --min-likes 30

For game SEO opportunities:

python scripts/twitter_trend_radar.py --topic "browser game" --days 30 --min-likes 20
python scripts/twitter_trend_radar.py --topic "made a game" --days 30 --min-likes 20
python scripts/twitter_trend_radar.py --topic "play online game" --days 30 --min-likes 20
python scripts/twitter_trend_radar.py --topic "indie game" --days 30 --min-likes 20

For Meccha/Super Chameleon-like research:

python scripts/twitter_trend_radar.py --topic "chameleon game" --days 60 --min-likes 5
python scripts/twitter_trend_radar.py --topic "hide and seek game" --days 60 --min-likes 10
python scripts/twitter_trend_radar.py --topic "paint game" --days 60 --min-likes 10

Query design

The script combines the topic with launch phrases:

just launched
introducing
built this
made this
made a site
shipped
launch day
now live
new app
new tool
new game
try it
play online

It adds link and engagement filters when supported by X search syntax:

filter:links min_faves:<N> since:<YYYY-MM-DD>

Output interpretation

High scores mean “investigate now”, not “build blindly”. Before building a page or product, manually verify:

Is the linked site real and accessible?
Is the product/game new or merely being rediscovered?
Are people asking for alternatives, online version, pricing, or tutorials?
Is there searchable demand in Google/YouTube/Reddit?
Is the keyword legally safe and not a protected brand misuse?

Recommended follow-up prompt

After running the skill, ask:

Based on this radar output, pick the top 5 SEO/GEO opportunities and generate landing-page briefs with target keywords, page structure, title, H1, FAQ, and CTA.