trend-ideas
Skill1 starsChanged 22 days agoMITView on GitHub
Give it a topic and it researches what is trending around it right now in your market and language, then recommends the top 10 video ideas ranked by viral potential, each with the evidence (real breakout videos with views, dates and outlier scores, plus dated news or trend sources) and why it fits now. Uses the YouTube Data API when YOUTUBE_API_KEY is set. Use when the user says give me video ideas for X, what should I make about X, what is trending in X, or top 10 viral ideas.
- Reads credentials
In this repository10 other files it gives its agents.
- publish-rulerSkill
- reels-researchSkill
- shorts-researchSkill
- similar-videosSkill
- source-checkSkill
- tiktok-researchSkill
- video-packagingSkill
- video-scoreSkill
- viral-checklistSkill
- youtube-researchSkill
The file
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 4 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 22
Angles: Write 6 to 10 search angles around the topic: sub-topics, famous books or people, problems people search, and fresh events.
- Distinctive: line 45
Call it viral potential, never a promise.
- Distinctive: line 51
Never copy a competitor's title word for word.
- Distinctive: line 52
Prefer ideas where small channels break out (an open door) over ideas owned by giant channels.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Trend ideas: the top 10 videos to make next
Finds what is breaking out around a topic right now and turns it into 10 ranked video ideas. Every idea shows its evidence, so you can judge it.
Inputs
| Input | Default |
| topic | ask |
| platform | YouTube (long videos); shorts for Shorts |
| market, language | US, en |
| window | last 90 days for trends (never more than 12 months) |
| your format and niche, if known | for example "animated narrator, book summaries"; used for fit |
Steps
- Angles. Write 6 to 10 search angles around the topic: sub-topics, famous books or people, problems people search, and fresh events. Use web search for news and trends from the last 90 days (dated sources only) to add angles that are new right now.
- Scan YouTube (needs
YOUTUBE_API_KEY):
For each video: views per day, outlier (views vs the channel's median of its last 10 uploads; 3x or more is a breakout), views per subscriber, engagement. For each angle: demand (median views per day of the top 10), breakouts, breakouts from channels under 100K subscribers, and rising (are newer videos doing better than older ones in the window). Without a key: use web search for recent top videos with dated view counts, and say the ranking is rough.python3 <skill_path>/scripts/trend_scan.py --angles "angle one" "angle two" ... --region US --lang en --days 90 --long --out scan.json - Clean the scan. Drop videos that are off-topic (search returns loose matches, like a band with a similar name), in another language, or made for another market. Say how many you dropped.
- Find the themes that win. Group the breakouts by what they promise (for example "fix your dopamine", "psychology of discipline"). Note the title patterns, lengths and formats that win, and the gaps: things people clearly want with few good videos.
- Write ideas. For the best themes, write specific video ideas. Each idea:
- working title (the winning pattern, never misleading)
- the hook (first 10 seconds) and the promise
- format and length (fit to the user's format if given)
- why now: the evidence (2 to 4 breakout videos with links, views, date, outlier) and any dated trend or news source
- risks: policy (no AI expert advice on health, money or law), copyright (books: own words, credit), saturation
- Rank. Score each idea 0 to 100 and show the parts:
part weight from demand 30% the theme's median views per day, scaled against the other themes breakouts 25% share of the theme's videos at 3x outlier or more open door 15% share of breakouts from channels under 100K subscribers (small channels can win) rising 15% newer videos beating older ones in the window fit 15% fits the user's format, characters and policies (your judgment, say why) Call it viral potential, never a promise. - Output
ideas.md(ranked list, one block per idea, sources at the end) andideas.json.
Rules
- Evidence comes only from the scan or from dated sources inside the window. No views from memory.
- Never copy a competitor's title word for word. Learn the pattern, write your own.
- Prefer ideas where small channels break out (an open door) over ideas owned by giant channels.
- Plain words.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

