agentleFS
TubeLabHQTubeLabHQ/youtube-research-skills/skills/niche-watchlist/SKILL.md

niche-watchlist

Skill91 starsChanged 11 days agoMITView on GitHub

Use when the user wants to monitor YouTube niches on a schedule and get a digest of new or breakout channels. Requires the TubeLab MCP.

The file

We split this skill into the rules it gives an agent

10 rules. About 0 are common, 6 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Read watchlist/config.json.

    line 19
  2. Very specific:

    Read watchlist/state.json.

    line 20
  3. Distinctive:

    Run each saved search once, exactly as saved.

    line 21
  4. Distinctive:

    Read each new channel and mover, which is free.

    line 23
  5. Very specific:

    Write watchlist/<date>.md on every run, then refresh the stored values.

    line 24
  6. Distinctive:

    If it runs long, cut channels, not the summary line or proof.

    line 64
  7. Very specific:

    Do not save a search without publishedAtFrom, or old channels crowd out the new ones.

    line 70
  8. Distinctive:

    Do not change a saved search between runs, or the differences are noise.

    line 71
  9. Distinctive:

    Do not call a channel new because it dropped out of the sample and came back.

    line 72
  10. Distinctive:

    Do not call a channel a breakout below 1K median views or 5 videos.

    line 73

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Niche Watchlist

Overview

Watch niches on a schedule and report only what changed since the last run. The two files below keep a fixed shape, so each run can be compared with the last one. For each new channel or mover, say what is worth learning from it.

Workflow

  1. Read watchlist/config.json. If it is missing, ask for 1 to 5 niches and write it.
  2. Read watchlist/state.json. If it is missing, this run is a baseline, not a wave of breakouts.
  3. Run each saved search once, exactly as saved. Compare only the channels it returns.
  4. A new channel is one first seen since the last run. A mover is a known channel whose median views rose more than 50 percent over the stored value.
  5. Read each new channel and mover, which is free. Look at its top videos by ratio, the topics it repeats and the title shapes it repeats. Fetch and view its best video's thumbnail, and say so if one fails to load.
  6. Write watchlist/<date>.md on every run, then refresh the stored values.

watchlist/config.json (the saved search can hold other settings, such as a language):

{ "niches": [ { "name": "budget travel", "search": { "query": ["budget travel tips"], "language": ["en"],
  "publishedAtFrom": "month", "sortBy": "foundAt", "sortOrder": "desc", "size": 20 } } ] }

watchlist/state.json:

{ "lastRun": "<date>", "niches": { "budget travel": { "<channelId>":
  { "handle": "@name", "subscribers": 0, "medianViews": 0, "firstSeen": "<date>" } } } }

Output Format

# 📡 Watchlist Digest: {date}

- **Since {last run}:** {n new and n movers, or "No change", or "Baseline"}

## {Niche}

### [{channel}]({url}) · {new or mover} · {subscribers} subs · {median} median views

[![]({best video thumbnail})]({video url})

- **Worth studying:** ✅ {yes, and why in one sentence}
- **Topics working:** {two or three topics}
- **Title formulas:** {template}, for example [{title}]({video url})
- **Best recent video:** [{title}]({video url}) · {views, ratio, date}

- [{channel}]({url}) · {new or mover} · ❌ not worth studying: {why in one sentence}
  • Length: about 600 words. Up to five channels per niche, strongest first. One line for a niche with no change, and one line for a channel not worth studying.
  • Language: plain words and short, complete sentences. Nothing compressed.
  • Bullets: one fact each, under 25 words, except quotes and verbatim titles. No extra sections or bullets.
  • If it runs long, cut channels, not the summary line or proof.
  • Links: link every video, channel, moment, brand and offer you cite, built from IDs, handles and links in the data. Never guess one.
  • Thumbnails: use the 320px mqdefault.jpg image, on its own line under the heading, linked to the video.

Common Pitfalls

  • Do not save a search without publishedAtFrom, or old channels crowd out the new ones.
  • Do not change a saved search between runs, or the differences are noise. An edited niche starts a new baseline.
  • Do not call a channel new because it dropped out of the sample and came back. Check its first-seen date.
  • Do not call a channel a breakout below 1K median views or 5 videos.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.