Stop eyeballing successful channels — paste one link, get 10 action items

• Paste a competitor channel URL and the skill auto-collects public metadata on its 100 most recent videos (adjustable) — no API key, no paid SaaS subscription.
• Upload weekday patterns, top-video title hooks, performance by length bracket, series structure, and the last 90 days' direction, organized into a 10-section report.
• The report doesn't end with impressions — it ends with 10 action items, each paired with its supporting data. Drop them straight into next week's upload plan.

"What is that channel doing right?" — if you've ever opened a competitor's channel and clicked through videos one by one, jotting down view counts, this skill replaces those hours. Scanning hundreds of videos by eye rarely gets you to structural conclusions like "this channel's Tuesday uploads perform 1.4x" or "8-15 minutes is their core length bracket." This skill chains collection (yt-dlp) → aggregation (Python stats script) → interpretation (Claude's language analysis) into one workflow that turns a channel link into a finished Markdown report.

📦 Real content preview

This is a new listing with no reviews yet, so instead of promises, here are the actual included files, quoted verbatim (translated from the Korean originals).

① How the report ends — the final action section of the included report template (report-template.md):

9. Ten action items

Each item follows the format action — supporting data. Example:

1. Lock uploads to Tue/Sat — Tuesday median views of {X}, {Y}x the overall median
2. …
   (All 10 must be grounded in collected data. If you can't produce evidence, replace the item.)

"If you can't produce evidence, replace the item" — that one line is this skill's character. Advice without data doesn't make it into the report.

② The full report outline — the 10 sections the same template enforces:

1. Collection summary · 2. Channel overview · 3. Upload cadence & weekday patterns · 4. Title hook analysis of top videos · 5. Video length distribution × performance · 6. Series/format structure · 7. Last 3 months' direction · 8. Channel comparison table (when analyzing 2+) · 9. Ten action items · 10. Data limitations (mandatory section — may not be omitted)

③ The rules the skill imposes on itself — from SKILL.md:

Prohibited

• No downloading video files (metadata collection only).
• Never hide collection failures — record the number of missing items in the report.
• Never invent numbers that aren't in the data. Every action item must be grounded in collected data.

📁 What's included (4 files total)

Both scripts use only the Python standard library. No pip packages to install — the sole external dependency is free, open-source yt-dlp. The raw data (videos.json/csv, stats.json) stays on disk, so you can run your own follow-up analysis after the report.

💬 Three usage scenes

Scene 1 — A growing YouTuber dissects a competitor

"This channel covers the same topic as mine but is doing way better. Analyze it: https://youtube.com/@somechannel"

→ The skill collects the 100 most recent videos and returns a report.md with patterns like "this channel uploads 2.3x per week, concentrated on Tue/Sat, and 70% of its top-video titles are number-driven ('3 ingredients')" — plus 10 action items.

Scene 2 — Scouting a niche before launching a channel

"Compare these 3 channels in the cooking niche. What length bracket and format should I start with?"

→ Per-channel stats.json files are bundled into a side-by-side comparison table (report section 8): weekly uploads, median views, core length bracket, signature formats, and 90-day trajectory. Supports up to 5 channels.

Scene 3 — An MCN manager's recurring report

"For my creator meeting, regenerate the competitor benchmark report with this month's data."

→ Run the same command again and you get a fresh report as of collection time. Collection failures and missing-item counts are recorded in the "Collection summary" section, so it's ready to present as-is.

👍 Buy this if / 👎 Skip this if

Buy this if you are

• A YouTuber who wants data — not vibes — on a successful competitor's upload patterns, title hooks, and length strategy
• A future creator comparing 3-5 top channels in a target niche before launching
• An MCN manager producing competitor reports on repeat, or a video editor/strategist who wants hard data in client proposals

Skip this if you need

• Watch time, CTR, traffic sources, or other channel-owner-only metrics — these can't be collected from public data, and this skill will not fabricate estimates for them
• View-count time series — views are captured once, as a cumulative value at collection time
• Revenue/RPM estimates — refusing baseless estimates is one of this skill's explicit rules
• A no-Claude workflow — the scripts alone handle collection and statistics, but hook-pattern interpretation and report writing are Claude's job

❓ FAQ

Q1. How do I receive and use it after purchase?
You download the skill folder (4 files) immediately after payment. Place the folder in your Claude Code skills directory, then paste a competitor channel URL into chat and ask for an analysis — collection → statistics → report runs from there. The only prerequisites are yt-dlp (brew install yt-dlp or pip install yt-dlp) and Python 3.9+; the skill checks for yt-dlp in its first step and tells you how to install it if it's missing.

Q2. How do updates work?
You receive the version current at purchase; when an improved version is published on this listing page, you can re-download it. yt-dlp itself is a separate free open-source project — keep it updated in your own environment.

Q3. Can I get a refund?
Refunds for this digital download follow Capafy's platform refund policy. Before buying, check the "Real content preview" and "Skip this if" sections above to confirm the fit.

Q4. Can I use it without AI (Claude)?
Partially. The collection and statistics scripts run standalone in a terminal and produce videos.json/csv and stats.json. However, title-hook categorization, series identification, and writing the report document are steps Claude performs inside Claude Code — so you need Claude Code to get the finished report.

Q5. Can I use it commercially?
Yes. Reports you generate can be used freely for commercial purposes — client proposals, team meetings, MCN recurring reports. Reselling or redistributing the skill files themselves is not permitted, and collection is limited to metadata YouTube exposes publicly.

Q6. What languages does this work in?
The tool itself is language-agnostic: it analyzes any public YouTube channel yt-dlp can read, in any language — Korean, English, Japanese, or otherwise. The skill's internal instructions and report template were authored in Korean, but the skill explicitly instructs the AI to serve and write everything — including your final report — in your conversation language. Talk to it in English, get an English report. Code, commands, and file formats are unaffected.

🛒 Purchase info

$3 · one-time payment · unlimited use. Analyze one channel or compare five every month — no extra cost. For the price of a coffee, turn hours of eyeballing competitor channels into "paste a link → get 10 action items."