youtube-analytics-reviewer
Skill2 starsChanged 13 days agoMITView on GitHub
Diagnose YouTube performance with a decision tree: is the problem packaging (CTR), the promise (first 30 s), the body (retention curve), distribution (traffic sources) or the topic? Reads the Analytics API or Studio CSV exports and returns 3 prioritised actions per video or for the channel. Not for full channel positioning (use youtube-channel-audit).
In this repository12 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- youtube-channel-skillsSkill
- youtube-channel-auditSkill
- youtube-comment-managerSkill
- youtube-content-plannerSkill
- youtube-idea-researchSkill
- youtube-metadata-optimizerSkill
- youtube-publisherSkill
- youtube-script-writerSkill
- youtube-shorts-repurposerSkill
- youtube-title-thumbnailSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 3 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
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 15
Before repackaging an old video (to confirm the problem is packaging)
- Distinctive: line 52
Never invent metrics. If CTR is missing, say the verdict on packaging is provisional.
- Distinctive: line 55
Do not blame "the algorithm".
≈ “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
YouTube Analytics Reviewer
Numbers into decisions. The output is never a dashboard; it is "the problem is X, here is the evidence, do Y next, check Z in 7-14 days".
When to use
- "Why did this video flop?", "review my last month", "what should I change?"
- Weekly or monthly review cadence
- Before repackaging an old video (to confirm the problem is packaging)
Input
- Tier 2: video URL(s) or "last 28 days"
- Tier 0/1: Studio exports (Advanced mode > Export > CSV) or screenshots of Reach, Engagement, Audience tabs
- CTR and impressions always come from Studio (not in the Analytics query API): ask for them
Steps
- Load the profile for the channel's norms if recorded, else compute norms from the last 10-20 videos.
- Pull data (Tier 2):
- Channel:
python3 scripts/analytics_report.py top --days 28 - One video:
python3 scripts/analytics_report.py retention <url> --days 90andpython3 scripts/analytics_report.py traffic <url> --days 28
- Channel:
- Compare to the channel's own norm at the same age (first 7 / 28 days), never to other channels.
- Walk the decision tree in
references/diagnosis-tree.md. Stop at the first branch that explains the gap. - Read the retention curve (
audienceWatchRatiovselapsedVideoTimeRatio): initial drop (0-30 s), slope, cliffs (> 5-point drop in a few %), bumps (rewatch = valuable moment), end drop timing. - Write 3 actions max, ordered by impact, each tied to a skill: repackage →
youtube-title-thumbnail; hook/structure →youtube-script-writer; topic →youtube-idea-research; comments/community →youtube-comment-manager. - Set a check: which metric to re-read and when (7-14 days for repackaging).
Output
VIDEO / PERIOD: ...
VERDICT: [Packaging | Promise | Body | Distribution | Topic | Working: scale it]
EVIDENCE: 3 bullets with numbers vs channel norm
RETENTION READ: initial drop, cliffs (timestamps), bumps
ACTIONS (≤ 3): action → skill → expected signal
CHECK ON: date + metricFor a channel review: a table of the period's videos (views, % viewed, subs gained, CTR if provided, verdict) then the 3 channel-level patterns.
Hard rules
Global voice rules: see root SKILL.md §Voice rules.
- Never invent metrics. If CTR is missing, say the verdict on packaging is provisional.
- Small samples (< 1,000 impressions or < 100 views) are noise: say so and do not over-diagnose.
- One change at a time per video, so the effect can be read.
- Do not blame "the algorithm". Translate every statement into viewer behaviour.
Files
references/diagnosis-tree.md../../references/algorithm-heuristics.mdscripts/analytics_report.py
Related skills
youtube-title-thumbnail— repackagingyoutube-script-writer— fix hooks and structure for the next videoyoutube-channel-audit— when the problem is the whole channel, not a video
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.

