AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
SKILL verified MIT Self-run

Youtube Retention

skill-ravsau-youtuber-skills-youtube-retention · by ravsau

Timestamped fix list from a retention curve. Use when a video loses viewers and you paste the curve or the drop points.

— No reviews yet
0 installs
4 views
0.0% view→install

Install

$ agentstack add skill-ravsau-youtuber-skills-youtube-retention

✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • ✓ Prompt-injection patterns
  • ✓ Secret / credential exfiltration
  • ✓ Dangerous shell & filesystem operations
  • ✓ Untrusted network calls
  • ✓ Known-malicious package signatures

What it can access

  • ✓ Network access No
  • ✓ Filesystem access No
  • ✓ Shell / process execution No
  • ✓ Environment & secrets No
  • ✓ Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README, it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/skill-ravsau-youtuber-skills-youtube-retention)

Reliability & compatibility

✓ Security review passed
0 installs to date
— no reviews yet
● 3d ago

Declared compatibility

Claude CodeClaude Desktop

Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Youtube Retention? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

YouTube Retention

Input: retention curve points or a description of the video and its average view duration. Data source: if ~/.youtube-skills/metrics/.json exists (see youtube-connect), read retention and curve from it. Else ask the user to paste the curve points from Studio. Output: the curve shape, the problem timestamps, one cutting pattern, and a fix list in priority order.

Steps

  1. Name the curve shape. Cliff: most of the loss is in the first 30 seconds. Gradual: steady loss. Bump: a spike where viewers rewatch. Flat: holds all the way.
  2. Compare the average retention with the benchmark table below and label it good, great, or a problem.
  3. Audit the first 30 seconds: promise delivered by 0:05, no preamble, action started by 0:15, an interrupt at 0:25.
  4. Mark each timestamp where the curve drops faster than the segments around it. Write what happens on screen there.
  5. Pick one cutting pattern from the table below for the re-edit.
  6. If the video runs past 8 minutes, place a payoff just after 8:00.

Benchmarks (average retention)

| Length | Good | Great | Problem | |--------|------|-------|---------| | Under 5 min | 50% | 60% | Under 40% | | 5-15 min | 40% | 50% | Under 30% | | 15-30 min | 30% | 40% | Under 25% | | Over 30 min | 25% | 35% | Under 20% | | Shorts | 70% | 90% | Under 50% |

Cutting patterns

| Pattern | Fits | Rule | |---------|------|------| | Progressive rhythm | Explainer | Fast cuts to 3:00, then slower with b-roll, then bursts after 8:00 | | Contrast | Commentary | Calm 15-25 s cuts, a burst of 5-10 quick cuts every 2-3 min | | Narrative loop | Challenge, story | Restate the premise every 2-3 min | | Hybrid tempo | Tutorial | Cuts every 10-15 s while explaining, holds up to 40 s during examples | | Anchor | Personal, documentary | Cut on emotional beats, hold through full statements |

Output format

## Retention: "[title]"
Shape: [cliff / gradual / bump / flat]. Average: [X]%, [good / great / problem] for [length].

### Drops
1. [m:ss] what happens on screen, why viewers leave
2. ...

### Pattern: [name]. Why: one line.

### Fix list
1. First 30 seconds: [rewrite]
2. [m:ss]: [fix]
3. ...

### Rewatch spike: [m:ss or none]. Clip it as a Short: [yes / no]

Hand the drops and the pattern to youtube-edit-list for the cut rows.

Example

Input: 9-minute video, "MLX vs GGUF on the Same Mac"; 28% average; loses 38% by 0:30, flat to 6:00, bump at 6:40, slow fade after

Shape: cliff, then a bump. 28% is a problem for 5-15 min (under 30%). Drops: 0:00-0:30, the creator explains what MLX is before showing a number. Pattern: hybrid tempo, it is a tutorial. Fix list: 1. open on the two tok/s numbers side by side, then explain. 2. 6:40 has the rewatch spike where the quant setting flips, clip it as a Short. 3. Move the 7:30 summary to 8:10.

CRITICAL: Every fix names a timestamp and one on-screen change. A fix without a timestamp is not a fix.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet, be the first.

Versions

  • v0.1.0 Imported from the upstream source.