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

Youtube Description

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

Search-or-browse call, then a full YouTube description with title check, tags, and a fixed footer. Use before upload, or when views stall and the metadata needs an audit.

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

Install

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

✓ 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-description)

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 Description? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

YouTube Description

Input: title, transcript or notes, surface and metrics if known. If ~/.youtube-skills/metrics/.json exists (see youtube-connect), read CTR and traffic_sources from it. Output: the surface call, the title check, one paste-ready description (head, body in the surface's shape, fixed footer), tags and 3 long-tail keywords.

Source: an audit of four MrBeast uploads and eight AI teaching channels (RESEARCH.md). MrBeast runs one byte-identical footer, one objection paragraph per video, money on top, socials at the bottom, no chapters. Teaching channels add chapters (5 of 8) and rarely fix the footer (3 of 8).

Steps

  1. Name the buyer query. Write the exact phrase a viewer types when they need this video. If no phrase exists, the surface is browse.
  2. Pick the surface. Search: the query exists and the channel can beat the top 3 results. Browse: the topic is a want, not a need, and the viewer came from the feed or a Short.
  3. Check the title. Search: keyword inside the first 40 characters. Browse: keep the packaging line as is. Never keep a keyword the video does not answer in the first two minutes.
  4. Write the head: the one offer for this video (sponsor, product or booking), one line, label → url. Money above the fold.
  5. Write the objection line: name the top comment before it exists (what was estimated, not measured, or skipped). Mandatory.
  6. Write the body in the surface's shape (template below), then paste the fixed footer. Never edit the footer per video. Search body line 1 is keyword + claim, line 2 is what the viewer gets.
  7. List 5-10 tags and 3 long-tail keywords with a low, medium, or high competition guess. Two minutes, no more.
  8. If metrics were given, name the one problem: CTR under 4% is packaging, midpoint retention is editing, no session continuation is the end screen.
  9. Check: one objection line, one comment question, zero subscribe requests, no hand-written CTA inside the body.

Rules

  • One surface per video. A video that tries to win both usually wins neither.
  • Long-tail beats broad. "Claude Code weekly limit reset time" beats "Claude Code tips".
  • The footer never changes. A changing footer reads as a pitch; a fixed one reads as the brand.
  • Browse shape: no chapters, no question block. Search shape: chapters as micro-promises (a number, a comparison, a finding).
  • One ?utm_source= on owned links only. Business email once, at the bottom. Zero to three hashtags.
  • Every video ends with an end screen that names one specific next video. Label AI-generated content in the upload settings.

Template

Surface: [search | browse]. Buyer query: "[phrase]". Title: "[title]" ([chars] chars, keyword at char [n])

[HEAD]  → ?utm_source=

[BODY, browse]                          [BODY, search]
  
                        
  
                  ▶ CHAPTERS  0:00  ...
                                        ▶ QUESTIONS ANSWERED  
                                        ▶ LINKS AND SOURCES  
                                        
                                        #tag #tag

[FOOTER, identical on every upload]
🎓  → ?utm_source=
🧪  → ?utm_source=
💻 Code → 
🐦 X →     💼 LinkedIn → 
📧 Sponsorship inquiries: 

Tags: tag1, tag2, ...   Long-tail: "[phrase]" (low) | "[phrase]" (medium) | "[phrase]" (low)
Problem to fix first: [packaging | editing | end screen | none]

Example

Input: "$20 or $100 or $200 AI plan: here's how to decide"; chapters given; CTR 3.1%

Surface: search. Buyer query: "which claude plan should I buy". Title: keyword at char 0, kept.

📞 I'll help you make the $200 plan pay back, 55 min → https://example.io/meet?utm_source=abc123

I used $300 of tokens in one day on the $100 plan. This is the decision ladder: when $20 is enough, when $100 and $200 pay for themselves, and how limit resets change the math.
The token values per plan come from a third-party estimate, not from the vendor.

▶ CHAPTERS  0:00 $300 of tokens in one day, on a $100 plan | 1:35 The full ladder: 4x the limits for 2x the price
▶ QUESTIONS ANSWERED  Is the $100 plan worth it? When should you leave the $20 plan?
▶ LINKS AND SOURCES  Prompt pack: https://github.com/example/reset-prompts
Which plan are you on, and did it pay back? Tell me in the comments.
#codex #claudecode

🎓 All courses → https://learn.example.io/courses?utm_source=abc123
💻 Code from my videos → https://github.com/example
📧 Sponsorship inquiries: hello@example.io

Tags: claude plan, claude max plan, claude pro vs max, claude code limits, ai coding plan cost
Long-tail: "claude max plan worth it" (low) | "claude pro vs max" (medium) | "claude code weekly limit" (low)
Problem to fix first: packaging. CTR sits under 4%, so the thumbnail must show the price ladder.

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.