AgentStack
SKILL verified MIT Self-run

Youtube Seo Optimize

skill-deeployco-youtube-seo-skills-youtube-seo-optimize · by deeployCO

>

No reviews yet
0 installs
5 views
0.0% view→install

Install

$ agentstack add skill-deeployco-youtube-seo-skills-youtube-seo-optimize

✓ 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.

Are you the author of Youtube Seo Optimize? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

YouTube Metadata Optimizer (Advanced)

Produce paste-ready metadata that optimizes for the specific surface the user cares about, not a generic keyword-stuffing pass. Browse and Search require different strategies; generate both when the goal is mixed discovery.

Required Inputs (ask in one batch if missing)

  1. Video URL or raw topic + video outline/script
  2. Target surface: Browse (Home feed) / Search / Suggested / Shorts /

mixed — this changes the title and thumbnail strategy

  1. Primary target keyword (or ask for suggestions via

youtube-seo-keywords)

  1. Secondary keywords (2-4) for semantic coverage
  2. Target audience (beginner / intermediate / expert) and viewer

persona if known

  1. Channel niche + brand voice (serious / playful / contrarian /

warm)

  1. Is this long-form or Shorts? (and length if long-form)
  2. Top 3 audience countries/languages (for translated metadata)
  3. Competing videos: 3 URLs or keyword for SERP grid fetch

Optimization Strategy by Surface

| Surface | Title pattern | Thumbnail rule | Description priority | |---------|--------------|----------------|---------------------| | Browse | Emotional hook first, keyword second | Face + emotion + ≤3 words text | Above-fold selling the click | | Search | Keyword first, hook second | Clarity over drama | Full keyword + entity coverage | | Suggested | Adjacency to parent video | Same grid style as parent | Cross-link to parent video | | Shorts | ≤40 chars, visceral hook | Vertical crop, muted-readable text | Minimal — first 2 lines only | | Mixed | Browse title + Search-heavy description | Browse-style thumbnail | Full entity coverage |

Outputs (deliver all nine blocks)

1. Titles — 5 variants tagged by surface

Deliver as a table with length, surface tag, hook type, and primary keyword position:

| # | Title | Chars | Surface | Hook | KW Pos | |---|-------|-------|---------|------|--------| | A | ... | 64 | Browse | Curiosity | char 42 | | B | ... | 68 | Search | Authority | char 1 | | C | ... | 59 | Browse | Number | char 3 | | D | ... | 71 | Mixed | Contrarian | char 8 | | E | ... | 66 | Search | Benefit | char 5 |

Rules:

  • All ≤70 chars (mobile cutoff) unless intentional overflow justified
  • Primary keyword present in all
  • No bait-mismatch with the actual video content
  • Include an emotional driver tag: `[curiosity|urgency|benefit|fear|

authority|contrarian|novelty]`

  • At least 2 Browse-primary and 2 Search-primary variants

2. Description (full, paste-ready)

Template:

{HOOK LINE — 100-150 chars, primary keyword in first 60 chars, restates
title intent, gives a click-for-more reason. This is the above-the-fold.}

{BODY — 3-5 paragraphs, 1,500-3,500 chars total. Explains what the
viewer will learn, why it matters, who it is for, what they will be able
to do after watching. Mentions all secondary keywords and 8-15 Knowledge
Graph entities from the topic cluster (people, brands, concepts, places,
products). Natural density. Uses short paragraphs for scan-reading.}

⏱️ CHAPTERS
0:00 {descriptive keyword-rich label}
0:XX {label 2}
X:XX {label 3}
...

🔗 LINKS MENTIONED
- {link 1 — labeled}
- {link 2 — labeled}

📚 RESOURCES
- {resource 1}
- {resource 2}

📺 RELATED VIDEOS
- {related video 1}
- {related video 2}

👋 ABOUT {channel name}
{1-2 sentence channel pitch} → Subscribe: {channel URL}

📩 BUSINESS INQUIRIES
{email}

{FTC disclosure if affiliate/sponsored}

#primaryKeyword #secondaryTopic #niche

Rules:

  • First line contains primary keyword
  • 1,500-3,500 chars total
  • All secondary keywords covered naturally
  • Entity coverage list printed as a checklist before the description

so the user sees which entities were included

  • Exactly 3 bottom hashtags (first = strongest, will show above title)
  • Key Moments / chapter format: M:SS label or MM:SS label (Google

Search rich result compatible)

3. Tags (comma-separated, paste-ready)

primary keyword, secondary 1, secondary 2, long-tail 1, long-tail 2,
entity 1, entity 2, broad niche, channel brand

Rules:

  • Tag 1 = exact primary keyword
  • 5-15 tags
  • <500 chars total
  • Mix: 1-3 broad + 5-8 specific + 2-3 entity-based + 1-2 brand
  • Tags reference the same entities present in the description

4. Chapters (paste-ready, ≥3, first = 0:00)

Each chapter label is keyword-rich but natural, ≥10s apart, matches the actual video structure. Return as a plain block the user pastes into the description.

5. 15-Second Hook Script

The most under-optimized part of most videos. Write a ready-to-shoot opening that solves the 0:30 retention problem:

[0-2s]   VISUAL HOOK: {what is on screen}
         AUDIO HOOK:  "{first sentence, ≤12 words, payoff preview or
                       pattern-interrupt}"

[2-6s]   "{stake / why this matters — make the viewer feel the cost
          of not knowing}"

[6-10s]  "{credibility flash — 1 sentence, what gives you authority
          here}"

[10-15s] "{explicit promise — what the viewer will have by the end}"
         → CUT TO: first content beat

Rules:

  • No "Hey guys welcome back to the channel"
  • No long logo animation
  • No subscribe request before value delivery
  • Promise must match the title and thumbnail exactly

6. Pinned Comment (paste-ready)

A 1-2 sentence comment that drives replies (open question preferred). At most 1 link. Used as a soft ranking signal via reply velocity.

Example pattern: {specific question that viewers will have an opinion on} {optional link to lead magnet or next video}

7. End-Screen Plan

  • Slot 1 (best related video): {title suggestion} — why it

continues the session

  • Slot 2 (playlist): {playlist name} — the binge path
  • Slot 3 (subscribe)
  • Slot 4 (channel)
  • Card placements: at retention peaks (from curve if provided) or

at 25%, 55%, 75% marks as fallback

8. Translated Metadata (top 3 audience languages)

For each target language, produce:

  • Translated title (same surface strategy)
  • Translated description (hook line + full body) — NOT machine-

translated, culturally adapted

  • Translated caption file recommendation

YouTube's Custom Channel Translations apply this automatically to viewers with matching language preferences and typically lift international impressions 20-50%.

9. Schema Markup (JSON-LD for the embed page)

Produce a VideoObject JSON-LD block with hasPart[].Clip entries and a SeekToAction target so the video is eligible for Google Search "Key Moments" rich result when embedded on the creator's website:

{
  "@context": "https://schema.org",
  "@type": "VideoObject",
  "name": "{title}",
  "description": "{above-fold description}",
  "thumbnailUrl": "{thumbnail URL}",
  "uploadDate": "YYYY-MM-DDTHH:MM:SSZ",
  "duration": "PT{H}H{M}M{S}S",
  "contentUrl": "{watch URL}",
  "embedUrl": "{embed URL}",
  "publisher": {
    "@type": "Organization",
    "name": "{channel name}",
    "logo": { "@type": "ImageObject", "url": "{logo URL}" }
  },
  "hasPart": [
    {
      "@type": "Clip",
      "name": "{chapter 1 label}",
      "startOffset": 0,
      "endOffset": 60,
      "url": "{watch URL}&t=0s"
    }
    /* ... one per chapter */
  ],
  "potentialAction": {
    "@type": "SeekToAction",
    "target": "{watch URL}&t={seek_to_second_number}",
    "startOffset-input": "required name=seek_to_second_number"
  }
}

Rationale Block (always include)

After the nine blocks, include a "Why these choices" section (≤250 words) citing which Tier signals from youtube-seo each artifact targets (e.g., "Title A targets Browse CTR by leading with curiosity; Title B targets Search APV by leading with exact keyword and intent match."). Do not restate the full ranking model.

Keyword Discovery Fallback

If no primary keyword is given:

  1. Ask for the topic
  2. Run youtube-seo-keywords on the topic
  3. Present 5 candidate keywords ranked by opportunity score
  4. Wait for user selection before generating metadata

Shorts Template (when <60s vertical)

Replace blocks 1-5 with the Shorts template:

  • Title: ≤40 chars, visceral hook, optional #Shorts
  • Description: ≤500 chars, first 2 lines only matter. Include primary

keyword and 1-2 hashtags max.

  • Chapters: skip (Shorts don't use chapters)
  • Hook script: 0-1s visual + audio pattern-interrupt, loop ending
  • CTA: pinned comment to long-form video
  • Audio note: trending-library audio vs original remixable — pick

based on goal (discovery vs brand)

Error Handling

| Scenario | Action | |----------|--------| | No topic or URL provided | Ask the required-inputs batch above | | Non-English video | Produce all blocks in the video's native language, plus English translation as a secondary block | | Multi-language target | Generate the translated metadata block (output #8) for all target languages | | Sponsored / affiliate | Include FTC disclosure automatically; flag if user does not request it | | Made-for-Kids content | Skip pinned comment block (comments disabled); emphasize thumbnail + title |

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.