Index of every page on this site: llms.txt
All skills
sergebulaev avatar

/yt-title-optimizer

@f422cea
Run it yourself

Write high-CTR YouTube and YouTube Shorts titles under 100 characters using 2026 packaging formulas (curiosity-gap, number, how-I outcome, mistake/loss, transformation, versus). Balances curiosity with a real keyword, front-loads the click-deciding words for the 60-char mobile cutoff, and returns 3 to 5 A/B variants tagged by goal for Test and Compare. Use to title or rework a video. Not for the description (use yt-description-writer) or the thumbnail overlay (use yt-thumbnail-brief).

Use this Skill: https://skilld.dev/gh/sergebulaev/youtube-skills/yt-title-optimizer

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

SKILL.md

≈127 tokens for metadata: the name and description. ≈1.2k when used: this file. ≈562 more on demand in 1 file.

Description uses 6.4% of example budget

Before choosing a Skill, your Agent reads its name and description. All available Skills share that space.

  • A shorter description leaves more room for other Skills. This entry exceeds our 1% size suggestion.

In our Claude Code example, all Skill names and descriptions share 8,000 characters. This Skill uses ≈510 characters, or 6.4%.

The 1% threshold is a size suggestion. Longer descriptions can still fit.

Your model, settings, and other Skills decide how much text your Agent can read.

Example settings and source

The example uses a 200k-token context and default Claude Code settings. The count includes the name, description, separators, and when_to_use when present. Codex also counts local file paths.

Skit's source and limits: Codex 0.160.1, Claude Code 2.1.292.

YouTube Title Optimizer

Titles are half the click. This skill turns a topic, a draft title, or a video URL into a shortlist of titles that earn the click honestly, then leaves the final pick (and the A/B test) to you.

When to use

  • User says "give me title ideas for this video"
  • User has a flat or clickbait title and wants sharper, honest options
  • User wants titles tied to a specific goal (CTR for browse, or search-intent)
  • User wants A/B variants to drop into YouTube's Test & Compare

Formulas this skill uses (packaging shapes, Y1-Y6)

Code Formula Best for
Y1 Curiosity-Gap a surprising result whose mechanism is hidden
Y2 Number / Listicle a finite, scannable payoff
Y3 How-I Outcome a first-person result with a number and timeframe
Y4 Mistake / Negativity a costly thing the viewer is getting wrong
Y5 Transformation a before/after arc over a timeframe
Y6 Versus / Comparison two named options in tension (search-friendly)
Y11 Third-Party Success Story a named person/archetype who built a big number from zero

Full skeletons in ../../references/hook-formulas.md. Y11 was added from a 214-video corpus pull (July 2026) where the third-party rags-to-riches title was the most-frequent uncaptured shape; Y1/Y3/Y5 were the confirmed over-performers, and Y2/Y4/Y6 read weaker in the business/creator niche (noted in the reference).

Steps

  1. Gather inputs. The topic or rough title, the real payoff of the video, the target viewer, the main search phrase if any, and whether it is long-form or a Short. If a URL is given, parse it with lib/url_parser.py and ask the user to paste the current title and what the video actually delivers.
  2. Decide the job. Browse-first (lead with curiosity and emotion) or search-first (lead with the exact query phrase). Most videos want browse; how- to and comparison videos lean search.
  3. Shortlist 2 to 3 formulas that fit the payoff, then draft.
  4. Write 3 to 5 titles. Each one:
    • Under 100 chars, ideally 40 to 60, click-deciding words first.
    • One specific number where the claim allows it.
    • Curiosity AND a concrete noun. Never pure mystery, never a dry label.
    • Complements a thumbnail (note what the thumbnail should carry instead).
    • At most one em dash, and a title rarely wants one (a colon reads cleaner); no ALL CAPS, 0 to 1 emoji.
  5. Humanizer pass. Scrub 2026 AI vocabulary (a title is one line: any marker such as "leverage" or "game-changer" gets replaced), break stacked or hollow rule-of-three, strip reveal bridges ("Here's what nobody tells you") and sincerity framing ("honestly?", "real talk"). Replace an excess em dash with a colon, never a period.
  6. Tag each variant with its formula, char count, and primary goal, and flag the one to lead the Test & Compare with.
  7. Approval card. Show the variants with char counts. The user picks or asks for another round.
  8. On approval. Titles ride along when the video is published. If the user is also uploading via this bundle, pass the chosen title as the title kwarg to lib.publish(kind="video"|"short", ...); otherwise hand it over for Studio.

Hard rules

Global voice rules: see root SKILL.md Voice rules. Additional skill-specific rules:

  • 100-char hard cap. Aim 40 to 60 so mobile does not truncate the payoff.
  • Front-load: the words that decide the click go before any colon.
  • Title and thumbnail never repeat the same words. Note the split every time.
  • Curiosity must be honest. The title cannot promise more than the first 30 seconds delivers (a fast click-away is a ranking penalty).
  • Always return at least 3 variants. One title is not a recommendation.

Anti-patterns (skill will refuse)

  • Clickbait the video does not honor ("you won't believe what happened").
  • ALL CAPS titles or 3+ emoji for fake intensity.
  • Keyword-stuffing the title to chase search.
  • More than one em dash in a title, or a dash where a colon does the job.
  • Vague mystery with no concrete noun ("This changes everything").
  • Repeating the thumbnail's words in the title.

Resources

  • ../../references/hook-formulas.md - Y1-Y6 title skeletons and the pairing table
  • ../../references/algorithm-heuristics.md - CTR x retention, packaging, the 60-char cutoff
  • references/title-checklist.md - the per-title scrub and CTR fit list

Related skills

  • yt-thumbnail-brief - design the other half of the click
  • yt-hook-scripter - script the first 30 seconds so the title's promise lands
  • yt-description-writer - the SEO description and chapters
  • yt-content-planner - pair titles and thumbnails across a week

Source: SKILL.md on GitHub

No rule matched.

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

No alerts2mo3 checks · Risk SAFE
  • Gen Agent Trust Hub2mo

    The skill is functional for YouTube title optimization but contains a surface for indirect prompt injection because it processes external URLs and user-provided text while possessing the capability to publish content via an internal library.

  • Socket2mo

    No alerts

  • Snyk2mo

    Risk: LOW · No issues

Signed by skilld at f422cea. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub yesterday.

Activeupdated 2 months ago

README badge

README badge for sergebulaev/youtube-skills/yt-title-optimizer