cc-title-hooks

Generate high-CTR titles and opening hooks for a piece of content and score them against clarity and curiosity.

SKILL.md

GitHub

When to use

Use this when you have a topic and need a batch of title/hook options plus a scored recommendation — for a video, newsletter, or post. Works best after the content angle is decided.

Not for: thumbnail design (use cc-thumbnail-brief), writing the body, or SEO keyword research beyond the primary term.

Method

  1. Extract the core payoff and the target viewer. Every title must promise something specific to that person.
  2. Draft across proven angles: number/list, how-to, curiosity gap, contrarian, outcome/result, and question. Produce at least 8 so you can compare.
  3. Front-load the value word in the first 3-4 words — feeds and inboxes truncate.
  4. Score each on three axes, 1-5: Clarity (is the payoff obvious?), Curiosity (does it open a gap?), Credibility (is it believable, not clickbait?). Decision point: if Clarity and Curiosity conflict, keep Clarity — confusion loses more clicks than a smaller gap.
  5. Kill any title that over-promises what the content delivers; broken promises tank retention and trust.
  6. Recommend the top 1-2 and note which is best for A/B against the runner-up.

Example

Topic: budgeting app comparison. Options + scores (Clarity/Curiosity/Credibility):

  • "I tested 7 budgeting apps so you don't have to" — 5/4/5 = 14
  • "The budgeting app nobody recommends (but should)" — 3/5/4 = 12
  • "How to pick a budgeting app in 5 minutes" — 5/3/5 = 13
  • "This free app beat my $99/yr one" — 4/5/4 = 13 Recommendation: run option 1 vs option 4 (highest curiosity with intact credibility).

Pitfalls

  • Clickbait that the content cannot pay off — the click is worthless if retention dies.
  • Burying the value word after "In this video I'm going to..." where truncation hides it.
  • Generating one favorite and no alternatives, so there is nothing to A/B test.
  • Scoring on gut feel with no axes, which hides why one title beats another.

Output format

TOPIC: <...>  |  TARGET VIEWER: <...>  |  CORE PAYOFF: <...>

| # | Title | Clarity | Curiosity | Credibility | Total |
|---|-------|---------|-----------|-------------|-------|
| 1 | ...   | x/5     | x/5       | x/5         | xx    |
(8+ rows)

RECOMMENDATION: <#> — <why>
A/B: run <#> vs <#>
MATCHING HOOK (first spoken/written line): <...>