agentleFS
rahulcommercialrahulcommercial/claude-code-for-content-creators/skills/thumbnail-ab-tester/SKILL.md

thumbnail-ab-tester

Skill0 starsChanged 5 months agoMITView on GitHub

Generates multiple thumbnail concepts for the same clip and ranks them by hook strength using contrast, face presence, text legibility, and emotional cue heuristics. Trigger when user wants thumbnail options, A/B test thumbnails, or improve CTR on Reels/Shorts/YT.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 3 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Avoid feed-camouflage: if your thumbnail looks like 5 others on the explore grid, it loses.

    line 43
  2. Distinctive:

    Don't fabricate emotions / numbers in thumbnails.

    line 68
  3. Distinctive:

    Don't use ©'d game promotional art as a thumbnail.

    line 69

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Thumbnail A/B Tester

Generate 3-5 thumbnail concepts, score them, ship the winner.

When to use

  • User has a video and wants 3+ thumbnail variants.
  • User wants critique on an existing thumbnail.
  • User wants to know which still frame from a clip is strongest.

Variant generation strategy

For one clip, produce variants across these axes:

AxisOption AOption BOption C
FrameHook moment (start)Climax (mid)Payoff (end)
TextOne wordShort phraseNone (visual only)
FaceTight crop (eyes visible)Medium shotObject only
ContrastDark bg + bright subjectInvertedHigh saturation

Pick one option per axis → 3-5 thumbnails total. Don't generate 30 — too much noise.

Scoring rubric (0-10 each, sum / 40)

FactorWhat to check
Hook clarityCan a viewer read the story in < 1 sec?
Face / emotionIf a face: is the emotion strong & specific?
Text legibilityReadable at 320px wide (IG feed size)?
ContrastSubject pops from background?

Heuristics (the boring stuff that matters)

  • Faces with visible eyes outperform face-from-behind ~2:1 on IG.
  • Bright single-colour backgrounds (yellow, red) beat busy stills.
  • Text on thumbnail: ≤ 4 words. Beyond that, no one reads.
  • Avoid feed-camouflage: if your thumbnail looks like 5 others on the explore grid, it loses.

When to NOT use a thumbnail

  • IG Reels cover image is shown on grid only, not feed. Optimise for grid coherence + first-frame hook in the video, not feed clicks.
  • YT Shorts: thumbnail = first frame by default. Either accept it or pin a custom cover.

Output

Variant A — score 32/40
  Frame: t=2.4s (climax)
  Text: "₹4,300 in 14 min"
  Notes: high contrast, specific number, face in lower third
  Strongest factor: hook clarity

Variant B — score 28/40
  ...

Winner: A
Why: specific number + face emotion drives the highest CTR pattern.

What NOT to do

  • Don't fabricate emotions / numbers in thumbnails. Clickbait that doesn't deliver tanks your reach next post.
  • Don't use ©'d game promotional art as a thumbnail.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.