2.5k
GitHub Stars
2
Bundled Files
2 months ago
Catalog Refreshed
3 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill openclaw/skills --skill youtube-cover-nano-banana- _meta.json321 B
- SKILL.md6.7 KB
Overview
This skill creates high-converting YouTube thumbnail concepts, short overlay text, and Nano Banana image prompts from titles, hooks, scripts, or marketing copy. It focuses on mobile-legible, emotionally clear designs and can optionally produce an AI-generated cover image when image generation is enabled. The output prioritizes one dominant idea, strong contrast, and text that reads in one second.
How this skill works
The skill first analyzes source copy to extract core topic, audience, main promise, emotion, and any non-negotiable visual details. It chooses a single thumbnail strategy (expressive face, product close-up, before/after, curiosity gap, etc.), compresses overlay text to 2–6 words, and builds a 16:9 Nano Banana prompt describing subject, framing, lighting, color palette, and reserved text space. If generation is requested and API keys are available, it calls Nano Banana to render the image; otherwise it returns a final prompt plus generation notes.
When to use it
- You have a raw title, hook, script, or marketing blurb and need a thumbnail plan fast
- You need a Nano Banana-ready prompt tailored for YouTube thumbnails
- You want overlay headline text that’s readable on mobile
- You need a single clear visual strategy instead of a crowded collage
- You want an automated workflow that can optionally render the image
Best practices
- Provide the strongest claim or angle; the skill infers the thumbnail direction from that
- Specify any required person, product image, brand colors, or banned elements up front
- Keep overlay text to 2–6 words and avoid full-title copy on the image
- Prefer one dominant subject and a clear emotional cue for higher click-throughs
- Request photorealistic, illustrated, or stylized only when you need a specific look
Example use cases
- Turn a 2‑sentence hook into a thumbnail concept plus overlay text and Nano Banana prompt
- Generate a thumbnail plan for a product demo with a close-up product shot and bold headline
- Produce a curiosity-gap thumbnail for a teaser video from a short script excerpt
- Create a warning/threat-style thumbnail for a mistake-focused tutorial
- Output a full JSON thumbnail plan for integration into automated publishing pipelines
FAQ
No. A strong title, hook, or short excerpt is usually enough; the skill infers the best angle from the strongest claim.
Will the generated image exactly match the overlay text placement?
No. Prompts reserve space and describe placement, but image models don’t guarantee pixel‑perfect text rendering; export the prompt and add text in an editor for best results.