- Home
- Skills
- Resciencelab
- Opc Skills
- Banner Creator
banner-creator_skill
- Python
472
GitHub Stars
1
Bundled Files
2 months ago
Catalog Refreshed
4 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 resciencelab/opc-skills --skill banner-creator- SKILL.md5.2 KB
Overview
This skill creates professional banners using AI image generation and an iterative design process. It guides discovery, produces many variations, previews results in HTML, and crops the chosen design to target ratios. All outputs are organized into a dated .skill-archive folder for easy delivery.
How this skill works
I first gather purpose, target ratio/size, style, content elements, and color preferences. Then I generate a batch of wide (21:9) AI images, create an HTML preview, iterate on favored options with additional variations, and finally crop the approved banner to the requested target ratio and resolution. Generated files are saved under .skill-archive/banner-creator/<yyyy-mm-dd-summaryname>/ for review and delivery.
When to use it
- You need a banner, header, hero image, cover image, GitHub banner, Twitter/X header, or README banner.
- You want multiple visual directions quickly to choose from.
- You need a banner cropped to a specific platform ratio and resolution.
- You want to iterate on a preferred style or incorporate an existing logo.
Best practices
- Collect clear requirements first: purpose, exact ratio/size, style, text elements, and brand colors.
- Generate at an ultra-wide 21:9 ratio, then crop to target sizes for consistent framing.
- Produce large batches (10–20+) to surface distinct directions before iterating.
- Use the HTML preview to compare variations side-by-side and capture specific feedback (numbers and notes).
- When incorporating logos, provide a clean transparent asset and specify placement and scale.
Example use cases
- Create a GitHub README banner (target 2:1, 1280x640) with a minimalist brand lockup and tagline.
- Design a Twitter/X header (3:1, 1500x500) with an illustrated scene and brand colors.
- Generate website hero images (16:9, 1920x1080) with multiple stylistic options for A/B testing.
- Iterate on a favored banner: produce 10–20 variations of the chosen style and deliver a cropped final asset.
- Add an existing mascot/logo into banner layouts and export both wide originals and cropped deliverables.
FAQ
Start with 10–20 wide variations to explore styles; then iterate 10–20 more on favorites until you converge.
Why generate at 21:9 first instead of the target ratio?
Generating ultra-wide gives more compositional room and ensures important elements remain after cropping to specific platform ratios.