Claude, Find Me a Decent Image Prompt
This skill plugs a curated library of 10,000+ Gemini image prompts into AI assistants like Claude and OpenClaw so you can stop hunting Twitter for inspiration.

What it does
This is a skill — essentially a plugin — for AI assistants like Claude Code, OpenClaw, and Cursor. It connects them to a library of over 10,000 curated prompts for Google’s Gemini image model (branded Nano Banana Pro) so you can describe an image you need in plain English and get back the top three matches, complete with sample images and ready-to-use English prompts. A remix mode also lets you paste an article or video script and receive a prompt tailored to that specific content.
The interesting bit
The project is essentially a well-organized data pipeline dressed up as an agent skill: a SKILL.md file, category JSON files, and a TypeScript script that syncs prompts from Twitter/X twice daily via GitHub Actions. Search is deliberately token-efficient, using grep-style filtering instead of loading the entire library into the assistant’s context window, which keeps your API bill from ballooning while you hunt for the perfect thumbnail.
Key highlights
- 10,000+ prompts sorted into use-case categories (social media, avatars, e-commerce, game assets, etc.)
- Returns up to 3 recommendations with sample images, translated descriptions, and the exact English prompt
- Content remix mode ingests your text or script and generates a customized illustration prompt
- Auto-updated twice daily via GitHub Actions syncing community sources
- Token-minimal grep search; never dumps the full database into context
Caveats
- The category counts in the README appear inconsistent (Social Media Post alone is listed as 10,000+, roughly the claimed total), so the exact distribution is unclear.
- Requires a skill-compatible assistant (OpenClaw, Claude Code, Cursor, etc.); it is not a standalone CLI or web app.
- Prompts are tuned for Gemini’s Nano Banana Pro; other models like DALL-E or GPT Image may need adjustments.
Verdict
Handy if you generate images with Gemini and want your coding assistant to handle prompt gruntwork. Ignore it if you don’t use a skill-compatible AI assistant or prefer crafting every prompt by hand.
Frequently asked
- What is YouMind-OpenLab/nano-banana-pro-prompts-recommend-skill?
- This skill plugs a curated library of 10,000+ Gemini image prompts into AI assistants like Claude and OpenClaw so you can stop hunting Twitter for inspiration.
- Is nano-banana-pro-prompts-recommend-skill open source?
- Yes — YouMind-OpenLab/nano-banana-pro-prompts-recommend-skill is an open-source project tracked on heatdrop.
- What language is nano-banana-pro-prompts-recommend-skill written in?
- YouMind-OpenLab/nano-banana-pro-prompts-recommend-skill is primarily written in TypeScript.
- How popular is nano-banana-pro-prompts-recommend-skill?
- YouMind-OpenLab/nano-banana-pro-prompts-recommend-skill has 1.8k stars on GitHub.
- Where can I find nano-banana-pro-prompts-recommend-skill?
- YouMind-OpenLab/nano-banana-pro-prompts-recommend-skill is on GitHub at https://github.com/YouMind-OpenLab/nano-banana-pro-prompts-recommend-skill.