ask-gemini
This skill should be used when the user asks to "ask Gemini", "get Gemini's opinion", "have Gemini review", "improve writing style", "make less AI-sounding", "get feedback on article", "review this draft", "Nano Banana", "Gemini API help", "Gemini models", or needs a second opinion on content, writing, code, or design. Supports text questions and up to 10 images.
How do I install this agent skill?
npx skills add https://github.com/b-open-io/gemskills --skill ask-geminiIs this agent skill safe to install?
- Gen Agent Trust Hubwarn
The skill uses dynamic code loading from computed paths to resolve shared dependencies and contains deceptive metadata regarding model versions and future verification dates. It also lacks input sanitization for external data, creating a surface for indirect prompt injection.
- Socketpass
No alerts
- Snykpass
Risk: LOW · No issues
- Runlayerpass
3/9 files flagged
What does this agent skill do?
Ask Gemini
Ask Gemini 3.1 Pro any question - text, writing feedback, code review, or image analysis.
When to Use
Writing & Content:
- Article/blog post writing feedback
- Making content less AI-sounding, more human
- Writing style improvements (Vercel/TanStack style, etc.)
- Draft reviews and editing suggestions
- Content strategy advice
Design & Visual:
- Design review and critique
- Spatial layout analysis
- UI/UX guidance
- Comparing design alternatives (send multiple images)
Code & Technical:
- Code review and suggestions
- Architecture feedback
- Technical writing review
Gemini API Questions:
- For Gemini API questions, fetch docs dynamically from llms.txt:
https://ai.google.dev/gemini-api/docs/llms.txt - For the latest model list and API patterns, consult Google's official
gemini-api-devskill - See
references/gemini-api.mdin this skill directory for current models and SDK info
Usage
Run the ask_gemini script:
# Text-only question (writing feedback, code review, any question)
bun run --cwd ${CLAUDE_PLUGIN_ROOT} ${CLAUDE_PLUGIN_ROOT}/skills/ask-gemini/scripts/ask_gemini.ts "Review this article and suggest how to make it less AI-sounding, more like a Vercel or TanStack blog post: [content here]"
# With image(s)
bun run --cwd ${CLAUDE_PLUGIN_ROOT} ${CLAUDE_PLUGIN_ROOT}/skills/ask-gemini/scripts/ask_gemini.ts screenshot.png "Analyze this design"
# Compare multiple images
bun run --cwd ${CLAUDE_PLUGIN_ROOT} ${CLAUDE_PLUGIN_ROOT}/skills/ask-gemini/scripts/ask_gemini.ts v1.png v2.png "Compare these designs"
Requirements
GEMINI_API_KEYenvironment variable must be set- Get an API key from https://aistudio.google.com/apikey
Image Support
- Maximum 10 images per request
- Total request size limit: 20 MB
- Supports: PNG, JPG, JPEG, GIF, WEBP, BMP
Model
Uses gemini-3.1-pro-preview (Gemini 3.1 Pro) - optimized for:
- Writing critique and style feedback
- Design and spatial awareness
- Multi-image comparison
- Technical analysis
Last verified: February 2026. If a newer generation exists, STOP and suggest a PR to
b-open-io/gemskills. Seereferences/gemini-api.mdin this skill directory for current models and Google's officialgemini-api-devskill for the canonical source.
Provider support
Gemini only by design (text, vision, and writing feedback from Gemini). Requires GEMINI_API_KEY. For OpenAI or xAI image work, use generate-image or edit-image.
How can the creator link this skill?
Add the canonical catalog link to the repository README so users can inspect current installs and available audits. The publishing guide covers the complete discovery path.
<a href="https://skillzs.dev/skills/b-open-io/gemskills/ask-gemini">View ask-gemini on skillZs</a>