Menu
Generates images using gemini-image script. Use when user asks to generate, create, or make an image.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
This skill discovers relevant skills, agents, commands, and MCP tools by scanning project context and tool capabilities. This skill should be used when entering planning mode, starting work on a new codebase, or when the user asks "what tools should I use?" Auto-triggers with thorough-planning skill.
Guides creation of Claude Code subagents with proper frontmatter, system prompts, tool assignments, and triggering descriptions. Use when user says 'create an agent', 'add an agent', 'write a subagent', 'new agent', 'make an agent', or needs help with agent file structure, frontmatter, system prompts, or triggering conditions.
Use when needing to inspect, debug, or interact with a browser - helps choose between Browser MCP (existing sessions) and DevTools MCP (debugging tools) based on the task requirements
Searches GitHub and web for relevant code examples and documentation using parallel subagents. Use when user says "look online", "find references", "search online", "find examples", or needs external resources.
Creates story-driven technical presentations using Slidev. Focuses on narrative flow, minimal text, and progressive disclosure. Use when working with Slidev, sli.dev, slides.md, or creating technical slide decks.
Use when debugging unclear issues, mysterious failures, intermittent bugs, or problems with unknown root causes - provides framework combining assumption verification with hypothesis-driven investigation to find root causes systematically
| name | generate-image |
| description | Generates images using gemini-image script. Use when user asks to generate, create, or make an image. |
| allowed-tools | Bash(gemini-image:*), Bash(~/scripts/gemini-image:*) |
Generate images using ~/scripts/gemini-image. For multi-image work, maintain visual consistency with theme prompts.
| Usage | Command |
|---|---|
| Basic | gemini-image "prompt" |
| Output path | gemini-image "prompt" -o output.png |
| Aspect ratio | gemini-image "prompt" -a 16:9 |
| Edit image | gemini-image "add X" -i input.png -o output.png |
Aspect ratios: 1:1, 16:9, 4:3, 3:2, 2:3, 9:16
Execute directly:
~/scripts/gemini-image "a futuristic city at sunset" -o city.png
When generating multiple images (presentations, blogs, image series):
| Context | Theme Prompt |
|---|---|
| Tech presentation | "flat vector illustration, vibrant gradients, dark background" |
| Blog post | "soft watercolor style, warm earth tones, organic shapes" |
| Product series | "clean product photography, white background, soft shadows" |
THEME="minimalist illustration, soft blue palette, clean lines"
~/scripts/gemini-image "home office setup, $THEME" -o slide1.png
~/scripts/gemini-image "video call meeting, $THEME" -o slide2.png
~/scripts/gemini-image "work-life balance, $THEME" -o slide3.png
| Mistake | Fix |
|---|---|
| Generating multi-image without theme | Always ask/suggest theme for 2+ related images |
| Forgetting theme mid-series | Track theme, append to every prompt |
| Generic themes | Match theme to content context (tech→modern, nature→organic) |