メニュー
This skill should be used when the user asks to "generate an image", "create a picture", "make an image of", "edit this image", "transform this photo", or needs AI-generated visual content using Google's Gemini image generation model.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Produce single-file, self-contained HTML artifacts instead of Markdown for any output a human will actually read. Use whenever the user asks for a "report", "research summary", "write-up", "spec", "plan", "PR explainer", "code review", "design mockup", "prototype", "dashboard", "deck", or anything that benefits from tables, SVG diagrams, color coding, side-by-side comparisons, sliders, drag-and-drop, or shareable visuals. Also use when the user wants to share something visual with a teammate, when a Markdown plan would exceed ~100 lines, when interactive tweaking would help (sliders, copy-as-JSON, copy-as-prompt buttons), or when the model is about to resort to ASCII-art charts or Unicode color blocks in Markdown. Strongly prefer this skill for any "summarize and present" task — even if the user doesn't explicitly say "HTML".
Use this skill whenever the user asks to look at GitHub PR feedback, PR review comments, review bot output, Gemini/Claude review feedback, or pasted reviewer suggestions and wants a markdown table that separates what should be fixed from what does not need to be fixed. This skill should trigger for PR feedback triage, deciding whether review comments are actionable, and producing a concise markdown report with yes/no emoji decisions and reasoning.
Use this skill whenever the user wants to interact with Google Cloud's Agent Registry using gcloud commands. Triggers on any mention of "agent registry", "agent-registry", "mcp-servers", "gcloud agents", "register an agent", "list agents", "create a service", "agent service", or any request to manage agents, MCP servers, endpoints, or services in Google Cloud Agent Registry. Also triggers on requests to integrate or use the Google Agent Development Kit (ADK) with the Agent Registry.
Follow Kent Beck's "Tidy First" approach by strictly separating structural changes (renames, extractions, reorganizations that must not change behavior) from behavioral changes (features, bug fixes, logic changes). Use when refactoring, restructuring code, renaming variables/functions, extracting methods, separating concerns, preparing code for new features, or whenever you need structural and behavioral changes to land in separate commits.
Read this skill before making git commits
Automates the end-to-end Git workflow: branch creation, staging, committing, pushing, and opening or updating a GitHub pull request. Use when a feature or fix is ready for review and requires a structured PR, or when branch changes may need to be pushed to an existing PR for follow-up review.
| name | generate-image |
| description | This skill should be used when the user asks to "generate an image", "create a picture", "make an image of", "edit this image", "transform this photo", or needs AI-generated visual content using Google's Gemini image generation model. |
| version | 0.2.0 |
Generate images using Google's Gemini image generation model via Vertex AI. Supports both text-to-image and image-to-image generation.
Environment variables (can be in .env file in working directory):
GOOGLE_APPLICATION_CREDENTIALS - Path to service account JSON fileGOOGLE_CLOUD_PROJECT - GCP project IDRun the generation script using uvx:
# Text-to-image generation
uvx --with google-genai --with python-dotenv python ~/.claude/skills/generate-image/scripts/generate.py "PROMPT"
# Image-to-image generation (with input image)
uvx --with google-genai --with python-dotenv python ~/.claude/skills/generate-image/scripts/generate.py "PROMPT" -i /path/to/input/image.png
| Flag | Description | Default |
|---|---|---|
-i, --input PATH | Input image for image-to-image generation | None |
-o, --output PATH | Where to save the image | Temp file |
--aspect-ratio | "16:9", "1:1", "9:16" | "16:9" |
--size | "1K" or "2K" | "2K" |
# Basic text-to-image - saves to temp file, prints path
uvx --with google-genai --with python-dotenv python ~/.claude/skills/generate-image/scripts/generate.py "A serene mountain landscape at sunset"
# Save to specific location
uvx --with google-genai --with python-dotenv python ~/.claude/skills/generate-image/scripts/generate.py "A cute robot" -o ./robot.png
# Square format
uvx --with google-genai --with python-dotenv python ~/.claude/skills/generate-image/scripts/generate.py "Abstract art" --aspect-ratio 1:1
# High resolution
uvx --with google-genai --with python-dotenv python ~/.claude/skills/generate-image/scripts/generate.py "Detailed cityscape" --size 2K
# Image-to-image: Transform an existing image
uvx --with google-genai --with python-dotenv python ~/.claude/skills/generate-image/scripts/generate.py "Make this scene nighttime with stars" -i ./daytime_photo.jpg
# Image-to-image: Add elements to an image
uvx --with google-genai --with python-dotenv python ~/.claude/skills/generate-image/scripts/generate.py "Add a rainbow in the sky" -i ./landscape.png -o ./landscape_with_rainbow.png
# Image-to-image: Style transfer
uvx --with google-genai --with python-dotenv python ~/.claude/skills/generate-image/scripts/generate.py "Convert to watercolor painting style" -i ./photo.jpg
Generate multiple images in parallel using the batch script (uv inline script):
# Multiple prompts as arguments
uv run ~/.claude/skills/generate-image/scripts/generate_batch.py "A sunset over mountains" "A robot in a garden" "A cat astronaut"
# Read prompts from a file (one per line, # comments ignored)
uv run ~/.claude/skills/generate-image/scripts/generate_batch.py --from-file prompts.txt
# Save to a specific directory
uv run ~/.claude/skills/generate-image/scripts/generate_batch.py -o ./images "prompt1" "prompt2"
# Control concurrency (default: 4)
uv run ~/.claude/skills/generate-image/scripts/generate_batch.py -c 2 "prompt1" "prompt2" "prompt3"
# Combine file + CLI prompts, custom aspect ratio
uv run ~/.claude/skills/generate-image/scripts/generate_batch.py -f prompts.txt -a 1:1 -s 2K "extra prompt"
# Image-to-image batch (same input image, multiple transformations)
uv run ~/.claude/skills/generate-image/scripts/generate_batch.py -i ./photo.jpg "Make it nighttime" "Add rain" "Watercolor style"
| Flag | Description | Default |
|---|---|---|
prompts (positional) | One or more text prompts | - |
-f, --from-file PATH | Read prompts from a text file (one per line) | None |
-o, --output-dir DIR | Directory to save images | Temp dir |
-i, --input-image PATH | Input image for image-to-image (all prompts) | None |
-a, --aspect-ratio | "16:9", "1:1", "9:16" | "16:9" |
-s, --size | "1K" or "2K" | "2K" |
-c, --max-concurrent | Max parallel API requests (semaphore) | 4 |