← Back to Skills Marketplace
5857
Downloads
8
Stars
33
Active Installs
5
Versions
Install in OpenClaw
/install gemini-image-gen
Description
Generate and edit images via Google Gemini API. Supports Gemini native generation, Imagen 3, style presets, and batch generation with HTML gallery. Zero depe...
Usage Guidance
Install only if you are comfortable sending prompts and selected edit images to Google using your Gemini API key. Treat heartbeat use as requiring extra controls: do not allow automatic social posting, avatar changes, or memory saving unless you have explicit approval, account scope, budget limits, and retention rules. Avoid passing confidential files or non-image paths to --edit.
Capability Analysis
Type: OpenClaw Skill
Name: gemini-image-gen
Version: 1.3.1
The `scripts/gen.py` file contains a data exfiltration vulnerability. The `--edit` argument, intended for image editing, allows reading arbitrary files from the local filesystem via `load_image_as_base64`. The base64-encoded content of these files is then sent to the Google Gemini API (generativelanguage.googleapis.com) as part of the image editing request. This flaw could be exploited by a malicious prompt to exfiltrate sensitive local files to a third-party service.
Capability Assessment
Purpose & Capability
The Python script matches the stated image generation/editing purpose: it uses GEMINI_API_KEY, sends prompts and optional edit images to Google's Gemini/Imagen endpoints, and writes generated files locally. However, HEARTBEAT.md expands the skill into recurring agent self-expression, social posting, avatar updates, and memory saving, which are materially broader than image generation.
Instruction Scope
HEARTBEAT.md marks sharing as optional but tells agents to post to social platforms, update profile avatars, and save results to memory without requiring explicit per-action user consent, target account scoping, or retention guidance.
Install Mechanism
Installation is standard ClawHub or git clone setup. There is no install-time execution, dependency installer, obfuscation, or hidden persistence mechanism in the reviewed artifacts.
Credentials
A Gemini API key, outbound calls to generativelanguage.googleapis.com, local output writes, and reading a user-supplied edit image are proportionate for this skill. The documentation does not prominently warn that prompts and edit-image bytes leave the local machine, and the edit path is not restricted to verified image content.
Persistence & Privilege
Runtime output persistence is expected: generated images, prompts.json, and index.html are written to an output directory. The additional heartbeat instruction to save good results to memory is under-specified and lacks user consent, visibility, retention, or deletion controls.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install gemini-image-gen - After installation, invoke the skill by name or use
/gemini-image-gen - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.3.1
- Added guidance for integrating with the opencode-omo skill for workflow/pipeline image generation.
- Updated documentation to mention opencode-omo in the integration section.
- No changes to core image generation features or CLI options.
v1.3.0
Add INSTALL.md, enrich package.json with engines/style_presets/quickstart/security, cross-promotion with AgentGram and agent-selfie
v1.2.0
Fix SKILL.md frontmatter to proper YAML, replace {baseDir} placeholders with relative paths, add README.md, add .gitignore
v1.1.0
- Introduced style presets for creative control over image generation.
- Added --style and --styles CLI flags to apply and list style presets.
- Expanded documentation with a dedicated "Style Presets" section and improved examples.
- New files: HEARTBEAT.md and package.json added.
v1.0.0
- Initial release of gemini-image-gen.
- Generate and edit images using Google Gemini API; supports Gemini native and Imagen 3 engines.
- Batch image generation with prompt customization and output as an HTML gallery.
- Includes image editing (prompt-based) and style transfer (Gemini engine only).
- Pure Python standard library — zero external dependencies.
- Requires a GEMINI_API_KEY environment variable for usage.
Metadata
Frequently Asked Questions
What is Gemini Image Gen?
Generate and edit images via Google Gemini API. Supports Gemini native generation, Imagen 3, style presets, and batch generation with HTML gallery. Zero depe... It is an AI Agent Skill for Claude Code / OpenClaw, with 5857 downloads so far.
How do I install Gemini Image Gen?
Run "/install gemini-image-gen" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Gemini Image Gen free?
Yes, Gemini Image Gen is completely free (open-source). You can download, install and use it at no cost.
Which platforms does Gemini Image Gen support?
Gemini Image Gen is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Gemini Image Gen?
It is built and maintained by 김덕환 (@iisweetheartii); the current version is v1.3.1.
More Skills