← Back to Skills Marketplace
168
Downloads
0
Stars
1
Active Installs
2
Versions
Install in OpenClaw
/install nano-banana-pro-custom
Description
Generate or edit images using OpenAI-compatible API. Supports multi-image input, fine-tuned models, and multiple configuration sources (env vars, openclaw.js...
Usage Guidance
This skill appears to do what it claims, but review these before installing or running: 1) The API key (NANO_API_KEY) and images you send will go to whatever base URL you configure — confirm NANO_BASE_URL points to a trusted provider. 2) The skill reads ~/.openclaw/openclaw.json and a local config.json where an API key can be stored; don't put sensitive keys in files you don't trust. 3) There is no automatic installer — you must provide the uv runtime and install Python deps (openai, pillow, requests) yourself. 4) Inspect scripts/generate_image.py (included) before use; if you point base_url at a third-party service, that service will receive your API key and image data. If you want to be extra cautious, run the script with a read-only or scoped API key and test against a trusted endpoint first.
Capability Analysis
Type: OpenClaw Skill
Name: nano-banana-pro-custom
Version: 1.0.1
The skill bundle is a legitimate tool for generating and editing images using OpenAI-compatible APIs. The main script, scripts/generate_image.py, implements standard configuration loading from environment variables, local config files (~/.openclaw/openclaw.json), and command-line arguments. It uses the official openai Python library to interact with specified endpoints and includes robust image processing logic using Pillow. No evidence of data exfiltration, malicious execution, or prompt injection was found; the code is well-documented and its behavior aligns perfectly with its stated purpose.
Capability Assessment
Purpose & Capability
Name/description (OpenAI-compatible image generation) match the declared env vars (NANO_API_KEY, NANO_BASE_URL, NANO_MODEL), required binary (uv) and the included Python script. Required dependencies (openai, pillow, requests) are proportionate to image-generation functionality.
Instruction Scope
SKILL.md and the script limit behavior to: reading args/env/openclaw.json/skill config.json, loading image files, encoding them, and calling the configured image API. The config file paths the skill reads (~/.openclaw/openclaw.json and skill config.json) are declared in metadata and used by the script. No instructions tell the agent to read unrelated system files or exfiltrate data outside the configured base_url.
Install Mechanism
There is no automated install spec; the skill is instruction-only with an included script. This is low-risk: nothing in the registry metadata downloads or executes external installers. Dependencies must be fulfilled by the environment (Python packages listed), which is expected.
Credentials
Requested environment variables are limited to an API key, base URL, and model name — directly relevant to the stated purpose. The primary credential is NANO_API_KEY, which is appropriate. The script reads only these env vars and declared config files.
Persistence & Privilege
always is false and the skill does not request system-wide changes or to modify other skills. It only reads its own config files and the user's ~/.openclaw/openclaw.json per its declared configPaths.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install nano-banana-pro-custom - After installation, invoke the skill by name or use
/nano-banana-pro-custom - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.1
- Added python version and required package dependencies (openai, pillow, requests) to metadata.
- Documented python and package requirements directly in the skill metadata.
- Specified config file search paths for configuration in metadata.
- No changes to user-facing usage or behavior.
v1.0.0
Initial release of nano-banana-pro-custom
- OpenAI-compatible image generation and editing skill with multi-image and fine-tuned model support.
- Comprehensive multi-source configuration: CLI flags, environment variables, openclaw.json, and skill config.json (in priority order).
- Supports prompt-based image creation, single/multi-image editing, and OpenAI-style API endpoints.
- Includes detailed configuration and quick-start instructions in both CLI and openclaw.json formats.
- Compatible with OpenAI, Azure OpenAI, OpenRouter, and other OpenAI-style APIs.
- Advanced options for output size, image quality, response format, and batch generation.
Metadata
Frequently Asked Questions
What is Nano Banana Pro Custom?
Generate or edit images using OpenAI-compatible API. Supports multi-image input, fine-tuned models, and multiple configuration sources (env vars, openclaw.js... It is an AI Agent Skill for Claude Code / OpenClaw, with 168 downloads so far.
How do I install Nano Banana Pro Custom?
Run "/install nano-banana-pro-custom" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Nano Banana Pro Custom free?
Yes, Nano Banana Pro Custom is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Nano Banana Pro Custom support?
Nano Banana Pro Custom is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Nano Banana Pro Custom?
It is built and maintained by CHENXCHEN (@chenxchen); the current version is v1.0.1.
More Skills