Generate images with Google's Nano Banana Pro (Gemini 3 Pro Image) using Application Default Credentials (ADC) or a GEMINI_API_KEY. Use this when the user asks to create, generate, draw, or render an image via Google / Vertex AI / Gemini.
Scanned 5/27/2026
Install via CLI
openskills install hummer98/nanobanana-adc---
name: nanobanana-adc
description: Generate images with Google's Nano Banana Pro (Gemini 3 Pro Image) using Application Default Credentials (ADC) or a GEMINI_API_KEY. Use this when the user asks to create, generate, draw, or render an image via Google / Vertex AI / Gemini.
---
# nanobanana-adc
Image generation CLI built on Gemini 3 Pro Image (Nano Banana Pro). Unlike other
Claude Code image skills, this one works with **Application Default Credentials
(ADC)** on Vertex AI — required in enterprise / CI / Cloud Run / Cloud Build
environments where `GEMINI_API_KEY` is not available.
## When to use
Trigger this skill when the user wants to:
- Create / generate / draw / render an image via Google or Vertex AI.
- Use Gemini 3 Pro Image (`gemini-3-pro-image-preview`) from a corporate GCP
project that uses ADC instead of a bare API key.
- Fall back to `GEMINI_API_KEY` when ADC is not set up.
## Quick examples
```bash
# ADC mode (default — uses gcloud / metadata server)
nanobanana-adc --prompt "a cat astronaut on mars, cinematic" --output cat.png
# 16:9 2K
nanobanana-adc --prompt "futuristic tokyo skyline" --aspect 16:9 --size 2K -o tokyo.png
# API key mode (explicit override)
nanobanana-adc --prompt "..." --api-key "$GEMINI_API_KEY" -o out.png
```
## Options
| Flag | Default | Notes |
|-----------------|------------------------------|-----------------------------------------------------------------------------------|
| `-p, --prompt` | (required) | Prompt text. |
| `-o, --output` | `output.png` | Output file path. |
| `-a, --aspect` | `1:1` | `1:1` / `16:9` / `9:16` / `4:3` / `3:4` / `3:2` / `2:3` / `21:9` / `9:21` / `5:4`. |
| `-s, --size` | `1K` | `1K` / `2K` / `4K`. |
| `-m, --model` | `gemini-3-pro-image-preview` | Override model id. |
| `--api-key` | — | Gemini API key. Falls back to `GEMINI_API_KEY` then ADC when omitted. |
## Environment variables
| Variable | Required (ADC) | Purpose |
|----------------------------------|----------------|--------------------------------------------------|
| `GOOGLE_CLOUD_PROJECT` | ✓ | GCP project id. |
| `GOOGLE_CLOUD_LOCATION` | ✓ | Region (e.g. `us-central1`). |
| `GOOGLE_GENAI_USE_VERTEXAI` | ✓ | Must be `true`. |
| `GOOGLE_APPLICATION_CREDENTIALS` | optional | Path to ADC key file (else falls back to gcloud).|
| `GEMINI_API_KEY` | fallback | Used when ADC is not configured. |
## Authentication priority
1. `--api-key` flag
2. `GEMINI_API_KEY` env var
3. ADC (`google-auth-library`)
## Troubleshooting
- **`Could not load the default credentials`** — run `gcloud auth application-default login` and ensure `GOOGLE_CLOUD_PROJECT` / `GOOGLE_CLOUD_LOCATION` / `GOOGLE_GENAI_USE_VERTEXAI=true` are set.
- **`PERMISSION_DENIED`** — ensure the ADC principal has `roles/aiplatform.user` on the project, and that Vertex AI image generation is enabled.
- **No image returned / unexpected model error** — confirm the model id (`--model gemini-3-pro-image-preview`) is available in your region, or explicitly fall back to an API key with `--api-key "$GEMINI_API_KEY"`.
No comments yet. Be the first to comment!