Skill

banana

Generates or edits images via Google Gemini from text prompts or up to 14 reference images for text-to-image and multi-image editing.

Python

Gemini

Bash

ai-ml

npx claudepluginhub agentic-dev3o/devx-plugins --plugin nano-banana-ultimate

Tool Access

This skill is limited to using the following tools:

Bash(uv:*)Bash(ls:*)Read

Preview

**Input:** `$ARGUMENTS` (optional plain-language description of what to generate)

Supporting Assets

references/prompts.md

SKILL.md

Similar Skills

generate-image

247

Generates AI images from text prompts, edits images, and composes from multiple references using Gemini models. Supports t2i, i2i, product mockups, and stickers.

4 files3 tools

claude-content

Nano Banana Image Generation

Generate and edit images using Google Gemini models via nano-banana CLI. Supports text-to-image, editing, style transfer. Use for AI image creation or modification requests.

5 files

nano-banana

nanobanana-skill

1.5k

Generates or edits images using Google Gemini API via nanobanana.py Python script. Supports prompts, aspect ratios like 9:16, models such as gemini-3.1-flash, and resolutions up to 4K.

2 files9 tools

nanobanana-skill

Stats

Parent Repo Stars10

Parent Repo Forks1

Last CommitApr 5, 2026

Actions

View Source View Plugin View on GitHub View README

Help us improve

Share bugs, ideas, or general feedback.

Nano Banana Ultimate — Image Generation

Input: $ARGUMENTS (optional plain-language description of what to generate)

If $ARGUMENTS is provided, use it as the user's intent and skip to prompt building. If empty, ask a clarifying question.

Requirements

uv installed
GEMINI_API_KEY environment variable set (get one at https://aistudio.google.com/apikey)

Workflow

Understand — Parse the user's request. If the subject or intent is unclear, ask ONE clarifying question.
Build prompt — Construct a detailed Gemini prompt using the formula below.
Configure — Choose aspect ratio, resolution, and model. Use defaults unless the user specifies otherwise.
Generate — Run the script.
Deliver — Report the saved file path. Do NOT read the image file back. Offer to iterate: adjust style, composition, or prompt.

Prompt Formula

Build every prompt using this structure:

[Subject] + [Action/Pose] + [Environment/Location] + [Composition/Framing] + [Style/Medium]

Rules

Positive framing: Describe what IS in the image, never what is absent. Say "empty street" not "no cars".
Strong verb opener: Start with Generate, Create, Capture, Render, Design, Compose.
Be specific: Include concrete details — materials, lighting direction, time of day, textures.
Camera/lens (photorealistic): "shot on 85mm f/1.4 lens", "low angle", "aerial view".
Lighting: "golden hour side light", "studio softbox", "dramatic chiaroscuro", "overcast diffused".
Materiality: "brushed aluminum", "matte ceramic", "glossy lacquer", "translucent glass".
Text in images: Enclose any text to render in quotes. Specify font style. Note: text rendering may be imperfect.

Generation

uv run "${CLAUDE_PLUGIN_ROOT}/scripts/banana.py" \
  --prompt "<detailed prompt>" \
  --output "<YYYY-MM-DD-HH-MM-SS-descriptive-name>.png" \
  --resolution <1K|2K|4K> \
  --aspect-ratio <ratio> \
  --model <model>

Editing (with input images)

When the user provides an existing image to modify:

uv run "${CLAUDE_PLUGIN_ROOT}/scripts/banana.py" \
  --prompt "<edit instruction: what to change AND what to keep>" \
  --output "<output-path>.png" \
  --input-image "/path/to/source.png"

For edits, describe both the change and what must stay identical:

"Change the background to a sunset beach. Keep the subject, their clothing, and pose exactly the same."

Defaults

Parameter	Default
Resolution	1K
Aspect ratio	1:1
Model	gemini-3-pro-image-preview

Models

Model	Best for
`gemini-3-pro-image-preview`	Final output, complex scenes, high quality (Nano Banana Pro)
`gemini-3.1-flash-image-preview`	Fast drafts, iterations, simple images (Nano Banana 2)

Aspect Ratios

Ratio	Use case
1:1	Social squares, logos, icons
3:2	Photography standard
2:3	Book covers, tall portraits
4:3	Landscape, presentations
3:4	Portrait photos, Pinterest
16:9	YouTube thumbnails, widescreen
9:16	Stories, Reels, TikTok
4:5	Instagram portrait
5:4	Large format print
21:9	Ultrawide banners

Output Naming

Always use timestamped filenames: YYYY-MM-DD-HH-MM-SS-descriptive-name.png

Save to the current working directory unless the user specifies a path.

References

See references/prompts.md for ready-to-use example prompts covering landscapes, abstract art, interiors, sci-fi, nature, fantasy, still life, and underwater scenes.