# Generate and Edit Images with Gemini

Creating consistent images requires model selection and precise command inputs. This skill prepares and runs Gemini image workflows through the inference.sh CLI.

## Install

```bash
npx skillstore add inferen-sh/nano-banana
```

## Metadata

- Status: approved
- Slug: inferen-sh-nano-banana
- Skillstore revision: r2
- Version status: missing
- Tree hash: 37156bd66ad4487376d3760d5fe32e3418f5d468eda3f6412f35e1b3b60a87b0
- Author: inferen-sh
- GitHub username: inferen-sh
- License: MIT
- Repository: https://github.com/inferen-sh/skills/tree/main/tools/image/nano-banana/
- Ref: 4121de961d1b6f2ffca856260e239505c302452c
- Supported tools: Claude, Codex, Claude Code
- Audit status: complete
- Agent install advisory: allowed
- Manual install advisory: allowed
- Artifact signature: available
- Audit attestation: unavailable
- Human verification: not\_verified
- Risk factors: external\_commands, network
- Quality score: 50
- Quality tier: warning
- Public page: https://skillstore.pages.dev/skills/inferen-sh-nano-banana
- Manifest: https://skillstore.pages.dev/api/skills/inferen-sh-nano-banana/manifest

## Capabilities

- Generates images from text prompts with supported Gemini image models.
- Edits an existing image from a supplied image URL and instruction.
- Requests multiple image variations in one operation.
- Configures aspect ratio, resolution, and output format.
- Enables Google Search grounding for current information.
- Creates sample input files for configurable CLI workflows.

## Use Cases

- Create Campaign Visuals: Generate several branded concepts with defined colors, composition, and aspect ratios.
- Prepare Product Mockups: Edit product images by changing backgrounds, lighting, or surrounding elements.
- Prototype Editorial Art: Produce high-resolution illustrations grounded in current information for review.

## Prompt Templates

### Generate a Basic Image

```
Generate a photorealistic image of [subject] in [setting] with [lighting]. Use a square aspect ratio.
```

### Create Multiple Concepts

```
Create four logo concepts for [brand] using [colors] and a minimalist style. Use the Gemini 2.5 Flash Image model.
```

### Edit a Source Image

```
Edit [image URL] by adding [change] while preserving composition, identity, and lighting. Return one image in 16:9 format.
```

### Create a Grounded 4K Scene

```
Use Gemini 3 Pro Image to create a 4K [aspect ratio] scene grounded in current information about [topic].
```

## Limitations

- Requires the inference.sh belt CLI, authentication, and network access.
- Model access, pricing, speed, and output quality depend on external services.
- Source images must be accessible to the remote service.
- The skill does not verify copyright, privacy, or factual accuracy.

## Best Practices

- Confirm the selected model, image count, resolution, and expected cost before execution.
- Describe subject, style, composition, lighting, colors, and required details explicitly.
- Use only source images that may be lawfully transmitted to external services.

## Anti Patterns

- Do not submit confidential images, personal data, credentials, or private URLs.
- Do not enable search grounding when current information is unnecessary.
- Do not request high resolution or multiple images before validating the prompt.

## Security Audit

- Audited at: 2026-08-06T10:40:18.891\+00:00
- Summary: The skill legitimately invokes the belt CLI and remote Gemini image services, which can transmit prompts and source images externally. Most static findings are Markdown formatting, model identifiers, placeholder URLs, or documentation links without executable behavior.

## Stats

- Views: 227
- Downloads: 13
- Favorites: 0
- Popularity score: 0