# Generate AI Images with Inference.sh

Image teams need quick access to many generation and editing models without writing provider-specific workflows. This skill supplies belt CLI examples for text-to-image, edits, upscaling, and model selection.

## Install

```bash
npx skillstore add inference-skills/ai-image-generation
```

## Metadata

- Status: approved
- Slug: inference-skills-ai-image-generation
- Skillstore revision: r1
- Version status: missing
- Tree hash: 2541dbb4377bfc113c1653482af66a00a1b7668a5e6642d192b77cd7fb85a028
- Author: inference-skills
- GitHub username: inference-skills
- License: MIT
- Repository: https://github.com/inference-skills/skills/tree/main/tools/image/ai-image-generation/
- Ref: a25199bc7d6b82598536822d1738eb5d5f54025b
- 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: 69
- Public page: https://skillstore.pages.dev/skills/inference-skills-ai-image-generation
- Manifest: https://skillstore.pages.dev/api/skills/inference-skills-ai-image-generation/manifest

## Capabilities

- Runs belt CLI examples for GPT-Image-2, FLUX, Gemini, Grok, Seedream, Reve, and other image apps.
- Shows text-to-image prompts with quality, aspect ratio, and model-specific options.
- Shows image editing and inpainting prompts using supplied image URLs.
- Lists examples for upscaling and stitching image outputs.
- Points users to model discovery and inference.sh documentation.

## Use Cases

- Create Campaign Visuals: Generate draft and final images for ads, landing pages, and social channels using model-specific prompts.
- Prototype Product Mockups: Create studio-style product photos, edit backgrounds, and upscale selected outputs for review.
- Compare Image Models: Test GPT-Image-2, FLUX, Gemini, Seedream, Reve, and Grok prompts from one CLI workflow.

## Prompt Templates

### Create a Basic Image

```
Generate a square image of [subject] in [style] with [lighting] and [background]. Use a fast general image model.
```

### Edit a Product Photo

```
Use this image URL: [image URL]. Change [specific element] to [new result] while preserving shape and perspective.
```

### Compare Model Options

```
Recommend a model for [goal], [aspect ratio], [quality level], and [budget]. Then provide the belt command to run.
```

### Build a Campaign Set

```
Create prompts for [campaign] across [formats]. Include model choices, aspect ratios, and editing steps for consistent visual style.
```

## Limitations

- Requires an installed and authenticated inference.sh belt CLI.
- Sends prompts and referenced images to external inference services.
- Does not include local model weights or offline generation.
- Model availability, pricing, and output quality depend on provider settings.

## Best Practices

- Confirm users approve sending prompts and image references to external services.
- Use specific subject, style, lighting, camera, aspect ratio, and quality instructions.
- Start with fast models for drafts, then use higher quality models for final images.

## Anti Patterns

- Do not include private or regulated images unless the user approves third-party processing.
- Do not ask for vague images when brand, style, or aspect ratio matters.
- Do not install related skills automatically from the examples without explicit user approval.

## Security Audit

- Audited at: 2026-07-05T17:00:52.999\+00:00
- Summary: Most static command-execution alerts are Markdown backtick and code-fence false positives; no prompt injection or hidden executable logic was found in SKILL.md. The skill still sends prompts and image references to third-party inference services and embeds one remote image asset.

## Stats

- Views: 75
- Downloads: 10
- Favorites: 0
- Popularity score: 0