# Generate Images from Natural Language

Creating image requests manually requires prompt formatting and API setup. This skill converts text or image references into requests for a third-party image generation API.

## Install

```bash
npx skillstore add ceeon/gemini-image
```

## Metadata

- Status: approved
- Slug: ceeon-gemini-image
- Skillstore revision: r3
- Version status: missing
- Tree hash: c36295c20a304b5f0f0ba6671994113474c6c1a83107dfc2511ab7faadfd0fc5
- Author: Ceeon
- GitHub username: Ceeon
- License: MIT
- Repository: https://github.com/Ceeon/gemini-image-skill/tree/main/
- Ref: c43861a65bb95efcae259cd161c9d6f4dc7eec6f
- Supported tools: Claude, Codex, Claude Code
- Audit status: complete
- Agent install advisory: confirmation\_required
- Manual install advisory: allowed
- Artifact signature: available
- Audit attestation: unavailable
- Human verification: not\_verified
- Risk factors: filesystem, external\_commands, network, env\_access
- Quality score: 38
- Quality tier: warning
- Public page: https://skillstore.pages.dev/skills/ceeon-gemini-image
- Manifest: https://skillstore.pages.dev/api/skills/ceeon-gemini-image/manifest

## Capabilities

- Formats natural-language descriptions as text-to-image prompts.
- Accepts one image URL for image-to-image generation.
- Accepts multiple image URLs as visual references.
- Posts generation requests to the api.apicore.ai endpoint with curl.
- Extracts the first generated image URL from the API response.
- Documents three Gemini preview model variants and ten aspect ratios.

## Use Cases

- Create Concept Illustrations: Generate visual concepts from concise descriptions for early design exploration.
- Produce Campaign Variations: Create image variations in specified aspect ratios for social and promotional assets.
- Prototype Image Workflows: Test text-to-image and reference-image requests through a conversational agent.

## Prompt Templates

### Generate a Simple Image

```
Generate an image of [subject] in [style].
```

### Set Composition and Ratio

```
Create a [aspect ratio] image of [subject], with [lighting], [composition], and [color direction].
```

### Transform a Reference Image

```
Use [image URL] as a reference. Recreate it with [requested changes] while preserving [important elements].
```

### Combine Multiple References

```
Use [first image URL] and [second image URL] as references. Combine [element one] with [element two] using [model and aspect ratio].
```

## Limitations

- Requires a third-party API key and outbound network access.
- Referenced configuration, secret, and upload guidance files are absent from the reviewed package.
- Image-based requests require accessible image URLs before generation.
- The instructions do not define error handling, retries, content moderation, or response validation.

## Best Practices

- Describe the subject, style, lighting, composition, and aspect ratio explicitly.
- Use accessible image URLs and confirm that sharing them with the API is permitted.
- Review generated images for accuracy, rights, privacy, and policy compliance before use.

## Anti Patterns

- Do not place real API keys directly in prompts, chat messages, or version-controlled files.
- Do not send private, licensed, or sensitive images without authorization.
- Do not assume placeholder models, files, or endpoints are available without verification.

## Security Audit

- Audited at: 2026-07-23T13:14:38.163\+00:00
- Summary: Most static alerts are false positives caused by Markdown code spans, placeholder URLs, standard installation paths, and plain README content. Confirmed behavior includes shell execution, external API requests, and transmission of an API credential to api.apicore.ai. The differing registration and API domains create an unclear credential trust boundary, and the registration link contains an undisclosed affiliate parameter.

## Stats

- Views: 204
- Downloads: 75
- Favorites: 1
- Popularity score: 0