# Generate and Edit Images with P-Image

Image teams need fast generation and editing across several visual workflows. This skill runs Pruna P-Image models through the inference.sh belt CLI.

## Install

```bash
npx skillstore add inferen-sh/p-image
```

## Metadata

- Status: approved
- Slug: inferen-sh-p-image
- Skillstore revision: r2
- Version status: missing
- Tree hash: f528b3468fc6ec7b4b5f47e94d04996a1503071fd0ad7ae8673d2021c0096130
- Author: inferen-sh
- GitHub username: inferen-sh
- License: MIT
- Repository: https://github.com/inferen-sh/skills/tree/main/tools/image/p-image/
- Ref: 4121de961d1b6f2ffca856260e239505c302452c
- 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: external\_commands, network
- Quality score: 38
- Quality tier: warning
- Public page: https://skillstore.pages.dev/skills/inferen-sh-p-image
- Manifest: https://skillstore.pages.dev/api/skills/inferen-sh-p-image/manifest

## Capabilities

- Generates images from text prompts with the P-Image model.
- Applies listed LoRA presets through P-Image-LoRA.
- Edits supplied images using natural-language instructions.
- Combines multiple image URLs into one composition.
- Selects standard aspect ratios and lists other Pruna image apps.

## Use Cases

- Create Product Visuals: Generate square studio images with controlled lighting and clear product details.
- Edit Campaign Assets: Replace backgrounds or restyle existing images while preserving the source subject.
- Prototype Visual Concepts: Compare aspect ratios, style presets, and models during rapid design exploration.

## Prompt Templates

### Generate a Basic Image

```
Create an image of [subject] in [setting] with [lighting] using P-Image.
```

### Apply a Style Preset

```
Generate [subject] at [aspect ratio] with the [LoRA preset] style and [visual details].
```

### Edit an Existing Image

```
Edit [image URL] by changing [target area] to [new appearance] while preserving [elements].
```

### Build a Multi-Image Composite

```
Combine [image URLs] into [composition], match [lighting and palette], and use [aspect ratio].
```

## Limitations

- Requires the belt CLI, an inference.sh login, and network access.
- Sends prompts and image references to an external inference service.
- Documents built-in LoRA presets but does not show custom LoRA loading.
- Does not validate output quality, source-image rights, or provider content policies.

## Best Practices

- Describe the subject, setting, lighting, composition, and desired aspect ratio.
- Use shell-safe serialization for every prompt and image URL.
- Review provider terms and remove sensitive data before remote processing.

## Anti Patterns

- Do not place secrets, credentials, or private URLs in prompts.
- Do not concatenate untrusted text directly into shell commands.
- Do not assume a style preset preserves every source-image detail.

## Security Audit

- Audited at: 2026-08-06T10:46:49.709\+00:00
- Summary: Most static detections are Markdown backticks, fenced examples, or documentation links and are false positives. Unpinned skill installation commands and remote image loading remain real risks. The command templates also require safeguards against shell injection and disclosure of sensitive prompts or image URLs.

## Stats

- Views: 146
- Downloads: 9
- Favorites: 0
- Popularity score: 0