Skip to content
Back to skills

Document Image State

ASecurity

"Guides Krita AI Diffusion document, layer, region, Image, Mask,

  • 247 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added September 8, 2026
documentationpythonbashdebuggingapi

Works with

  • api

Security analysis

A100/100

Pro scans all 5 files and shows the line behind each finding

Scanned September 8, 2026

npx -y skills add VectorSpaceLab/AREX-Skill --skill document-image-state --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Document Image State?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Document Image State
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/vectorspacelab-document-image-state/badge)](https://www.skillsdirectory.com/skills/vectorspacelab-document-image-state)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: document-image-state
description: "Guides Krita AI Diffusion document, layer, region, Image, Mask,
  Bounds, prompt, style, metadata, and persistence state tasks."
disable-model-invocation: true
metadata:
  disco-role: operating
license: GPL 3.0
---

# document-image-state

Use this sub-skill when the task is about Krita document/layer state, image and
mask helpers, prompt/style processing, LoRA tags, regions, persistence, or how
canvas data enters/leaves workflows.

Trigger examples:

- Converting between Krita pixel data and `ai_diffusion.image.Image`, `Mask`,
  `Bounds`, `Extent`, or `ImageCollection`.
- Debugging active selection, layer bounds, region masks, control layers, or
  apply behavior after generation.
- Understanding prompt comments, `{wildcards}`, `<lora:name:weight>`,
  `<layer:name>` tokens, style prompt merge, negative prompt merge, or PNG
  metadata.
- Working with `Style`, `Styles`, sampler presets, settings JSON, document
  persistence, or observable persisted properties.

## Safe entry point

```bash
python sub-skills/document-image-state/scripts/inspect_prompt_style.py --prompt "cat <lora:fur:0.6> # note" --style-prompt "cinematic {prompt}" --lora-id fur --metadata
```

The helper analyzes prompt/style behavior offline. It does not launch Krita,
connect to ComfyUI, read private images, download models, or run generation.

## References

- [references/document-layer-image-reference.md](references/document-layer-image-reference.md):
  document/layer/image/mask/bounds APIs and data flow.
- [references/prompt-style-persistence.md](references/prompt-style-persistence.md):
  prompt, style, LoRA, wildcard, metadata, and persistence behavior.
- [references/troubleshooting.md](references/troubleshooting.md): selection,
  layer, image, prompt, style, metadata, and persistence failure recovery.
- [scripts/inspect_prompt_style.py](scripts/inspect_prompt_style.py): bundled
  offline prompt/style inspector.

## Boundaries

- For final request dataclasses and workflow cost/pass behavior, route to
  `inference-workflows`.
- For UI controls and workspace state that decide when document data is
  collected, route to `ui-workspaces`.
- For custom graph placeholders that request layers/masks/canvas data, route to
  `custom-graphs` too.

Files in this skill

  • SKILL.md2.2 KB
  • references/document-layer-image-reference.md3.6 KB
  • references/prompt-style-persistence.md3.2 KB
  • references/troubleshooting.md2.9 KB
  • scripts/inspect_prompt_style.py6.5 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.