Art-direction engine for text-to-image generation. Separates Style from Content to keep aesthetic identity consistent across an image series. Use when generating, rendering, or illustrating images (especially across a series), capturing the style of a reference image for reuse, or producing an image in a named/saved style. Not for captioning, image analysis for decisions (use ocas-look), or photo editing.
Skill name: ocas-imagine
Version: 1.0.5
Type:
Layer: creative
Author: Indigo Karasu
Art-direction engine for text-to-image generation. Separates Style from Content to keep aesthetic identity consistent across an image series. Use when generating, rendering, or illustrating images (especially across a series), capturing the style of a reference image for reuse, or producing an image in a named/saved style. Not for captioning, image analysis for decisions (use ocas-look), or photo editing.
imagine.generate --style <name|prompt> --content <description>— generate an image in a specific style.imagine.extract --image <path|url>— analyze an image and produce a structured Style Prompt.imagine.library.list— list all available predefined and saved styles.imagine.style.save --name <name> --prompt <prompt>— save a new custom style to the library.imagine.journal— read the most recent run record (final prompt + resulting image URL).
See SKILL.md for outputs, journals, and persistence rules.
| File | Purpose |
|---|---|
SKILL.md |
Skill definition |
references/ |
Supporting documentation |
scripts/ |
Helper scripts |
- [1.0.4] - 2026-04-14
- Changed
- [1.0.3] - 2026-04-14
- Changed
- Added
- [1.0.2] - 2026-04-14
- Added
- Changed
Read SKILL.md for operational details, schemas, and validation rules.
Read references/ for detailed specifications and examples.
MIT License — see LICENSE for details.