opticlisted

# OPTIC — AI Image Generation Pipeline *From prompt to pixel. Text-to-image, image-to-image, and surgical compositing.* Developed for the [jord0.skills](https://github.com/jord0-cmd/jord0.skills) ecosystem. OPTIC encodes the complete AI image generation pipeline: prompt engineering, model invocation, multi-pass refinement, and pixel-level compositing. These techniques were developed through extensive experimentation and produce dramatically better results than naive single-pass generation. **Requirements**: Python 3.10+, `google-genai` package, `GEMINI_API_KEY` environment variable. Optional: `Pillow` + `numpy` for inpainting pipeline. --- ## Usage Describe what you want to generate, edit, or composite. OPTIC handles prompt engineering, model selection, and compositing pipelines. See detailed techniques below. --- ## When to Invoke This Skill - Generating images from text descriptions (text-to-image) - Editing or transforming existing images (image-to-image) - Writing effective prompts for Gemini image models - Multi-pass scene generation (Sequential Grounding) - Surgical inpainting (crop, edit, colour match, composite) - Hand-drawn guide line techniques for geometry control - Any workflow where AI-generated imagery is a component --- ## What OPTIC Builds — Capability Lookup | Need | Technique | Section | |------|-----------|---------| | "Generate an image from description" | Text-to-image via Generator | The Generator | | "Edit th