imagegen

Source excerpt starting at line 1.

1# Prompting best practices2 3These prompting principles are shared by both top-level modes of the skill:4- built-in `image_gen` tool (default)5- explicit `scripts/image_gen.py` CLI fallback6 7This file is about prompt structure, specificity, and iteration. Fallback-only execution controls such as `quality`, `input_fidelity`, masks, output format, and output paths live in the fallback docs.8 9## Contents10- [Structure](#structure)11- [Specificity policy](#specificity-policy)12- [Allowed and disallowed augmentation](#allowed-and-disallowed-augmentation)13- [Composition and layout](#composition-and-layout)14- [Constraints and invariants](#constraints-and-invariants)15- [Text in images](#text-in-images)16- [Input images and references](#input-images-and-references)17- [Iterate deliberately](#iterate-deliberately)18- [Transparent images](#transparent-images)19- [Fallback-only execution controls](#fallback-only-execution-controls)20- [Use-case tips](#use-case-tips)21- [Where to find copy/paste recipes](#where-to-find-copypaste-recipes)22 23## Structure24- Use a consistent order: scene/backdrop -> subject -> key details -> constraints -> output intent.25- Include intended use (ad, UI mock, infographic) to set the level of polish.26- For complex requests, use short labeled lines instead of one long paragraph.27 28## Specificity policy29- If the user prompt is already specific and detailed, normalize it into a clean spec without adding creative requirements.30- If the prompt is generic, you may add tasteful detail when it materially improves the output.31- Treat examples in `sample-prompts.md` as fully-authored recipes, not as the default amount of augmentation to add to every request.32- For photorealism, include `photorealistic` directly when that is the goal, plus concrete real-world texture such as pores, wrinkles, fabric wear, material grain, or imperfect everyday detail.33 34## Allowed and disallowed augmentation35 36Allowed augmentation for generic prompts:37- composition and framing cues38- intended-use or polish-level hints39- practical layout guidance40- reasonable scene concreteness that supports the request41 42Do not add:43- extra characters, props, or objects that are not implied44- brand palettes, slogans, or story beats that are not implied45- arbitrary side-specific placement unless the surrounding layout supports it46 47## Composition and layout48- Specify framing and viewpoint (close-up, wide, top-down) and placement only when it materially helps.49- Call out negative space if the asset clearly needs room for UI or copy.50- Avoid making left/right layout decisions unless the user or surrounding layout supports them.51- For people, describe body framing, scale, gaze, and object interactions when they matter (`full body visible`, `looking down at the book`, `hands naturally gripping the handlebars`).52 53## Constraints and invariants54- State what must not change (`keep background unchanged`).55- For edits, say `change only X; keep Y unchanged` and repeat invariants on every iteration to reduce drift.56 57## Text in images58- Put literal text in quotes or ALL CAPS and specify typography (font style, size, color, placement).59- Spell uncommon words letter-by-letter if accuracy matters.60- For in-image copy, require verbatim rendering and no extra characters.61- In CLI fallback mode, use `medium` or `high` quality for small text, dense infographics, data-heavy slides, multi-font layouts, legends, axes, and footnotes.62 63## Input images and references64- Do not assume that every provided image is an edit target.65- Label each image by index and role (`Image 1: edit target`, `Image 2: style reference`).66- If the user provides images for style, composition, or mood guidance and does not ask to modify them, treat the request as generation with references.67- If the user asks to preserve an existing image while changing specific parts, treat the request as an edit.68- For compositing, describe how the images interact (`place the subject from Image 2 into Image 1`).69 70## Iterate deliberately71- Start with a clean base prompt, then make small single-change edits.72- Re-specify critical constraints when you iterate.73- Prefer one targeted follow-up at a time over rewriting the whole prompt.74 75## Transparent images76- Ask built-in `image_gen` for a genuinely transparent background and preserve its alpha.77 78## Fallback-only execution controls79- `quality`, `input_fidelity`, explicit masks, output format, and output paths are fallback-only execution controls.80- Do not assume they are built-in `image_gen` tool arguments.81- If the user explicitly chooses CLI fallback, see `references/cli.md` and `references/image-api.md` for those controls.82- In CLI fallback mode, `gpt-image-2` is the default. It supports `quality=low|medium|high|auto`; use `low` for fast drafts and thumbnails, and move to `medium`, `high`, or `auto` for final assets.83- `gpt-image-2` always uses high fidelity for image inputs, so do not set `input_fidelity` with that model.84- CLI `gpt-image-2` does not support `background=transparent`; ask before using `gpt-image-1.5` unless the user explicitly requested that model.85- If the user asks for 4K-style output with `gpt-image-2`, use `3840x2160` for landscape or `2160x3840` for portrait.86 87## Use-case tips88Generate:89- photorealistic-natural: Prompt as if a real photo is captured in the moment; use photography language (lens, lighting, framing); call for real texture; avoid over-stylized polish unless requested.90- product-mockup: Describe the product/packaging and materials; ensure clean silhouette and label clarity; if in-image text is needed, require verbatim rendering and specify typography.91- ui-mockup: Describe the target fidelity first (shippable mockup or low-fi wireframe), then focus on layout, hierarchy, and practical UI elements; avoid concept-art language.92- infographic-diagram: Define the audience and layout flow; label parts explicitly; require verbatim text; prefer higher quality in CLI mode for dense labels.93- logo-brand: Keep it simple and scalable; ask for a strong silhouette and balanced negative space; avoid decorative flourishes unless requested.94- ads-marketing: Write like a creative brief; include brand positioning, audience, desired vibe, scene, and exact tagline if text must appear.95- productivity-visual: Name the exact artifact (slide, chart, workflow diagram), define the canvas and hierarchy, provide real labels/data, and ask for readable typography and polished spacing.96- scientific-educational: Define audience, lesson objective, required labels, scientific constraints, arrows, and scan-friendly whitespace.97- illustration-story: Define panels or scene beats; keep each action concrete.98- stylized-concept: Specify style cues, material finish, and rendering approach (3D, painterly, clay) without inventing new story elements.99- historical-scene: State the location/date and required period accuracy; constrain clothing, props, and environment to match the era.100 101Edit:102- text-localization: Change only the text; preserve layout, typography, spacing, and hierarchy; no extra words or reflow unless needed.103- identity-preserve: Lock identity (face, body, pose, hair, expression); change only the specified elements; match lighting and shadows.104- precise-object-edit: Specify exactly what to remove/replace; preserve surrounding texture and lighting; keep everything else unchanged.105- lighting-weather: Change only environmental conditions (light, shadows, atmosphere, precipitation); keep geometry, framing, and subject identity.106- background-extraction: Request a clean cutout on a genuinely transparent background; preserve fine edges and label text; no halos or restyling.107- style-transfer: Specify style cues to preserve (palette, texture, brushwork) and what must change; add `no extra elements` to prevent drift.108- compositing: Reference inputs by index; specify what moves where; match lighting, perspective, and scale; keep the base framing unchanged.109- sketch-to-render: Preserve layout, proportions, and perspective; choose materials and lighting that support the supplied sketch without adding new elements.110 111## Where to find copy/paste recipes112For copy/paste prompt specs (examples only), see `references/sample-prompts.md`. This file focuses on principles, specificity, and iteration patterns.113