Skip to content

cumabozkurt/god-of-design / skills/god-imagegen

god-imagegen

Verified

Illustration and image-generation prompting for Midjourney, GPT Image, Gemini/Imagen, FLUX, Stable Diffusion, Ideogram, Recraft and Firefly - prompt anatomy, style vocabulary (techniques, photography, lighting, camera, materials), aspect ratios, consistency, negative prompts, text-in-image, editing and ethics. Use when the user needs illustrations, photos, mockups, textures or reference boards generated by AI.

God Image Generation & Illustration

Prompt anatomy (works across models)

[Subject + action] , [setting/context] , [style/medium] , [composition/camera] ,
[lighting] , [colour palette] , [mood] , [details/materials] ,
[technical: aspect ratio, quality]

Example: An elderly ceramicist glazing an İznik-style tulip plate, sunlit workshop in Kütahya, documentary photography, 50mm lens at eye level, shallow depth of field, soft window light from the left, cobalt and turquoise with warm wood tones, calm and focused, visible glaze drips and brush texture, 4:5

Model notes (late 2026; check current docs because they change fast)

ModelStrengthsPrompting tips
Midjourney (v7+)Aesthetic quality, stylesShort, evocative phrases; --ar 4:5, --stylize, --sref style refs, --cref/omni-reference for characters; --no text
GPT Image (ChatGPT)Instruction following, text rendering, editingNatural sentences; describe the layout explicitly; iterate conversationally; good for mockups with text
Google Gemini image (Nano Banana family) / ImagenEditing, consistency, multi-image fusion, textDescribe edits precisely ("keep the face, change background to…"); supports reference images
FLUX (Black Forest Labs)Photorealism, prompt adherence, open weights (dev/schnell)Detailed natural language; LoRAs for brand styles
Stable Diffusion / SDXLOpen, controllable (ControlNet, LoRA, inpainting)Weighted tags (keyword:1.2), negative prompts, ControlNet for pose/layout
IdeogramTypography in images, postersPut the exact text in quotes; specify the font style
RecraftVector/SVG output, brand styles, iconsAsk for "vector illustration, flat"; export SVG
Adobe FireflyCommercially safe training data, Photoshop integrationUse for client work needing indemnity

Style vocabulary

  • Illustration techniques: flat vector, line art, risograph, screen print, linocut, woodcut, engraving/etching, gouache, watercolour, ink wash (sumi-e), pastel, charcoal, collage, paper cut, claymation, isometric, low-poly, pixel art, cel-shaded anime, children's-book, botanical illustration, technical cutaway, blueprint, editorial conceptual.
  • Photography: documentary, editorial, studio product, macro, aerial/drone, street, fashion, still life, food (overhead flat lay / 45°), architectural (tilt-shift).
  • Lighting: golden hour, blue hour, overcast soft, hard noon sun, Rembrandt, rim light, backlit silhouette, softbox studio, neon, candlelight, volumetric fog, chiaroscuro.
  • Camera: 24mm wide (space), 35mm (story), 50mm (natural), 85mm (portrait), 100mm macro, aperture f/1.8 bokeh vs f/11 deep focus, low angle (power), high angle (vulnerability), Dutch tilt (tension), bird's-eye, worm's-eye.
  • Materials: brushed aluminium, frosted glass, ceramic glaze, linen, terrazzo, travertine, walnut, felt, chrome, iridescent film, recycled paper.
  • Styles from the atlas: every god-styles entry has a Prompt fragment. Combine it with your subject.

Workflow for design work

  1. Generate reference boards (4–8 variations) to lock the direction; pick one.
  2. Lock consistency with style references / seeds / LoRA / character references.
  3. Leave negative space where text will go. Add text afterwards in HTML/SVG/Figma (crisp, editable, correctly spelled, especially Turkish ı/İ/ş/ğ, which models often get wrong).
  4. Upscale and clean up (remove artefacts, fix hands and text).
  5. For illustration systems, define: line weight, colour palette (hex), shading method, character proportions, perspective. Then prompt every image with the same "style block".

Ethics and legal

  • Don't prompt for living artists' names to imitate their signature style for commercial work. Describe the technique instead ("bold flat shapes, limited palette, risograph texture").
  • No real people's likenesses without consent. No trademarks or logos you don't own.
  • Disclose AI generation where platforms or laws require it (EU AI Act transparency duties, platform labels).
  • Don't generate sacred or First Nations art (see god-styles world-traditions protocol). Commission artists instead.
  • Check each model's commercial-use terms.

See references/prompt-library.md for ready prompts by design area.

reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall