design-image-prompt-engineer

Installation
SKILL.md

Image Prompt Engineering Guide

Overview

This guide covers how to translate visual concepts into precise, structured prompts for AI image generation platforms. It bridges real photography knowledge (lens selection, lighting setups, composition) with platform-specific syntax for Midjourney, DALL-E, Stable Diffusion, and Flux.

The Five-Layer Framework

Every prompt should address these five layers in order:

  1. Subject: Who/what, specific attributes, pose, expression, materials, textures.
  2. Environment: Location, background treatment, atmospheric conditions, time of day.
  3. Lighting: Source type, direction, quality (hard/soft), color temperature.
  4. Camera: Focal length, aperture, angle, depth of field, exposure style.
  5. Style: Genre, era, post-processing, color grade, reference photographer.

Use correct photography terminology throughout: "shallow depth of field, f/1.8 bokeh" not "blurry background." Keep lighting direction consistent with shadow descriptions within the same prompt. Include negative prompts when the platform supports them to exclude unwanted artifacts.

When a requested effect is physically implausible in real photography, flag it and suggest a plausible alternative.

Related skills

More from peterhdd/agent-skills

Installs
3
GitHub Stars
8
First Seen
Mar 4, 2026