skills.rest
fralapo
fralapoCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

image-gen-prompts

Generate image prompts with model-specific syntax and structured constraints.

1|Updated Apr 15, 2026
Search Tags:#image editing#model routing#structured prompts#identity preservation#image prompt#inpainting workflow
One-click install
npx skills add https://github.com/fralapo/awesome-agent-skills --skill image-gen-prompts
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: image-gen-prompts
Source: https://github.com/fralapo/awesome-agent-skills/tree/main/skills/image-gen-prompts
Command: npx skills add https://github.com/fralapo/awesome-agent-skills --skill image-gen-prompts

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

It solves the problem of generating high-quality, consistent image prompts across many different image generation and editing models without losing control over subject, edits, identity, or formatting.

Core Features & Use Cases

  • Universal prompt engineering for image generation/editing: Provides a shared prompt anatomy (subject, environment, lighting, camera/lens, style, format, preserve/negative constraints) that works across major models.
  • Model-aware routing and syntax guidance: Automatically switches to per-model reference files (e.g., Midjourney flags, GPT Image 2 structured prompts, Nano Banana natural-language reference handling) when the user names a model.
  • Editing workflows with non-destructive identity preservation: Supports background swaps, inpainting/masked edits, object add/remove, and multi-image fusion with explicit “preserve” clauses.
  • Structured prompts, templating, and typography reliability: Includes guidance for JSON/YAML/XML prompt formats and stronger text rendering modes when you need in-image labels or posters.

Quick Start

Use the image-gen-prompts skill to create an identity-preserving edit prompt: keep the face and pose from the uploaded photo while changing only the background to a cinematic golden-hour city street, for GPT Image 2.

Frequently Asked Questions about image-gen-prompts

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I write structured prompts for image generation across different models?▼

To create an inpainting prompt for object removal, define the subject and enforce explicit preserve-only change regions to protect the background, while using model-appropriate structured or prose prompt formats to specify the masked edit area for the target image model.

What is the best way to preserve facial identity when swapping image backgrounds?▼

Preserving facial identity during background swaps requires explicit preserve clauses in your image prompt that lock the face and pose from reference photos, allowing non-destructive edits by enforcing constraints that only modify the targeted background region.

Can I use structured JSON or YAML prompt formats for typography and text rendering?▼

Yes, structured JSON, YAML, and XML prompt formats support typography-oriented layouts by providing stronger text rendering modes, enabling reliable in-image labels and posters when generating images with text-heavy visual elements.

How does multi-image fusion work with natural-language reference handling?▼

Multi-image fusion works by combining multiple reference images using natural-language instructions, where the model-specific syntax routing translates your shared prompt anatomy into the correct structured constraints for the designated image generation model.

When do I need model-aware syntax routing for image editing workflows?▼

Model-aware syntax routing for image editing is needed when switching between major models for background swaps, inpainting, or object add/remove tasks, ensuring your structured prompts use the correct flags and reference handling syntax for each specific platform.

Related Skills

Explore More Design & Creative Skills→
packetalienpacketalien

art-bible

Coordinate the Cursor art bible workflow across Phases 1–9 to draft visual standards.

Community
Advanced
daanteijema-beepdaanteijema-beep

banner-creator

Generate AI banner images and crop them to platform-specific aspect ratios.

Community
Intermediate
agentspace-soagentspace-so
31

image-inpainting

Perform mask-driven image inpainting to remove or replace image regions.

Official
Intermediate
runcomfy-comruncomfy-com
12

ai-image-generation

Generate and edit images from natural-language prompts via RunComfy endpoints.

Official
Advanced