higgsfield-image-gen

What problem does it solve?

This skill prevents generic, off-brand AI ad imagery by forcing structured, brand-aware briefs and a deliberate iteration loop that converges on production-ready visuals.

Core Features & Use Cases

• 8-field Higgsfield briefing: Collects subject, pose/action, setting, lighting, palette, aesthetic reference, composition, and aspect ratio to reduce drift and improve consistency.
• Per-brand aesthetic rules: Applies brand-specific do/don’t guidance for Abundria / Palm Aura and Pit Smith Co. (including example mood references and palette directives).
• Iteration loop for winners: Plans Round 1 (4 variants), Round 2 (refine picked variant(s) with deltas), and Round 3 (final polish) to avoid “get it right in one go” failure modes.
• Overlay text compliance guidance: Specifies generating imagery without burned-in text first, then reserving composition space for later typography with brand fonts and WCAG contrast intent.
• Safety and compliance guardrails: Blocks common ad-creation rejection triggers such as identifiable people without permission, medical-claim visuals, fake testimonials, and misleading before/after concepts.

Quick Start

Use higgsfield-image-gen to generate a 3-round Higgsfield brief and output plan for a specific brand, concept, and aspect ratio, then create 4 variants for Round 1.