What problem does it solve? AI-generated images often look generic because no visual brand identity exists, so every image prompt starts from scratch and produces inconsistent results across sessions. ## Core Features & Use Cases - Three Definition Modes: Extract a visual style from an existing website URL, Build one through a structured interview, or synthesize one from mood boards and reference images. - Persistent Brand Memory: Writes a Visual Brand Style section (aesthetic, lighting, composition, mood, avoidances, reference prompts) into brand/creative-kit.md so downstream skills like /image-gen produce on-brand visuals. - Progressive Enhancement: Reads voice-profile.md and positioning.md when available to align visuals with brand personality and positioning angles. - Use Case: A startup founder runs the skill, answers five questions about aesthetic, lighting, and composition, and receives ready-to-paste reference prompts that make every future generated image recognizably on-brand. ## Quick Start Ask the agent to define your brand's visual style by analyzing your website or interviewing you about your aesthetic preferences.