What problem does it solve?
Creative teams need multiple visual directions before committing to campaign assets, but generating isolated chat images produces no reviewable, reusable board. This Skill turns a light brief into a structured mood board of 16-30 generated visual references with intake qualification, saved run state, and a build handoff.
Core Features & Use Cases
- Intake-Gated Board Creation: Runs a qualification scale (Level 0-3) via an empty-board MCP intake widget or chat questions before writing the board spec.
- Batch Image Generation: Writes a JSON spec with per-item prompts and remix suggestions, then generates images through Codex exec workers in batches of up to 64 via a local mood-board server.
- Inline MCP Review and Remix: Renders the saved run through render_moodboard_board_widget with selection, expand, copy, annotation-based edits, and slot-based Remix controls.
- Use Case: A restaurant owner asks for visual directions for a new brand. The Skill runs intake, generates 18 single-scene reference images across three territories, and presents an inline board so the owner picks a direction before ads or menus are built.
Quick Start
Ask the agent to build an image-first mood board for your business brief, answer the intake questions, then review the generated board and pick a direction.