What problem does it solve?
Game teams need a stable full-screen visual anchor to guide scene direction and evaluate visual output, but ad-hoc image generation produces inconsistent, unvalidated results that cannot be audited or reused in a production pipeline.
Core Features & Use Cases
- Provider-driven image generation: Produces non-pixel-art full-screen references through configured providers (native, codex, gemini, openai) with role-preserving reference image attachments.
- Deterministic production pipeline: Records prompt, raw source, and provider/finalize reports at fixed paths, then resizes and center-crops the source to the exact target canvas with aspect validation.
- Use Case: A designer needs a 1920x1080 visual reference for a game's main battle screen. The Skill builds a prompt naming the game, camera viewpoint, HUD safe regions, and style, generates the image via the selected provider, finalizes it to the exact canvas, and validates the PNG output.
Quick Start
Ask the agent to generate a 1920x1080 screen reference for your game's main menu scene using the screen-reference asset type with your chosen image provider.