Use this skill whenever the user wants to generate, create, edit, or restyle an AI image through the agent's native image generation capability or configured image APIs such as GPT-image2, Nano Banana-compatible APIs, SiliconFlow Qwen Image, or OpenAI GPT Image-compatible APIs. Direct mode sends the user's prompt directly to the model. Pro mode only checks whether the user's prompt has the basic elements needed for generation, asks for missing information, then sends the completed user-approved prompt to the model.
2026-07-17