from praisonaiagents.tools import multimodal_content, image_part, text_part

def screenshot() -> ToolResult:
    png_bytes = capture()
    return multimodal_content(
        text_part("Here is the current screen:"),
        image_part(png_bytes, mime="image/png", name="screen.png"),
    )

def multimodal_content() -> 'ToolResult'