if I'm not mistaken, the agent should detect that a user has sent a message of type 'image' and parse the image automatically, if the agent is enabled