Self-adapting Robotic Agents through Online Continual Reinforcement Learning with World Model Feedback | ScienceToStartup | ScienceToStartup