One Model for ALL: Low-Level Task Interaction Is a Key to Task-Agnostic Image Fusion | Read Paper on Bytez