My understanding of Marr's Three Levels is that the Computational Problem often comes from the human researcher. As a result, the intentions or goals of the system being studied are in the mind of the researcher but seldom considered in the mind of the system. Although the goals can themselves be implicit or explicit. Secondly, our behavior often seems to be dependent on certain drives (eg: I'm sleepy vs hungry vs thirsty vs angry vs calm vs excited). If I am amidst a violent riot, I might use a packet of food or a shoe I see lying around as a weapon. I can also imagine fitting both these aspects into the Marr's Levels of Analysis and I love the framework for the wide-applicability and helping us systematize the phenomena under study, but I wonder if there is any recommended literature that incorporates these other aspects.
"Vehicles" (Valentino Braitenberg): https://mitpress.mit.edu/9780262521123/vehicles/