Use-value is definitely context dependent and agent dependent. This is fine. But we have to acknowledge that the reason why people assign exchange-value to goods is because they have actual use-value: economics describes physical, real world activities.
I can agree that no one has presented a good dynamical theory of use-value. That's sensible: it would require at least a full description of reality, and a deep understanding of nonlinear mathematics. One quibble i have is i don't think Marx tried to establish such a theory: his observations about use-value and exchange-value can be made without a full theory of use-value or a full theory of exchange-value. The relationship between the two can be examined in the abstract and that is Marx's program.
I can still perform calculations with use-value to derive exchange value: why is water worth so little in exchange-value even though it is so vital to life? The marginalist point of view says that the price of water is determined by equilibrium between marginal cost and utility. Why is the marginal utility of water so low?
Well, if we're talking about people drinking water, then it is obvious, isn't it? You only have to drink so much water. When you're not thirsty, water has none of that use-value anymore.
Ignoring the relationship between use-value and exchange-value is... weird. It's replacing the true and difficult reality with a nice fiction amenable to manipulation with our primitive mathematics. Do markets define exchange-value or do they compute exchange-value from the aggregate use-values of everyone participating?
Edit: as for business cycles, i think they are caused by structural flaws in property rights and financial institutions, but if you asked 5 people why the business cycle occurs, you'll get 6 conjectures and 3.5 theories.