An astronaut probably wouldn't have been helpful in fixing Hubble's cable problem, though he might have determined the problem more quickly through simple visual inspection.
Even if an astronaut had been able to cut the cable and splice in a longer one, the engineering fix that was actually performed was far simpler and just as effective: merely avoid swinging the antenna through the problem area. It becomes an operational constraint just like avoiding gimbal lock on a 3-gimbal platform.
And then there are all the other issues involved in astronaut repair: the accessibility of the spacecraft orbit (the shuttle could only reach low altitudes), the exorbitant cost of shuttle missions (one of the main reason the program was ended), the difficulty of doing any kind of unanticipated repair work while wearing a pressure suit, the vastly increased safety requirements associated with human space flight, etc, etc.
Quite frankly, although the on-orbit servicing of the Hubble is frequently held up as an example of the versatility of the shuttle, I strongly suspect it would have been cheaper to just build a half dozen Hubbles and launch them on expendables as replacements as needed. HR Haddon's comment in Contact ("First rule in government spending: why build one when you can have two at twice the price?") is a bit too cynical; in actual fact, you can almost always negotiate a considerably lower cost per unit when you commit to several rather than just one. Not only is the NRE (non-recurrent engineering) cost amortized over more units, but there's almost always a manufacturing learning curve that brings down the variable costs of additional units.
