I should have said wherever rather than at all times when. I don't try to discuss this a lot so I guess I'm not as precise at organizing my thoughts as I'd like to be. Thanks. Yeah, "superbugs" are disproportionately at hospitals, and garden variety bacterial species; they cause problems because patients tend to have wounds or compromised immune systems, and because the frequency of antibiotic use at hospitals encourages thriving populations of resistant strains.
Especially if we can find ways sufficiently weighted towards bacteria. The major thing that saves us with them is that their significant low level differences from eukaryotic cells, e.g. those smaller and different ribosomes that provide a handy target for various antibiotic families (they're the organelles that synthesize proteins). Or their plant style cell walls.
(I stopped paying attention to the details of immune system in the late '70s when I noted how fast we were learning about it, i.e. for me it's something to learn/review "on demand", so I can't make any really specific suggestions as to what to do or judge the feasibility.)
Another promising research avenue is bacteriophages. I don't know why it's being pursued more in the former Soviet Union than it is in the West. I suppose even with targeted bacteriophages there's some risk that they could mutate and start targeting beneficial bacteria, which would be a problem.
A cute CS-inspired theoretical approach would be to throw variable layers at the problem. With genetic engineering, create bacteria that function as a supplemental immune system, but in a very narrow way; the bacteria would, possibly in collaboration with the host immune system, generate specific bacteriophages for any pathogenic bacteria they come into contact with. Obviously it wouldn't work against pathogenic viruses, but other than the (perhaps impossible) difficulty of having a cell able to generate novel bacteriophages, it doesn't seem so outlandish. If we get to the point where we can write code for a wide variety of biological activity and get it translated into DNA, then such a bacteriophage-generating bacterium might be possible.
I realize there are pathologies imaginable with any possible solution. The question is how likely is it that natural or engineered pathogens will target whatever new targets are available. Bacteriophage-generating bacteria could be hijacked by viruses, mutated by gene transfer with harmful bacteria to do bad things, or even give their bacteriophage-generating properties to pathological bacteria which might then gain mutations to generate human-targeting viruses... wouldn't that be fun.
Note that one possible flip side of this is auto-immune diseases.
"immune response is bad at all times when there's no active threatening infection to fight"
I don't think that's at all true. Our immune systems tend to work wonderfully most of the time given that we're constantly bathed in potentially pathogenic bacteria. In this discussion scythe says of "the ESKAPE pathogens, which are the primary superbugs" that "in every case except S aureus, infection is basically unheard of when an individual has a functioning immune system; they are all opportunistic pathogens". I take that to mean that for a lot people, temporary immune system dysfunction, e.g. due to other insults to the body, are a big factor.
I noticed on Wikipedia that one of the ESKAPE bacteria, Acinetobacter baumannii (http://en.wikipedia.org/wiki/Acinetobacter_baumannii), "Colloquially, A. baumannii is referred to as 'Iraqibacter' due to its seemingly sudden emergence in military treatment facilities during the Iraq War. It has continued to be an issue for veterans and soldiers who serve in Iraq and Afghanistan. Multidrug resistant (MDR) A. baumannii has spread to civilian hospitals in part due to the transport of infected soldiers through multiple medical facilities."
That I assume is largely due to gross trauma (and lots more soldiers surviving because of much better body armor (even helps in non-comat accidents I've read) and ever better treatment) ... and e.g. car crashes in the US tell us we're not soon going to be rid of that problem....