Mine life for molecules to treat bacteria
Let bacteria try to infect various plants and animals, see which adapt, and why (what proteins or molecules do they evolve), and apply it to developing new treatments.
While the universal decline in the effectiveness of antibiotics has generated new interest in phage therapy, the repertoire of means to deal with superbugs mustn't end there: Earth has evolved a diversity of life forms (besides viruses), that may have encountered the same bacterial tricks before, and found ways to deal with them, so, searching for natural immunity to certain bacteria doesn't have to stop with just evolving viruses to infect them.
Bacteria, being a rather ubiquitous creature, might have historically interacted with a variety of plants, some of which may have evolved immunity to them as well. Imagine that in the future, simply drinking juice of a currently unknown plant could be equivalent to drinking antibiotics, because that plant happens to produce a molecule that's harmless to us, but puts an end to a superbug, or, imagine that a bat or a snake, that gets infected with a bacteria, recovers from the infection by producing an antibody: we could thus harvest the protein that does the trick in a similar way that we train horses to produce anti-bodies to neutralize snake venom.
The point is, looking for animals and plants that are naturally resistant (or quickly develop immunity) to human known superbugs is a general strategy that this idea is about, and while this strategy to an extent has been tried, it's probably worth thinking about afresh.