I recall reading about this idea (but not using CRISPR, natch) in New Scientist back in the 1990's.
They covered off the precautionary principle / unintended consequences concern by proposing to retain a breeding stock of the target species in a closed system (say a greenhouse).
You'd be looking at a few years to deplete or render extinct the wild stock, during which time you could be sure the malaria pathogen was absent from your breeding stock.
Once malaria, and the modified mosquitoes, are confirmed absent from the wild, you have the option to re-introduce. Naturally the option also exists at any earlier time if the species turns out to be more important than originally determined.
They covered off the precautionary principle / unintended consequences concern by proposing to retain a breeding stock of the target species in a closed system (say a greenhouse).
You'd be looking at a few years to deplete or render extinct the wild stock, during which time you could be sure the malaria pathogen was absent from your breeding stock.
Once malaria, and the modified mosquitoes, are confirmed absent from the wild, you have the option to re-introduce. Naturally the option also exists at any earlier time if the species turns out to be more important than originally determined.