I'm glad you know so much about this, but don't let your knowledge prevent you from doing the right thing. All I know is that Givewell seems extremely thorough, and holistic in their thoroughness. If there were a major effect such as the one you describe (warlords looting villages for bed nets), I expect that Givewell would recognize it and incorporate it into their estimates of cost-effectiveness for charities. Givewell is really very good. (If you want evidence for this, read some of their in-depth charity analyses and see how many factors they consider.)

And if the effect of warlords looting villages for their bed nets is not large enough to make a significant negative impact on the charity's effectiveness, then it still seems like one of the best possible uses of money to do good in the world.