Antimicrobial peptides (AMPs) are host-encoded antibiotics that combat invading microbes. These short immune effectors are conserved in plants, animals, and fungi. Early work showed that AMPs killed bacteria in generalist fashions in vitro: i.e. AMPs that killed Escherichia coli also killed many Gram-negative bacteria when tested. At the genome level, AMP gene families rapidly expand or contract, which suggested single genes were unlikely to be important. At the signalling level, AMPs are induced as a suite of peptides by conserved NF-κB immune pathways across organisms (e.g. Toll, Imd). Together these observations led to the assump-tion that individual genes contributed only small effects, and instead the cumulative cocktail of AMPs was key to a successful de-fence response. This idea was never robustly tested in vivo owing to technical limitations. In 2015, two studies stumbled onto re-markable effects of fruit fly immune effectors. In one case a polymorphism in a single AMP gene (Diptericin A or DptA) greatly af-fected the fly defence against a specific bacterium (Providencia rettgeri). In another case, deleting just the Bomanin gene family caused immune susceptibility mimicking loss of Toll signalling generally. The prevailing model of generalist AMP action was ill equipped to explain these findings. In my PhD, I have systematically deleted the AMP genes of fruit flies to clarify AMP defences in vivo. This confirmed AMPs can act in generalist or redundant fashions in some cases. However some AMP-pathogen interactions are remarkably specific. Deletion of just the Drosocin gene explains much of the susceptibility of NF-κB/Imd immune deficient flies to Enterobacter cloacae infection. Meanwhile deletion of just the two Diptericins recapitulates the susceptibility of Imd mutants to P. rettgeri. Contrary to previous assumptions, our findings suggested AMPs are not simple generalist peptides. There are many more short peptide immune genes waiting to be characterized that may be relevant to specific infections. I next investigated three lesser-characterized genes: Baramicin A, DptB, and Drosocin. The Baramicin A gene encodes multiple products, including one peptide (IM22) that was first annotated in my study. I found BaraA is key to the fly defence against pathogenic fungi. Next, I dissected the roles of individual Diptericin genes. Surprisingly, DptB alone is required for survival after infection by a lab isolate of Acetobacter. Finally, I identified Drosocin as the source gene for IM7, a mystery peptide first detected in 1998. A polymorphism in IM7 previously obscured its identification. This polymorphism affects fly defence against Providencia burhodogranariea, where one immune-poor allele effectively rivals deletion of IM7 entirely against this microbe. These AMP-microbe interactions reveal that survival after infection can be mediated at the level of single AMP genes or even com-mon alleles of those genes. Contrary to previous assumptions, it appears the AMP response is composed of silver bullets, wooden stakes, and other specialized defence tools required to fight specific enemies. In this light, the diverse AMPs induced upon infection may be an evolutionary solution to optimize survival: general microbe patterns induce immune signalling, there is a need for timely production of antimicrobial peptides, and so a battery of antimicrobial peptides are produced immediately even though only a few are likely to be relevant for any given pa