Summary: Cancer cells overexpress a diversity of anti-apoptotic BCL-2 family proteins, such as BCL-2, MCL-1, and BFL-1/A1, to enforce cellular immortality. Thus, intensive drug development efforts have focused on targeting this class of oncogenic proteins to overcome treatment resistance. Whereas a selective BCL-2 inhibitor has been FDA approved and several small molecule inhibitors of MCL-1 have recently entered phase I clinical testing, BFL-1/A1 remains undrugged. Here, we developed a series of stapled peptide design principles to engineer a functionally selective and cell-permeable BFL-1/A1 inhibitor that is specifically cytotoxic to BFL-1/A1-dependent human cancer cells. Because cancers harbor a diversity of resistance mechanisms and typically require multi-agent treatment, we further investigated BFL-1/A1 co-dependencies by mining a genome-scale CRISPR-Cas9 screen. We identified ataxia-telangiectasia-mutated (ATM) kinase as a BFL-1/A1 co-dependency in acute myeloid leukemia (AML), which informed the validation of BFL-1/A1 and ATM inhibitor co-treatment as a synergistic approach to subverting apoptotic resistance in cancer. : Guerra et al. constructed an exquisitely selective BFL-1 inhibitor capable of covalent BFL-1 targeting and cellular penetrance without membrane disruption. Mining a genetic dependency database revealed a spectrum of BFL-1 dependency in cancer and an ATM co-dependency in AML, prompting the combination of BFL-1 and ATM inhibitors to achieve synergistic cytotoxicity. Keywords: BFL-1, A1, BCL-2 family, apoptosis, stapled peptide, covalent inhibitor, dependency, ATM, AML, cancer