The anti-viral and anti-cancer STING innate immune pathway can exacerbate autoimmune and neurodegenerative diseases when aberrantly activated, emphasizing a key unmet need for STING pathway antagonists. However, no such inhibitors have advanced to the clinic because it remains unclear which mechanistic step(s) of human STING activation are crucial for potent and context-independent inhibition of downstream signaling. Here, we report that C91 palmitoylation, the mechanistic target of a potent tool compound, is not universally necessary for human STING signaling, making it a poor target for drug development. Instead, we discover that evolutionarily conserved C64 is basally palmitoylated and is crucial for preventing unproductive STING oligomerization in the absence of cGAMP stimulation. The effects of palmitoylation at C64 and C91 converge on the control of intra-dimer disulfide bond formation at C148. Importantly, we show for the first time that signaling-competent STING oligomers are composed of a mixture of two species: disulfide-linked STING dimers that stabilize the oligomer, and reduced STING dimers that are phosphorylated to actuate interferon signaling. Given this complex landscape and cell type specificity of palmitoylation modifications, we conclude that robust STING inhibitors must directly inhibit the oligomerization process. Taking inspiration from STING’s natural autoinhibitory mechanism, we identified an eight amino acid peptide that binds a defined pocket at the inter-dimer oligomerization interface as a proof-of-concept human STING inhibitor, setting the stage for future therapeutic development.