SoK: On the Fragility of Memory Error Exploit Mitigations

The perennial war in memory has long been shaped by a continuous arms race: defenses are deployed, bypasses emerge, and stronger mitigations follow, only for the cycle to repeat. This pattern persists in part due to the fragility of many defenses, which often fail when assumptions change. Such fragility reflects how security guarantees are assessed, often through ad hoc reasoning tied to specific threat models. In a fast-evolving landscape of polyglot applications and heterogeneous systems, manually re-evaluating these guarantees for every new context is both labor-intensive and error-prone.

This SoK advocates for a more systematic, adversary-aware approach. We introduce a graph-based framework for evaluating the fragility of memory safety defenses by modeling the progression of memory corruption exploits, and where, how, and under what assumptions defenses intervene and may fail. We demonstrate the utility of this model by revisiting flaws in prominent defenses and showing how their shortcomings could have been anticipated. Finally, we release open-source tooling that implements our model and supports systematized and semi-automated fragility testing.