AEMB: An Automated Exploit Mitigation Bypassing Solution
Ruipeng Wang, Zulie Pan, Fan Shi, Min Zhang
Abstract
Open-access reader
Ruipeng Wang, Zulie Pan, Fan Shi, Min Zhang
Abstract
Open-access reader
Modern operating systems set exploit mitigations to thwart the exploit, which has also become a barrier to automated exploit generation (AEG). Many current AEG solutions do not fully account for exploit mitigations, and as a result, they are unable to accurately assess the exploitability of vulnerabilities in such settings.This paper proposes AEMB, an automated solution for bypassing exploit mitigations and generating useable exploits (EXPs). Initially, AEMB identifies exploit mitigations in the system based on characteristics of the program execution environment. Then, AEMB implements exploit mitigations bypassing the payload generation by modeling expert experience and constructs the corresponding constraints. Next, during the program’s execution, AEMB uses symbol execution to collect symbol information and create exploit constraints. Finally, AEMB utilizes a solver to solve the constraints, including payload constraints and exploit constraints, to generate the EXP. In this paper, we evaluated a prototype of AEMB on six test programs and seven real-world applications. Furthermore, we conducted 54 sets of experiments on six different combinations of exploit mitigations. Experiment results indicate that AEMB can automatically overcome exploit mitigations and produce successful exploits for 11 out of 13 applications.
OpenAlex reports 5 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Modern operating systems set exploit mitigations to thwart the exploit, which has also become a barrier to automated exploit generation (AEG). Many current AEG solutions do not fully account for exploit mitigations, and as a result, they are unable to accurately assess the exploitability of vulnerabilities in such settings.This paper proposes AEMB, an automated solution for bypassing exploit mitigations and generating useable exploits (EXPs). Initially, AEMB identifies exploit mitigations in the system based on characteristics of the program execution environment. Then, AEMB implements exploit mitigations bypassing the payload generation by modeling expert experience and constructs the corresponding constraints. Next, during the program’s execution, AEMB uses symbol execution to collect symbol information and create exploit constraints. Finally, AEMB utilizes a solver to solve the constraints, including payload constraints and exploit constraints, to generate the EXP. In this paper, we evaluated a prototype of AEMB on six test programs and seven real-world applications. Furthermore, we conducted 54 sets of experiments on six different combinations of exploit mitigations. Experiment results indicate that AEMB can automatically overcome exploit mitigations and produce successful exploits for 11 out of 13 applications.
Key concepts: Exploit, Computer science, Computer security