Scheduler for LibAFL
- Mentors
- Andrea Fioraldi
- Organization
- AFLplusplus
Fuzzing has been a popular vulnerability detection tool, and researchers have developed various techniques to build the "best" fuzzer. It is claimed, however, that the fuzzer adjusted for the target program that one wants to fuzz is the most effective one.
LibAFL offers those functionalities, with which users can customize the fuzzer or even extend the fuzzer tailored for each target program. However, since it is still at the early stage of development, some of the powerful scheduling algorithms have not been incorporated into LibAFL yet.
This proposal is to implement two effective scheduling algorithms, the AFLFast, and the MOpt algorithm into LibAFL. With these techniques, LibAFL can offer users a faster way to discover vulnerabilities and a more efficient mutator selection strategy.