An example on how to do performant in-memory fuzzing with AFL++ and Frida. Currently works only on Linux, not Android.
To run use the same option of unicorn_mode, -U.
afl-fuzz -U -i in -o out -m none -- python3 frida_driver.py
If you want to adapt this script to other binaries that are not the simple test binary shipped within this repo you have to hack a bit fuzz.py to use your binary and frida_script.js to correctly call your target function.
You can also use showmap to test the instrumentation (and run frida_driver.py).
afl-showmap -U -o - -- python3 frida_driver.py
