I'm trying to fuzz an ASCII file format. Specifically, I'm defining some HTML5/HTML structures to be used as definition file for a smart fuzzer to fuzz web browsers. This smart fuzzer allows to exclude some of the included mutators. The number of iterations and the slowness of my fuzzers invites me to exclude some of them.
I would like to know your opinion about mutators that change string case or inject unicode strings in value fields. Knowing that HTML is an ASCII file format and that browsers are working well with this,
Do you think that I should exclude these mutators in order to focus the fuzzer in mutation that have higher probability of show vulnerabilities? Do you think it is important to mantain these mutators?
I obvously will mantain mutators that do the following things: Duplicate nodes Duplicate nodes recursively (test Mutators that mess with sizes (height, width) Mutators that mess with the order of the nodes Mutators that remove nodes
Do you recommend other type of mutators that will be interesting to include when fuzzing HTML/HTML5/XML file formats?