4 ms·
Uses movfuscator, one of the most painful tools when you're faced with reverse engineering. Also one of the most fun, from the other side of the aisle.
by shakna 9y ago
Uses movfuscator, one of the most painful tools when you're faced with reverse engineering. Also one of the most fun, from the other side of the aisle.
- chii 9y agoi love the design of the movfuscator. Most obfuscation tools tries to add more complexity and branching, in the hope that it will fool the decompiler. But it turns out, having absolutely no branching, and no 'features' makes it so much more of a nightmare to reverse!
- shakna 9y agoI think you can see a similar trend in estoric languages. Malboge is so complicated that nobody can really write a working program. Brainfuck is so simple that nearly anybody can write it, but reverse-engineering it is a massive time-sink.
- elago 9y agoHopefully this isn't too off topic, but could you explain how to even start approaching this from a RE/malware analysis perspective? I'm guessing there's no drag-and-drop de-obfuscate tool like there is for some of the common .NET obfuscators. Do you just rely on behavioral/dynamic methods?
- problems 9y agoFor even the nastiest of obfuscators there are often attempts at deobfuscators... https://github.com/kirschju/demovfuscator https://github.com/kirschju/demovfuscator
- shakna 9y agoI was just looking to replace a specific function call, because the company weren't using VC, and no longer had a working codebase. So I was hunting a particular pattern to replace, that I knew probably existed. A simple combination of demovfuscator, regex on the "original" asm and eyeballing it succeeded in the end.
- CapacitorSet 9y agoHas it ever been used against reverse engineering in the wild? Its lack of compiler features (it's based on lcc IIRC) and the sheer slowness of any operation makes it quite unsuitable for production usage.
- userbinator 9y agoThe comment you're referring to implies yes, and it's often applied to very specific pieces of code instead of the entire program. I have also seen a similar obfuscation, although perhaps not using The Movfuscator, many years ago in some shareware. You could probably guess which parts of the code used it; it even looks very distinctive when you open the binary in a text editor for a quick glance.
- shakna 9y agoIn my case, it's sort of a yes and no. I reverse engineered partly, but I was only looking to replace whole function calls. Company I came in as a contractor to... Sort of repair codebases and retrain staff, had a product they were getting paid buckets to support. Unfortunately, someone had broken the codebase quite severely, and they weren't using version control. All we had was a movfuscated release (yay for paranoid managers who go to tech conferences), but needed to fix a certain feature ASAP. So I used demovfuscator, someone's memory of what the code once looked like, and some ASM know-how to tear out references to the old and inject the new. Took a couple months of going nowhere fast, but I was pretty far out of my depth. They don't use movfuscator anymore.
- rootw0rm 9y agoAnd Tigress is even more painful...
- earenndil 9y agoApparently someone started writing a decompiler for movfuscator that looked for the patterns of movs generated. Probably possible to obfuscate past that still though with an extra layer of indirection whereby all instructions generate the same basic pattern of movs, with one small difference.