Themida is a popular commercial software obfuscator which provides code virtualization and code mutation features. While Themida's code mutation is unanimously considered a weaker obfuscation scheme than code virtualization, there's little to no public information on the feature's implementation. As a result, it's difficult to estimate the code mutation's impact on an attacker's reverse engineering flow. In this talk we fill a bit of that gap by studying Themida's code mutation in details and looking for potential shortcomings. We'll use Binary Ninja and Python to understand how the code mutation works for x86-64 executables, ultimately automating its deobfuscation using Miasm and symbolic execution.