RFC: DotNet Assembly Runner

[sempervictus]

Similar to #576, this request for comment seeks to assess the usability and cost/benefit ratio is implementing a CLR assembly execution harness which would permit us to pull in various post-exp frameworks' executable "modules" written in .NET.

Unlike #576, there are a number of ways to skin this cat including leveraging our own PSH extension which is a CLR runtime of its own. This example uses the native CLR runtime, and i think ideally we would permit users to decide whether to use the runtime from our extension or what's on-host.

[zeroSteiner]

Isn't this functionality provided by https://github.com/rapid7/metasploit-framework/blob/master/modules/post/windows/manage/execute_dotnet_assembly.rb ? rapid7/metasploit-framework#12405.

[zeroSteiner]

Now that #601 is landed, you can also run .NET assemblies using the BOF loader and anthemtotheego/InlineExecute-Assembly.

[sempervictus]

I think the BOF and module cover the use case of leveraging the CLR runtime native to the host. Thanks @zeroSteiner for linking those.
The second part of the consideration is around "what do we do if we want to avoid hooked/tripped tartget-local CLR?" The Python meterpreter extension carries an RDI-compatible runtime under the attacker's control into the target system; and since .NET core is Open Source, it seems to me that a similar approach should be possible to permit execution of .NET code without ever touching the host's DLLs, tripping hooks on loading of the library, etc.