Add ModuleExecution and related data models

[cdelafuente-r7]

This adds ModuleExecution, ModuleExecutionError and ModuleExecutionEvent data models as part of the reporting refactor in Metasploit Framework.

(I'll update this description once the related PR in Framework is ready)

[smcintyre-r7]

I don't think "external" is a module type in the same sense as the others. You can have for example an external exploit or auxiliary module that's implemented in Python. I would suggest removing it or tracking it as a flag so you can note something is an external-exploit, external-auxiliary, etc.

[smcintyre-r7]

If unhandled_exception is the inverse of expected_failure, it might be more intuitive to track it as unexpected_failure.

[cdelafuente-r7]

Let me give more details about these status:

• expected_failure: the module explicitly declared failure via fail_with. It is an error the module author has anticipated with a proper failure_reason.
• unhandled_exception: a Ruby exception that escaped the module's rescue chain (NoMethodError, Rex::ConnectionError outside a rescue, etc.). The author did not anticipate this; there's an exception_class and backtrace.

So, unhandled_exception and expected_failure aren't strict inverses. For example, unhandled_exception will help answer question like "which modules are buggy?"

[smcintyre-r7]

Suggested change

	ORIGINATING_UIS = %w[console rpc json_rpc mcp external import plugin autocheck].freeze
	ORIGINATORS = %w[console rpc json_rpc mcp external import plugin autocheck module].freeze

I'd suggest adding 'module' as an originator to account for modules running other modules. Additionally since some of these aren't exactly user-interfaces, it may be better to use a slightly more generic name like ORIGINATORS vs ORIGINATING_UIS.

[cdelafuente-r7]

You're right it's not accurate. I would prefer using ORIGINATING_INTERFACES, which is more programmatic. ORIGINATORS sounds like a user, which seems slighly wrong with things like autocheck. I'm also fine if you disagree, I'll change it to ORIGINATORS, no problem.

[smcintyre-r7]

Events should be immutable, so let's remove the updated_at column so once it's created, it stays the same.

Suggested change

	# @!attribute [rw] updated_at
	# Last time this event row was updated.
	#
	# @return [DateTime]

[smcintyre-r7]

This should also be immutable.

Suggested change

	# @!attribute [rw] updated_at
	# Last time this error row was updated.
	#
	# @return [DateTime]

[cdelafuente-r7]

Yes, this makes no sense for Error and Events. I removed these fields. Thanks!