Create testbed to test rules, get rules prevalence, and find regressions

[mr-tz]

Let's discuss the creation of a testbed of samples (in addition/separate from the tests/data) to:

• test (new) rules
• obtain an approximation of rule hit prevalence
• find regressions

Some rough tasks:

• sketch out desired functionality and requirements
• collect 1000 - 10,000 samples (benign and malicious)
• create and automate frz file (frozen features) or BinExport creation to speed up processing
• create scripts and automation to facilitate the desired use-cases

CC: @Aayush-Goel-04 who helped to come up with this idea

[Aayush-Goel-04]

Seems like issue #424 is relevant to this discussion. I'll link it here for reference.