Good Questions in Threat Modeling?

Techniques & Tooling
1

I am designing a Threat Modeling tool. It is good at (many to many) relationships!

At it’s heart, it has a set of rules with source type, question/prompt and target type. I can configure this really flexible.

A simple set and example is

HOME => What are we working on? => PROJECT
PROJECT => What can go wrong with <X>? => THREAT
THREAT => What helps with <X>? => MITIGATION

This helps create a tree of PROJECT, THREAT and MITIGATION nodes.

If I want likelihood/impact, I might add

THREAT => How likely is <X>? => LIKELIHOOD
THREAT => How bad is <X>? => IMPACT

If I am concerned about MITIGATION bypasses, I might add

MITIGATION => How could someone bypass <X>? => THREAT

(Notice how I introduced a circular thing?)

I’m fiddling around with different variants.

Call for ideas

What would you think would be a promising set of source type => question => target type rules, so that Threat Modeling will be best supported?

2

There’s some response in OWASP #threat-modeling Slack.

@AviD suggests asking “Tell me more.”

@izar advocates asking “How important is that to you”

…

3

See also OWASP TM Slack

Discussion spawned by @irene221b if we even need generator questions / is TM science and/or art.

4

More discussion about what static questions can do opposed to creative tailored questions…

OWASP thread evolves interestingly.