How would you break into your own application? What’s the worst thing that can happen? Is there anything you are secretly worried about? How would you try to abuse your system? Are there undocumented workarounds? If you were an evil version of yourself, what’s the worst you could do?
Pick a service or asset to model.
For example:
- An API recently built
- An older system that you worry about
- Something more personal, like your mailbox or bank account
Assess risks and potential impact
- What if the service becomes unresponsive?
- Is there sensitive data? What happens if that data is deleted or exfiltrated?
- What can a bad actor do if he gets in?
- Accept: decide that the business impact is acceptable, and document who has chosen to accept the risk
- Eliminate: remove components that make the vulnerability possible
- Mitigate: add checks or controls that reduce the risk impact, or the chances of its occurrence
- Transfer: Transfer risk to an insurer or customer.