STRIDE
STRIDE is a threat modelling mnemonic used in secure design reviews. It names six failure modes to reason about for each component of a system: spoofing, tampering, repudiation, information disclosure, denial of service and elevation of privilege. Each one is the negation of a security property the design is supposed to hold.
The value of STRIDE is not the list itself but the discipline of walking it. You take a data flow diagram, and for every process, store, flow and trust boundary you ask the six questions in order. Spoofing tests authentication, tampering tests integrity, repudiation tests logging and non repudiation, information disclosure tests confidentiality, denial of service tests availability, and elevation of privilege tests authorisation. A component that survives all six has an argument behind it rather than an assumption.
Two limits are worth stating plainly. STRIDE finds design flaws, not implementation bugs: it will tell you that a trust boundary has no authentication, and it will say nothing about the injection flaw in the code behind it. And its output is only as good as the diagram, so a data flow diagram that omits the batch job nobody documented produces a clean model of a system that does not exist.
In practice we use it in threat modelling workshops to structure the conversation, then carry each credible elevation of privilege into a concrete attack path to test. In a report it appears as the category label on a design finding, not as a finding of its own.