The product uses an API function, data structure, or other entity in a way that relies on properties that are not always guaranteed to hold for that entity.
This can lead to resultant weaknesses when the required properties change, such as when the product is ported to a different platform or if an interaction error (CWE-435) occurs.
Fuzz testing (fuzzing) is a powerful technique for generating large numbers of diverse inputs - either randomly or algorithmically - and dynamically invoking the code with those inputs. Even with random inputs, it is often capable of generating unexpected results such as crashes, memory corruption, or resource consumption. Fuzzing effectively produces repeatable test cases that clearly indicate bugs, which helps developers to diagnose the issues.
Effectiveness: High