Chosen-name Attacks: An Overlooked Class of Type-flaw Attacks

In the context of Dolev-Yao style analysis of security protocols, we consider the capability of an intruder to dynamically choose and assign names to agents. This capability has been overlooked in all significant protocol verification frameworks based on formal methods. We identify and classify new type-flaw attacks arising from this capability.Several examples of protocols that are vulnerable to this type of attack are given, including Lowe's modification of KSL. The consequences for automatic verification tools are discussed.