Embedded Threats: A Deep Dive Into The ESIM World

The talk began with a question: could an eSIM serve as an out-of-band command-and-control channel, separate from the network connections that security tools usually monitor? Markus Vervier walks through the eSIM stack, from the eUICC chip with its isolated profile domains to the remote provisioning backend, which is certified and protected by mutual TLS plus an additional end-to-end encrypted layer.

On the attack side, he looks at the Windows Mobile plans app, which renders web content from operator sites, so an attacker in a privileged network position could alter what a user sees. Using a test profile from a provider that allows it, he demonstrates an SMS-based command channel between two eSIM-equipped machines. It works but is slow, and the mobile carrier eventually blocked the test lines. He also installs a custom Java Card applet remotely over SMS and shows SIM-generated dialogs on an iPhone that the operating system does not control. Without the management keys, he examines SCP02 key diversification and a tool for offline key guessing.

The takeaway is that eSIM is a capable platform for attackers as well as a security control for operators. Applets running on the eUICC are largely invisible to phone-level forensics, so defenders should consider whether they control their eUICC and which provisioning paths are trusted. The speaker said the test applet and attack tools would be released.