Detection of JavaScript injection eavesdropping on WebRTC communications

WebRTC is a Google-developed project that allows users to communicate directly. It is an open-source tool supported by all major browsers. Since it does not require additional installation steps and provides ultra-low latency streaming, smart city and social network applications such as WhatsApp, Facebook Messenger, and Snapchat use it as the underlying technology on the client-side both on desktop browsers and mobile apps. While the open-source tool is deemed to be secure and despite years of research and security testing, there are still vulnerabilities in the real-time communication application programming interface (API). We show in this paper how eavesdropping can be enabled by exploiting weaknesses and loopholes found in official WebRTC specifications. We demonstrate through real-world implementation how an eavesdropper can intercept WebRTC video calls by installing a malicious code onto the WebRTC webserver. Furthermore, we identify and discuss several, easy to perform, ways to detect wiretapping. Our evaluation shows that several indicators within webrtc-internals API traces can be used to detect anomalous activities, without the need for network monitoring tools.