Kaspersky experts discuss the Model Context Protocol used for AI integration. We describe the MCP's architecture, attack vectors and follow a proof of concept to see how it can be abused.| Securelist
In this article we analyze social engineering aspects of the XZ backdoor incident. Namely pressuring the XZ maintainer to pass on the project to Jia Cheong Tan, and then urging major downstream maintainers to commit the backdoored code to their projects.| securelist.com