In a concerning development for the rapidly evolving landscape of AI-assisted development, cybersecurity researchers have identified a novel attack vector that specifically targets AI coding agents. This emerging threat, dubbed "Agentjacking" by the team at Tenet Security, represents the first known class of attack designed to manipulate these increasingly popular development tools into executing malicious code on unsuspecting developers' machines.
The Agentjacking attack exploits the way AI coding agents interact with error-reporting systems. By crafting a sophisticated fake error report using Sentry, an open-source error-tracking platform widely adopted by development teams, attackers can