Agentjacking Attack Tricks AI Coding Agents Into Running Malicious Code

Agentjacking is a vulnerability/exploit where attackers send specially crafted Sentry error events that AI coding assistants misinterpret as trusted system guidance, causing the agents to execute attacker-controlled code on developers' machines and potentially exfiltrate credentials and sensitive environment data. Tenet Security found thousands of exposed DSNs and achieved high exploitation success rates in controlled tests; Sentry has applied limited filtering but declined a full fix.