Researchers at UC Santa Barbara discovered that LLM API routers used for cost optimization can act as malicious intermediaries, with 9 routers actively injecting malware into AI responses and 17 stealing credentials. The risk is amplified when autonomous agents execute code without human approval, potentially allowing attackers to compromise systems through compromised routers or prompt injections.
Researchers at UC Santa Barbara discovered that LLM API routers used to reduce costs and balance loads are vulnerable to man-in-the-middle attacks. They found that 9 routers inject malicious code into AI responses and 17 steal credentials, with autonomous agent execution enabling immediate code exploitation without human approval.