Security researchers have discovered "Underminr," a new vulnerability affecting approximately 88 million domains that allows attackers to hide malicious connections behind legitimate, trusted domains. This technique is a variant of domain fronting that exploits shared content delivery network infrastructure by presenting one domain's credentials while routing requests to a different, potentially malicious tenant on the same server. Experts warn that as AI-generated malware becomes more sophisticated, Underminr could soon appear in every attack attempting to evade protective DNS systems.