Threat actors are exploiting a vulnerability in shared content delivery network (CDN) infrastructure to hide connections to malicious domains.
Dubbed Underminr, the issue is a variant of domain fronting, a now-mitigated type of attack that enabled threat actors to place an allowed domain in the SNI and TLS certificate validation fields of an HTTPS request, while embedding a different target domain in the TLS tunnel’s encrypted HTTP host header.
Because CDNs routed requests internally based on the host headers, the request reached the hidden destination, while traffic would appear to be going to a reputable front domain.
Instead of using a front domain, Underminr presents the SNI and HTTP Host of a domain while forcing a request to the IP address of another tenant on the same shared edge.
The mismatch, ADAMnetworks reports, has been exploited in attacks targeting large-scale hosting providers, including those that have implemented mitigations against domain fronting.
“This abuse permits connections that appear to go to a trusted domain to actually connect to another domain that could be used for malicious intent,” the web security firm explains.
Threat actors can abuse Underminr to hide connections to command-and-control (C&C) servers, as well as VPN and proxy connections, and to circumvent network egress policies.
“In the simple form, the detection gap appears when DNS decisions, edge IPs, SNI, Host headers, and CDN tenant routing are not correlated. The endpoint sees an allowed DNS lookup while the connection can complete against a different hosted name,” ADAMnetworks says.
According to the company, the attack technique has been abused in attacks to connect to domains hosted on CDN infrastructure shared with allowed domains, mostly via TCP connections on port 443, in which SNI exposes the intended TLS hostname.
The Underminr vulnerability can be exploited using four different strategies to circumvent the DNS query monitoring and filtering service Protective DNS (PDNS).
In real-world scenarios, attackers can launch attacks using malicious applications and shell scripts. The vulnerability can also be abused in ClickFix attacks, ADAMnetworks says.
There are approximately 88 million domains potentially affected by Underminr, with internet infrastructure in the US, the UK, and Canada most impacted. Threat actors’ increased reliance on AI is expected to lead to a surge in attacks.
“Once Underminr becomes parametric information for AI-generated malware, we could expect to see it in every attack that needs to evade protective DNS as part of the attack chain,” ADAMnetworks CEO David Redekop says.
Related: $10 Domain Could Have Handed Hackers 25k Endpoints, Including in OT and Gov Networks
Related: US Disrupts Russian Espionage Operation Involving Hacked Routers and DNS Hijacking
Related: Complex Routing, Misconfigurations Exploited for Domain Spoofing in Phishing Attacks
Related:Microsoft Warns of ClickFix Attack Abusing DNS Lookups
