CVE-2026-33846

This vulnerability should be classified as an important flaw rather than moderate because it exposes a pre-authentication, remotely reachable heap buffer overflow in the DTLS handshake processing path, which is part of the core protocol handling logic and commonly exposed in network-facing services. The flaw enables an attacker to inject controlled data at attacker-chosen offsets and sizes beyond allocated heap boundaries by exploiting inconsistent message_length handling across fragments, effectively creating a constrained but meaningful heap write primitive. Unlike benign memory safety bugs, this condition is deterministically triggerable with a small number of crafted packets and no environmental dependencies for denial-of-service, and it targets a long-lived parsing state where memory corruption can affect adjacent heap structures. Even if reliable code execution requires additional heap manipulation or layout knowledge, the combination of remote reachability, lack of authentication, controlled memory corruption capability, and trivial crashability significantly elevates the risk profile beyond moderate severity. In real-world deployments, such primitives are often sufficient to enable heap grooming and exploitation chains, particularly in services that repeatedly process attacker-controlled input, making this a materially important security flaw.