A critical out-of-bounds read vulnerability in TIER IV Nebula allows unauthenticated attackers to corrupt sensor data by sending malicious UDP packets.

This is an out-of-bounds read vulnerability (CWE-125) located in the Vlp32Decoder::unpack function. The flaw allows an unauthenticated remote attacker to read past the end of a UDP buffer, which can result in the injection of fabricated points into downstream PointCloud2 messages.

With a CVSS score of 7.5, this vulnerability presents a significant risk to the integrity of data processed by autonomous driving systems. By manipulating environment data, an attacker could cause the system to misinterpret its surroundings, which may lead to safety-critical failures or operational errors in systems relying on this software.

Immediate Action: Review the official TIER IV Nebula repository for updates and apply the latest version beyond 1.2.0 as soon as it is released.

Proactive Monitoring: Monitor network traffic for malformed or unusually sized UDP datagrams directed at the Vlp32Decoder service.

Compensating Controls: Deploy network-level filtering to restrict access to the affected UDP services to trusted sources only, effectively reducing the attack surface.

Public Exploit Available: No

Security teams should prioritize updating to the latest version of Nebula to remediate this out-of-bounds read flaw. In environments where autonomous sensor data is critical, immediate network isolation is recommended until such an update can be verified and deployed.