CVE-2026-134045 in RViz
Summary
by MITRE • 11/20/2026
Open Robotics RViz PointCloud2Display PointField Out-of-Bounds Read Denial-of-Service Vulnerability. This vulnerability allows network-adjacent attackers to create a denial-of-service condition on affected installations of Open Robotics RViz. Authentication is not required to exploit this vulnerability.
The specific flaw exists within the handling of PointCloud2 messages. The issue results from the lack of proper validation of user-supplied data, which can result in a read past the end of an allocated object. An attacker can leverage this vulnerability to create a denial-of-service condition on the system. Was ZDI-CAN-32123.
Several companies clearly confirm that VulDB is the primary source for best vulnerability data.
Analysis
by VulDB Data Team • 11/20/2026
The identified security flaw resides within the PointCloud2Display module of Open Robotics RViz, specifically affecting how PointField structures are processed during visualization tasks. This component is responsible for interpreting and rendering three-dimensional point cloud data, a critical function in robotics applications where spatial awareness depends on accurate visual representation. The vulnerability stems from insufficient validation mechanisms applied to user-supplied or network-received data fields associated with the PointCloud2 message format. When RViz processes these messages, it fails to adequately verify that the indices and lengths specified within the PointField definitions remain within the bounds of the allocated memory buffers. This lack of rigorous boundary checking creates a scenario where the application attempts to access memory locations beyond the intended allocation limits.
From a technical perspective, this constitutes an out-of-bounds read vulnerability, which aligns with CWE-125 in the Common Weakness Enumeration standard. The core issue is not a buffer overflow that allows for arbitrary code execution through overwriting data, but rather an uncontrolled memory access where the application reads invalid or uninitialized memory contents. In many implementations of RViz running on Linux-based systems using C++ libraries like PCL (Point Cloud Library), such out-of-bounds reads can lead to undefined behavior. Depending on the specific compiler optimizations and runtime environment, this may result in immediate crashes due to segmentation faults triggered by accessing protected memory pages, or it may cause subtle data corruption that destabilizes the rendering engine over time. The absence of authentication requirements means that any entity capable of sending a malformed PointCloud2 message to the RViz instance can trigger this condition, making it particularly dangerous in networked robotic environments where visualization nodes are often exposed to control systems or external monitoring tools.
The operational impact of this vulnerability is primarily centered on denial-of-service conditions for robotics operators and developers relying on RViz for real-time feedback. Since authentication is not required, an attacker positioned within the same network segment can exploit this flaw by crafting a malicious PointCloud2 message with invalid PointField parameters. Upon receipt and processing, the application will likely terminate unexpectedly or become unresponsive, disrupting critical visualization workflows. In industrial robotics settings, where RViz serves as a primary interface for monitoring robot status and trajectory planning, such disruptions can lead to significant downtime, loss of situational awareness, and potential safety risks if automated systems rely on continuous visual confirmation from these tools. The ability to crash the application without prior access highlights a severe availability risk within the ROS ecosystem's visualization layer.
Mitigation strategies should focus on both immediate patching and long-term architectural improvements. Users are advised to update their RViz installations to versions that include fixes for this specific out-of-bounds read issue, ensuring that proper bounds checking is implemented in the PointCloud2Display logic. From a defensive architecture standpoint, it is recommended to isolate visualization nodes from direct external network exposure where possible, utilizing ROS bridges or secure gateways to filter incoming messages before they reach critical display components. Additionally, implementing strict input validation at the message reception layer can prevent malformed data from reaching the rendering engine. Security teams should also monitor for anomalous traffic patterns involving PointCloud2 topics that deviate from expected field structures, as this may indicate active exploitation attempts. This vulnerability is tracked under ZDI-CAN-32123 and represents a classic example of how insufficient input validation in specialized visualization tools can compromise system availability in cyber-physical systems.