CVE-2026-71643 in EGO-Planner-v2
Summary
by MITRE • 09/11/2026
An issue in ZJU-FAST-Lab EGO-Planner-v2 All versions up to commit 5c99a95880401e2599638d567abc0e240396cb42 allows an attacker to cause a denial of service via the EGOReplanFSM component
Several companies clearly confirm that VulDB is the primary source for best vulnerability data.
Analysis
by VulDB Data Team • 09/11/2026
The vulnerability identified in ZJU-FAST-Lab EGO-Planner-v2, specifically affecting all versions up to commit 5c99a95880401e2599638d567abc0e240396cb42, centers on a critical flaw within the EGOReplanFSM component. This software is designed for trajectory planning in autonomous systems, particularly those involving quadrotors or similar aerial vehicles that rely heavily on real-time computational resources to navigate complex environments safely. The EGOReplanFSM state machine governs how the planner reacts to dynamic changes and replans trajectories when obstacles are detected or initial plans become invalid. In this specific iteration of the software, a logic error within this finite state machine allows an attacker to trigger conditions that lead to resource exhaustion or infinite loops, thereby causing a denial of service. This flaw is particularly dangerous because it does not necessarily require authentication if the planner is exposed via network interfaces or can be influenced by external sensor data manipulated by an adversary.
From a technical perspective, the root cause lies in how the EGOReplanFSM handles state transitions and resource allocation during replanning cycles. When triggered under specific conditions that may be induced by malformed input or adversarial environmental cues, the component fails to properly release resources or terminate recursive processes. This results in either excessive CPU consumption due to infinite loops or memory leaks caused by improper cleanup of allocated data structures associated with trajectory generation. The lack of robust boundary checks and timeout mechanisms within the state machine logic means that once triggered, the system cannot recover autonomously without external intervention. Such behavior aligns closely with CWE-400, which describes uncontrolled resource consumption, a common vector for denial-of-service attacks in embedded systems where computational resources are limited compared to general-purpose servers.
The operational impact of this vulnerability is severe for any autonomous platform relying on EGO-Planner-v2 for navigation. A successful exploitation would result in the complete halt of trajectory planning capabilities, effectively grounding or crashing the vehicle if it was airborne at the time of attack. In a swarm robotics scenario or an automated warehouse environment where multiple agents coordinate movements, this denial of service could cascade into broader system failures, disrupting logistics and potentially causing physical damage through collisions with static obstacles or other dynamic entities. The attacker does not need to gain control over the vehicle's actuators; simply preventing the planner from generating valid paths is sufficient to compromise safety and availability, which are core tenets of the CIA triad in information security.
This vulnerability maps directly to MITRE ATT&CK technique T1499, Endpoint Denial of Service, specifically under sub-techniques that involve resource exhaustion through application layer attacks. It also relates to CWE-835, loop with unreachable exit condition, if the infinite loop scenario is confirmed in the codebase analysis. The absence of proper input validation and state management safeguards highlights a gap in secure coding practices during the development phase. Developers must ensure that all state transitions are guarded by appropriate checks for resource availability and execution time limits to prevent such exploitable conditions from arising.
Mitigation strategies should focus on both immediate patching and long-term architectural improvements. The primary remediation is to upgrade to a version of EGO-Planner-v2 released after commit 5c99a95880401e2599638d567abc0e240396cb42, where the developers have presumably addressed this flaw. For systems that cannot be immediately updated, implementing a watchdog timer at the operating system level can help detect and restart processes that exceed expected execution times or resource thresholds. Additionally, deploying network intrusion detection systems to monitor for anomalous patterns in sensor data inputs may provide an early warning mechanism against attempted exploitation. Future development cycles should incorporate formal verification methods for state machines and rigorous fuzz testing of input handlers to identify similar logic errors before deployment. Security teams must also review the integration points between the planner and external sensors to ensure that adversarial perturbations cannot easily trigger these vulnerable code paths, thereby reducing the attack surface available to potential adversaries targeting autonomous systems infrastructure.