Ultralytics Parse Model Vulnerability Allows Remote Code Execution
A critical code injection vulnerability in Ultralytics' parse_model function enables remote attackers to execute arbitrary code on affected systems.
A critical code injection vulnerability, identified as ZDI-26-815, has been discovered in the popular Ultralytics software, potentially allowing remote attackers to execute arbitrary code on vulnerable systems. The vulnerability, which carries a high CVSS score of 7.8, requires user interaction for exploitation, such as tricking a user into visiting a malicious webpage or opening a specially crafted file.
The specific flaw lies within the parse_model function, where the software fails to adequately validate user-supplied input before using it in Python code execution. This oversight allows an attacker to inject malicious code that can then be executed with the privileges of the current user, leading to a full system compromise if the user has elevated permissions.
The Zero Day Initiative (ZDI) reported the vulnerability to the vendor on June 18, 2026, and followed up with inquiries on June 20 and October 19. After receiving no substantial updates, ZDI notified the vendor of their intention to publish an advisory on October 22, leading to the coordinated public release of the advisory on October 28, 2026.
Given the nature of this vulnerability, which relies on code injection through user-supplied data, the primary mitigation strategy recommended is to restrict interaction with the product. This means users should be cautious about opening files or accessing content generated by untrusted sources that might be processed by Ultralytics.
This discovery highlights a recurring theme in software security: the critical importance of robust input validation. Functions that process external data, especially when that data is used in code execution contexts, must implement stringent checks to prevent malicious payloads from being injected. The vulnerability was discovered by Bongeun Koo, who is credited for their work in identifying and reporting this flaw.
While the advisory does not specify a CVE ID, its disclosure as a "0-day" advisory by ZDI indicates that it was publicly disclosed before a patch was widely available or known. Users of Ultralytics are advised to monitor for any official security advisories or updates from the vendor to address this vulnerability.
The potential impact of this vulnerability is significant, as it allows for arbitrary code execution. This could lead to a wide range of malicious activities, including data theft, system disruption, or the installation of further malware. The requirement for user interaction, however, provides a layer of defense, making it less likely to be exploited in automated attacks.
As of the advisory's publication, the only effective mitigation is to limit exposure to untrusted inputs processed by the parse_model function. Users should exercise extreme caution and ensure they are only processing data from known and trusted sources until a permanent fix is released by Ultralytics.
This advisory from Zero Day Initiative details the specific vulnerability ZDI-26-817, which stems from improper validation of user-supplied strings before their use in executing Python code within the parse_model function. The vulnerability requires user interaction, such as visiting a malicious page or opening a malicious file, and carries a CVSS score of 7.8.