CVE-2025-62364 is a Local File Inclusion (LFI) vulnerability categorized under CWE-59 (Improper Link Resolution Before File Access) found in the open-source project oobabooga's text-generation-webui, a web interface for running Large Language Models. The vulnerability exists in versions up to 3.13 within the character picture upload functionality. Specifically, the application fails to properly validate or sanitize uploaded files that contain symbolic links. An attacker can craft a text file embedding a symbolic link pointing to any arbitrary file path on the server. Upon upload, the application follows the symbolic link and reads the target file's contents, subsequently serving this data through the web interface. This flaw allows an unauthenticated attacker to read sensitive files on the server, potentially exposing critical information such as system configuration files, credential stores, or other confidential data. The vulnerability does not require any privileges or user interaction, making it easier to exploit in environments where the vulnerable version is deployed. The issue was publicly disclosed on October 13, 2025, and fixed in version 3.14 of the software. No known exploits have been reported in the wild, but the medium CVSS score of 6.2 reflects the significant confidentiality impact and ease of exploitation via local access vectors. The lack of workarounds means patching is the primary mitigation strategy.

For European organizations using oobabooga text-generation-webui versions 3.13 or earlier, this vulnerability poses a significant risk of unauthorized disclosure of sensitive internal files. Exposure of configuration files or credentials could lead to further compromise, including privilege escalation or lateral movement within networks. Organizations deploying this software in research, AI development, or production environments may face confidentiality breaches impacting intellectual property, personal data, or operational security. Since the vulnerability can be exploited without authentication, any exposed instance accessible to attackers—even internally—could be leveraged. This risk is heightened in sectors with strict data protection regulations such as GDPR, where unauthorized data exposure can result in legal and financial penalties. The absence of known exploits in the wild currently limits immediate widespread impact, but the vulnerability's nature makes it a prime target for attackers once discovered. European entities relying on this software should prioritize patching to prevent potential data leaks and maintain compliance with cybersecurity standards.

The definitive mitigation is to upgrade oobabooga text-generation-webui to version 3.14 or later, where the vulnerability is fixed. Until upgrade is possible, organizations should restrict access to the upload feature by implementing strict network segmentation and access controls, limiting exposure to trusted users and systems only. Monitoring and logging upload activities can help detect attempts to exploit symbolic link uploads. Employing file integrity monitoring on server directories involved in uploads can alert administrators to suspicious symbolic link creations. Additionally, deploying web application firewalls (WAFs) with custom rules to detect and block symbolic link payloads in uploads may provide temporary protection. Organizations should also review and harden server file permissions to minimize the impact of any unauthorized file reads. Finally, conducting regular security assessments and penetration tests focusing on file upload functionalities will help identify residual risks.