Unrestricted File Upload Vulnerability in Anil-matcha Open-Generative-AI

A security vulnerability (CVE-2026-90603) has been identified in Anil-matcha Open-Generative-AI versions up to 1.0.11 and 2.0.0. The vulnerability resides in the S3 Upload component within the /api/upload-binary file. An unauthenticated remote attacker can exploit this flaw by manipulating the x-proxy-target-url parameter, which bypasses intended file validation and allows for the upload of arbitrary files to the server. This vulnerability poses a significant risk as it can lead to remote code execution (RCE) if an attacker is able to upload malicious scripts or executables to a web-accessible directory. A patch (f013270957f75e439eaf97eb2a93decb32a4543e) has been released to address this issue.

Impact

Successful exploitation of this vulnerability allows remote attackers to upload arbitrary files to the server infrastructure, potentially resulting in full system compromise, data theft, or persistent malware installation. The vulnerability affects all deployments of Open-Generative-AI running versions 1.0.11 or earlier, or the 2.0.0 release.

Recommendation

Prioritize the following actions to secure the environment:

• Apply the vendor-provided patch f013270957f75e439eaf97eb2a93decb32a4543e to all instances of Open-Generative-AI immediately.
• Audit server logs for requests to /api/upload-binary containing unexpected patterns in the x-proxy-target-url parameter.
• Restrict access to the /api/upload-binary endpoint using network-level controls if immediate patching is not possible.