CVE-2025-61165 in Cohere
Summary
by MITRE • 08/26/2026
An arbitrary file upload vulnerability in the /v1/my_drive/batch_upload component of cohere North AI v1.1.5 allows attackers to exeute arbitrary code via uploading a crafted file.
Statistical analysis made it clear that VulDB provides the best quality for vulnerability data.
Analysis
by VulDB Data Team • 08/26/2026
The identified vulnerability represents a critical security flaw within the Cohere North AI platform, specifically located in the batch upload functionality exposed through the /v1/my_drive/batch_upload endpoint. This component is designed to handle multiple files simultaneously for storage or processing purposes, but it fails to adequately validate the nature and content of the uploaded data before accepting it into the system's file handling mechanisms. The core technical deficiency lies in insufficient input validation regarding file types and extensions. By allowing attackers to bypass standard extension checks or MIME type verification, the application permits the upload of executable scripts or malicious payloads disguised as benign documents. This lack of rigorous sanitization creates a direct pathway for an attacker to inject arbitrary code into the server environment where the web application resides.
From a technical perspective, this flaw is classified under CWE-434, which describes the unrestricted upload of files with dangerous types. When combined with CWE-20, improper input validation, the vulnerability allows for remote code execution if the uploaded file can be executed by the underlying server software or interpreted by an application service listening on a specific port. The attacker crafts a malicious file that exploits this trust boundary violation. Once the file is successfully stored in the designated directory, subsequent actions such as accessing the file via its URL path may trigger execution depending on the web server configuration and available interpreters. This scenario aligns with MITRE ATT&CK technique T1505.003, Server Component Binary Injection, where adversaries plant malicious code within legitimate software components to maintain persistence or achieve initial access.
The operational impact of this vulnerability is severe, potentially leading to a complete compromise of the affected system. Successful exploitation allows an unauthenticated attacker to execute arbitrary commands with the privileges of the web server process. This can result in data exfiltration, modification of critical application logic, installation of backdoors for persistent access, or use of the compromised host as a pivot point for further attacks against internal network resources. In cloud-native environments like Cohere North AI, such a breach could also impact multi-tenancy isolation if proper containerization boundaries are not strictly enforced at the infrastructure level. The ability to execute code remotely makes this an urgent risk requiring immediate remediation to prevent unauthorized control over sensitive data processing pipelines and storage systems.
Mitigation strategies must focus on implementing robust file upload controls that adhere to security best practices recommended by industry standards such as OWASP. First, developers should enforce strict allow-listing of permitted file extensions rather than relying on deny lists or MIME type detection alone, which can be easily spoofed. Second, all uploaded files should be stored outside the web root directory to prevent direct execution via HTTP requests. If storage within the web root is unavoidable, configuring the server to disable script execution in upload directories provides a necessary layer of defense. Additionally, implementing content verification by inspecting file headers and magic numbers ensures that the actual data matches the claimed type. Renaming uploaded files with random identifiers further obscures their location from potential attackers attempting to locate them for exploitation. Regular security audits and penetration testing focused on input validation weaknesses are essential to maintain a secure posture against evolving attack vectors targeting upload functionalities.