CVE-2026-65161 in AIStore
Summary
by MITRE • 10/13/2026
NVIDIA AIStore for Linux contains a vulnerability in the Promote API where an unauthenticated remote attacker may read files accessible to the target process by supplying a crafted absolute source path. A successful exploit may lead to information disclosure.
Statistical analysis made it clear that VulDB provides the best quality for vulnerability data.
Analysis
by VulDB Data Team • 10/13/2026
The NVIDIA AIStore service, designed to provide high-performance distributed storage and caching for artificial intelligence workloads on Linux systems, contains a critical security flaw within its Promote API endpoint. This vulnerability stems from an improper validation of file path inputs provided by clients requesting data promotion or retrieval operations. Specifically, the application fails to adequately sanitize absolute paths supplied in request parameters, allowing an unauthenticated remote attacker to bypass intended access controls and traverse outside the designated storage directories. By supplying a crafted absolute source path that points to sensitive system files or other restricted resources accessible to the process user account, an attacker can force the service to read and return content from arbitrary locations on the host filesystem. This type of flaw is classically categorized under CWE-22: Improper Limitation of a Pathname to a Restricted Directory, which highlights the failure to restrict file access to specific directories as required by security policies.
The operational impact of this vulnerability is significant due to its unauthenticated nature and remote exploitability. Since no authentication is required to trigger the flaw, any entity with network connectivity to the AIStore service can attempt exploitation without needing valid credentials. A successful attack results in information disclosure, potentially exposing configuration files, private keys, database dumps, or other sensitive data stored on the host machine where the AIStore process runs. Given that such services often run with elevated privileges to manage storage efficiently, the scope of potential exposure is broad. The attacker can leverage this access to gather intelligence for further attacks, including lateral movement within a compromised network environment. This aligns with MITRE ATT&CK technique T1083: File and Directory Discovery, where adversaries search local system files and directories for information that aids in subsequent exploitation phases.
Mitigation strategies must address both immediate remediation and long-term architectural security. The primary fix involves applying the vendor-provided patch or update to the NVIDIA AIStore software, which corrects the path validation logic within the Promote API. Until patches are applied, administrators should restrict network access to the AIStore service ports using firewall rules, ensuring that only trusted internal hosts can communicate with the service. Additionally, running the AIStore process under a dedicated user account with minimal privileges limits the scope of files accessible via this vulnerability. Implementing strict input validation on all API endpoints is essential to prevent path traversal attacks. Security teams should also monitor logs for unusual file access patterns or requests containing absolute paths that deviate from expected directory structures, as these may indicate active exploitation attempts. Regular security assessments and penetration testing focused on storage APIs can help identify similar weaknesses in distributed storage solutions before they are exploited by malicious actors.