Summary
NVIDIA AIStore for Linux contains a path traversal vulnerability in the dSort archive extraction component. A remote attacker who can submit or influence a dSort job may use crafted archive member names to create or overwrite files outside the intended directory, potentially enabling code execution, information disclosure, data tampering, or denial of service.
NVIDIA AIStore for Linux contains a path traversal vulnerability in the dSort archive extraction component. A remote attacker who can submit or influence a dSort job may use crafted archive member names to create or overwrite files outside the intended directory, potentially enabling code execution, information disclosure, data tampering, or denial of service.
Affected context
The issue occurs in AIStore deployments that use the dSort archive extraction workflow and process attacker-influenced archive contents. Risk depends on whether remote users can submit dSort jobs, which filesystem locations are writable by the service, and what permissions the extraction process has.
The issue occurs in AIStore deployments that use the dSort archive extraction workflow and process attacker-influenced archive contents. Risk depends on whether remote users can submit dSort jobs, which filesystem locations are writable by the service, and what permissions the extraction process has.
Exploit preconditions
- Attack surface: Remote dSort job submission or another remote path that permits archive processing
- Required access: Sufficient access to create, submit, or influence a dSort job and its archive input
- Trigger point: Archive extraction when member names are resolved into filesystem paths
- Why exploitation is possible: Archive paths are not adequately constrained to the intended extraction directory, allowing traversal into other filesystem locations
- Attack surface: Remote dSort job submission or another remote path that permits archive processing
- Required access: Sufficient access to create, submit, or influence a dSort job and its archive input
- Trigger point: Archive extraction when member names are resolved into filesystem paths
- Why exploitation is possible: Archive paths are not adequately constrained to the intended extraction directory, allowing traversal into other filesystem locations
Why this happens
The vulnerability results from improper pathname restriction during archive extraction. When archive member names are combined with a destination directory without canonicalization and an effective containment check, traversal sequences can cause the resolved path to escape the intended directory. This behavior maps to CWE-22: Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal').
The vulnerability results from improper pathname restriction during archive extraction. When archive member names are combined with a destination directory without canonicalization and an effective containment check, traversal sequences can cause the resolved path to escape the intended directory. This behavior maps to CWE-22: Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal').
Potential impact
- An attacker may create or overwrite files outside the designated extraction directory
- The service may disclose files that should remain inaccessible
- Modified files could enable code execution under the service account
- Application data or extracted content may be tampered with
- The affected service may experience denial of service
- Broader compromise could occur if the process has excessive filesystem privileges
- An attacker may create or overwrite files outside the designated extraction directory
- The service may disclose files that should remain inaccessible
- Modified files could enable code execution under the service account
- Application data or extracted content may be tampered with
- The affected service may experience denial of service
- Broader compromise could occur if the process has excessive filesystem privileges
General mitigation guidance
Mitigation generally focuses on enforcing a strict extraction boundary and reducing the trust placed in archive metadata.
Common defensive patterns include:
- Canonicalizing each destination path before file creation
- Rejecting absolute paths, traversal components, and paths that resolve outside the extraction root
- Using extraction APIs or wrappers that enforce directory containment by default
- Applying strict archive validation and resource limits before extraction
- Running extraction workers with least-privilege filesystem and operating-system permissions
- Using isolated, temporary workspaces with controlled cleanup and read-only access to unrelated data
- Limiting which users or services may submit dSort jobs
Mitigation generally focuses on enforcing a strict extraction boundary and reducing the trust placed in archive metadata.
Common defensive patterns include:
- Canonicalizing each destination path before file creation
- Rejecting absolute paths, traversal components, and paths that resolve outside the extraction root
- Using extraction APIs or wrappers that enforce directory containment by default
- Applying strict archive validation and resource limits before extraction
- Running extraction workers with least-privilege filesystem and operating-system permissions
- Using isolated, temporary workspaces with controlled cleanup and read-only access to unrelated data
- Limiting which users or services may submit dSort jobs
How teams detect issues like this
Organizations can detect path traversal risks through:
- Static analysis rules for archive extraction, path joining, and file creation APIs
- Code review checks for canonicalization and extraction-root containment validation
- Tests using absolute paths, traversal components, platform-specific separators, and encoded path variants
- Runtime monitoring for unexpected writes outside designated job directories
- File-integrity monitoring on sensitive paths and policy enforcement restricting service-account write locations
- Automated security review tools that flag archive extraction without explicit path-boundary checks
Organizations can detect path traversal risks through:
- Static analysis rules for archive extraction, path joining, and file creation APIs
- Code review checks for canonicalization and extraction-root containment validation
- Tests using absolute paths, traversal components, platform-specific separators, and encoded path variants
- Runtime monitoring for unexpected writes outside designated job directories
- File-integrity monitoring on sensitive paths and policy enforcement restricting service-account write locations
- Automated security review tools that flag archive extraction without explicit path-boundary checks
Sources
| Source | Link |
|---|---|
| github. | github.com/ |
| nvd. | nvd.nist.gov/ |
| cve. | cve.org/ |
