CVE-2026-55832: CWE-22: Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal') in sonos tract
Tract is a tiny, no-nonsense, self-contained TensorFlow and ONNX inference toolkit. Prior to 0.21.17, 0.22.3, and 0.23.2, the tract-onnx crate passes the attacker-controlled external_data location from an ONNX model through onnx/src/tensor.rs get_external_resources and joins the value to the model directory without rejecting absolute paths or parent directory components. Loading an untrusted model through model_for_path can therefore make onnx/src/data_resolver.rs MmapDataResolver open an arbitrary local file and place the file contents into model tensors or inference output. Attacker-controlled offset and length fields can also select an out-of-range mapping slice and cause a denial of service, but the flaw does not write files or execute code. This issue is fixed in versions 0.21.17, 0.22.3, and 0.23.2.
AI Analysis
Technical Summary
The vulnerability arises because the tract-onnx crate passes an attacker-controlled external_data location from an ONNX model through onnx/src/tensor.rs get_external_resources and concatenates it with the model directory without validating or rejecting absolute paths or parent directory components. This allows an attacker loading an untrusted model via model_for_path to cause onnx/src/data_resolver.rs MmapDataResolver to open arbitrary local files and place their contents into model tensors or inference outputs. Crafted offset and length fields can cause out-of-range memory mapping slices leading to denial of service. The flaw does not enable writing files or code execution. The vulnerability is addressed in versions 0.21.17, 0.22.3, and 0.23.2.
Potential Impact
An attacker can cause the software to read arbitrary local files and load their contents into model tensors or inference outputs, potentially exposing sensitive data. The vulnerability also allows denial of service via out-of-range memory mapping slices. There is no impact on code integrity or availability beyond denial of service, and no code execution or file write capabilities are present.
Mitigation Recommendations
This vulnerability is fixed in sonos tract versions 0.21.17, 0.22.3, and 0.23.2. Users should upgrade to these or later versions to remediate the issue. No other mitigation steps are indicated by the vendor advisory.
CVE-2026-55832: CWE-22: Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal') in sonos tract
Description
Tract is a tiny, no-nonsense, self-contained TensorFlow and ONNX inference toolkit. Prior to 0.21.17, 0.22.3, and 0.23.2, the tract-onnx crate passes the attacker-controlled external_data location from an ONNX model through onnx/src/tensor.rs get_external_resources and joins the value to the model directory without rejecting absolute paths or parent directory components. Loading an untrusted model through model_for_path can therefore make onnx/src/data_resolver.rs MmapDataResolver open an arbitrary local file and place the file contents into model tensors or inference output. Attacker-controlled offset and length fields can also select an out-of-range mapping slice and cause a denial of service, but the flaw does not write files or execute code. This issue is fixed in versions 0.21.17, 0.22.3, and 0.23.2.
CVSS v3.1
Score 6.1medium
Affected software
sonos
tract
pkg:cargo/sonos/tract-onnxRun on your own infrastructure? Check whether these packages are installed with threat-finder — our free open-source scanner.
Weaknesses
AI-Powered Analysis
Machine-generated threat intelligence
Technical Analysis
The vulnerability arises because the tract-onnx crate passes an attacker-controlled external_data location from an ONNX model through onnx/src/tensor.rs get_external_resources and concatenates it with the model directory without validating or rejecting absolute paths or parent directory components. This allows an attacker loading an untrusted model via model_for_path to cause onnx/src/data_resolver.rs MmapDataResolver to open arbitrary local files and place their contents into model tensors or inference outputs. Crafted offset and length fields can cause out-of-range memory mapping slices leading to denial of service. The flaw does not enable writing files or code execution. The vulnerability is addressed in versions 0.21.17, 0.22.3, and 0.23.2.
Potential Impact
An attacker can cause the software to read arbitrary local files and load their contents into model tensors or inference outputs, potentially exposing sensitive data. The vulnerability also allows denial of service via out-of-range memory mapping slices. There is no impact on code integrity or availability beyond denial of service, and no code execution or file write capabilities are present.
Mitigation Recommendations
This vulnerability is fixed in sonos tract versions 0.21.17, 0.22.3, and 0.23.2. Users should upgrade to these or later versions to remediate the issue. No other mitigation steps are indicated by the vendor advisory.
Technical Details
- Data Version
- 5.2
- Assigner Short Name
- GitHub_M
- Date Reserved
- 2026-06-17T16:29:38.864Z
- Cvss Version
- 3.1
- State
- PUBLISHED
Threat ID: 6aa836a955bf5e2cf5684393
Added to database: 09/14/2026, 18:02:17 UTC
Last enriched: 09/14/2026, 18:17:12 UTC
Last updated: 09/15/2026, 03:35:01 UTC
Views: 8
Community Reviews
0 reviewsCrowdsource mitigation strategies, share intel context, and vote on the most helpful responses. Sign in to add your voice and help keep defenders ahead.
Want to contribute mitigation steps or threat intel context? Sign in or create an account to join the community discussion.
Actions
Updates to AI analysis require Pro Console access. Upgrade inside Console → Billing.
Need more coverage?
Upgrade to Pro Console for AI refresh and higher limits.
For incident response and remediation, OffSeq services can help resolve threats faster.
Latest Threats
Check if your credentials are on the dark web
Instant breach scanning across billions of leaked records. Free tier available.