CVE-2026-31229
Official description Straight from the sourceThe vendor's or NVD's own wording, published unedited. Authoritative, but often terse — it says what broke, rarely what to do.
NVD · uneditedThe Adversarial Robustness Toolbox (ART) thru 1.20.1 contains an insecure deserialization vulnerability (CWE-502) in its Kubeflow component's model loading functionality. When loading model weights from a file (e.g., model.pt) during robustness evaluation, the code uses torch.load() without the security-restrictive weights_only=True parameter. This allows the deserialization of arbitrary Python objects via the Pickle module. An attacker can exploit this by uploading a maliciously crafted model file to an object storage location referenced by the pipeline, or by controlling the model_id parameter to point to such a file. When the pipeline loads the model, the malicious payload is executed, leading to remote code execution.
Technical summary Written by usOur analysis, written from the advisory, the CVSS vector and the affected-version data. It adds context the advisory leaves out, and never invents facts that are not in the source.
dbcve analysis · high confidenceThe Adversarial Robustness Toolbox (ART) through version 1.20.1 has an insecure deserialization vulnerability (CWE-502) in its Kubeflow component. The code uses torch.load() to load model weights without the security-restrictive weights_only=True parameter, allowing arbitrary Python object deserialization via the Pickle module. An attacker can achieve remote code execution by uploading a malicious model file to object storage or controlling the model_id parameter.
Verify against the referenced sources before acting — the references below are authoritative for this CVE, this summary is not.
CVSS breakdown How the score is builtThe industry scoring standard. It rates how the flaw is reached, what it takes to exploit, and what an attacker gains — the score is derived from those, not the other way round.
From the vector- Attack vector
- Network
- Complexity
- Low
- Privileges
- None
- User interaction
- None
- Scope
- Unchanged
- Confidentiality
- High
- Integrity
- High
- Availability
- High
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
Am I affected? How to checkSteps we derive from the advisory and the affected-version data, so you can decide whether this CVE reaches your setup. They are a guide, not a scan — your own configuration is the authority.
dbcve checksWork through these to decide whether this CVE applies to you.
-
Check the installed ART versionRun 'pip show adversarial-robustness-toolbox' or 'pip list | grep -i art' to see the installed version numberAffected if The installed version is 1.20.1 or lower, or any version where the Kubeflow component uses vulnerable torch.load() calls
-
Verify Kubeflow component is in useCheck if your environment imports or uses the Kubeflow module from ART (e.g., 'from art.attack' or similar Kubeflow-related imports in your codebase)Affected if You have imported or are using the ART Kubeflow component for model loading or inference
-
Inspect torch.load calls in the Kubeflow codeSearch the ART installation for torch.load() calls within Kubeflow-related files: find $(python -c 'import art; print(art.__path__[0])') -name '*kubeflow*' -exec grep -l 'torch.load' {} \;Affected if Found torch.load() calls in Kubeflow component files without weights_only=True parameter
-
Check model loading configurationReview the model loading logic in your application or the Kubeflow component to see if it loads model weights from user-controlled sources or object storage without validationAffected if Your application loads model files via Kubeflow from untrusted sources or allows model_id parameter injection
-
Verify weights_only parameter usageExamine the torch.load() calls found and check if weights_only=True is set, or if the parameter is missing entirelyAffected if The torch.load() calls do NOT have weights_only=True set, leaving deserialization unprotected
You are affected if you use ART version 1.20.1 or lower with the Kubeflow component and any torch.load() call in that component lacks the weights_only=True parameter, especially when loading models from untrusted sources.
Generated from the published advisory. Verify against your own configuration.
Remediation Closing itWhat it takes to close this. Where a vendor fix exists we point at it; where none exists we say so plainly, and can build one. Effort estimates are scoped from the advisory, not from your codebase.
dbcve · scopedAdd weights_only=True to all torch.load() calls in the Kubeflow component, or migrate to a safer serialization format like safetensors that does not allow arbitrary code execution during deserialization.
Latest version of ART beyond 1.20.1 (check PyPI for availability)
- 1. Check the current version of the Adversarial Robustness Toolbox (ART) installed: pip show adversarial-robustness-toolbox
- 2. Upgrade to the latest version of ART that addresses this vulnerability: pip install --upgrade adversarial-robustness-toolbox
- 3. Verify the upgrade was successful and note the new version: pip show adversarial-robustness-toolbox
- 4. If a fixed version is not yet available via pip, manually patch the Kubeflow component code to add weights_only=True parameter to all torch.load() calls: torch.load(file_path, weights_only=True)
- 5. After patching, re-run any model loading operations to confirm the fix works and no errors occur
- 6. Audit other parts of the codebase for similar insecure torch.load() usage patterns and apply the same fix
Generated from the published advisory — verify against the referenced sources before acting.
- Consultation4.0 h
- Implementation6.0 h
- Testing4.0 h
- Review / QA2.0 h
An estimate, not a bill — we confirm scope with you before any work starts. Need it this week? Rush from $4,544.
Scan for this in your stack
Free · runs locallyCheck whether your project pulls in CVE-2026-31229 — or any other known-vulnerable package — straight from your lock files. Free and open source; it runs locally and uploads nothing.
References Go to the primary sourcePrimary sources — vendor advisories, patches and trackers. Where our summary and a reference disagree, the reference wins.
Primary sourcesPractitioner notes
ContributedPeer-ranked notes from engineers who’ve handled CVE-2026-31229 in production — separate from our analysis above.
The advisory tells you what broke. It rarely tells you what actually worked. If you’ve dealt with this one, that detail is what the next engineer is searching for.
- The version that genuinely resolved it — not the one the vendor claimed
- A config change or rule that shut the vector down
- A gotcha in the upgrade path that cost you an afternoon
No notes yet
Be the first to add a field note for this CVE — a mitigation you’ve verified, a version caveat, or a link to a working fix. Sign in above to contribute.
A place for practitioners to share what actually worked: a mitigation you’ve tested, a configuration change, a version- or environment-specific caveat, or a link to a verified patch. The most useful notes rise to the top as peers upvote them, so the signal stays high.
- Verified mitigations, workarounds, and config changes
- Version or environment caveats, and links to real fixes
- No weaponised exploit code, or anything meant to cause harm
- No spam, self-promotion, credentials, or personal data