CVE-2026-31238
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 Ludwig framework thru 0.10.4 is vulnerable to insecure deserialization (CWE-502) in its model serving component. When starting a model server with the ludwig serve command, the framework loads model weight files using torch.load() without enabling the security-restrictive weights_only=True parameter. This default behavior allows the deserialization of arbitrary Python objects via the pickle module. An attacker can exploit this by providing a maliciously crafted PyTorch model file, leading to arbitrary code execution on the system hosting the Ludwig model server.
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 Ludwig framework through version 0.10.4 is vulnerable to insecure deserialization (CWE-502) when serving models via the ludwig serve command. The framework loads model weight files using torch.load() without the security-restrictive weights_only=True parameter, which enables deserialization of arbitrary Python objects through the pickle module. An attacker can achieve arbitrary code execution by providing a maliciously crafted PyTorch model file to the model server.
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.
-
Identify Ludwig installation versionRun 'pip show ludwig' or 'ludwig --version' to confirm the installed version is 0.10.4 or earlierAffected if Installed version is 0.10.4 or any version prior to a fix that adds weights_only=True to torch.load() calls
-
Confirm ludwig serve command is in useCheck if the 'ludwig serve' CLI command is being used to serve models, or inspect any service that invokes the Ludwig serving APIAffected if The ludwig serve command or serving API is actively used to load and serve PyTorch model files
-
Locate torch.load() calls in Ludwig codebaseSearch the Ludwig installation directory for torch.load() calls without the weights_only parameter: grep -r 'torch.load' <ludwig_path> | grep -v weights_onlyAffected if torch.load() calls are found that omit the weights_only=True argument, which is required to prevent unsafe deserialization
-
Verify model loading without weights_only protectionInspect the specific code paths in Ludwig that handle model weight loading during the serve operation, confirming they do not pass weights_only=True to torch.load()Affected if Model weight files are loaded via torch.load() without the weights_only=True security restriction enabled
-
Assess model file trust boundaryDetermine whether untrusted or externally-supplied PyTorch model files (.pt, .pth) can be submitted to the ludwig serve endpointAffected if The serving endpoint accepts model files from untrusted sources without prior validation or integrity checking
A user is affected if they run ludwig serve with a version <= 0.10.4 that loads model files via torch.load() without weights_only=True and accepts model input 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.
From vendor dataUpgrade to a version that implements weights_only=True in torch.load() calls or ensure all untrusted model files are validated before loading. If immediate upgrade is not possible, implement model file integrity validation and network isolation for model serving endpoints.
- Consultation3.0 h
- Implementation6.0 h
- Testing6.0 h
- Review / QA3.0 h
An estimate, not a bill — we confirm scope with you before any work starts. Need it this week? Rush from $4,992.
Scan for this in your stack
Free · runs locallyCheck whether your project pulls in CVE-2026-31238 — 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-2026-31238 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