Pytorch LightningApplication · Lightningai
CVE-2026-58659
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 · uneditedPyTorch Lightning through 2.6.5, fixed in commit d710d68, contains a remote code execution vulnerability in the _load_state function that imports and executes attacker-controlled module names from checkpoint _instantiator hyperparameters. Attackers can craft malicious checkpoint files that bypass weights_only=True protections to execute arbitrary code when LightningModule.load_from_checkpoint is called.
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 confidencePyTorch Lightning versions through 2.6.5 have an RCE vulnerability in the _load_state function where checkpoint _instantiator hyperparameters contain attacker-controlled module names that get imported and executed. Malicious checkpoint files bypass the weights_only=True protection mechanism when load_from_checkpoint is called, allowing arbitrary code execution.
Verify against the referenced sources before acting — the references below are authoritative for this CVE, this summary is not.
Affected products & versions What the vendor confirmedThe version ranges the vendor confirmed as vulnerable. If your version sits inside a range here, treat yourself as exposed until you have upgraded.
NVD · CPE data
Pytorch LightningApplication
<= 2.6.5CVSS 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
- Local
- Complexity
- Low
- Privileges
- None
- User interaction
- Required
- Scope
- Unchanged
- Confidentiality
- High
- Integrity
- High
- Availability
- High
CVSS:3.1/AV:L/AC:L/PR:N/UI:R/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 installed PyTorch Lightning versionRun `pip show pytorch-lightning` or `import pytorch_lightning; print(pytorch_lightning.__version__)`Affected if Version is 2.6.5 or earlier (any version through 2.6.5)
-
Review checkpoint loading code for load_from_checkpoint usageSearch codebase for calls to `load_from_checkpoint` function, especially with weights_only=TrueAffected if Code loads checkpoints from files that could originate from untrusted or external sources
-
Audit checkpoint file sources in your environmentIdentify all checkpoint files loaded by PyTorch Lightning and verify their provenance - check if any are downloaded from untrusted URLs, user uploads, or third-party repositoriesAffected if Any checkpoints are loaded from untrusted or potentially malicious sources
-
Inspect saved checkpoint files for _instantiator entriesOpen checkpoint .ckpt or .pth files in a safe manner (as text/json) and look for dictionary keys containing '_instantiator' or similar nested objects with module path stringsAffected if Checkpoint files contain _instantiator configurations with attacker-controlled module names
-
Check for checkpoint validation before loadingReview whether your codebase validates checkpoint file integrity or source before passing to load_from_checkpointAffected if No validation or trust verification is performed on checkpoint files prior to loading
A defender is affected if they run PyTorch Lightning version 2.6.5 or earlier AND load checkpoint files from untrusted or potentially malicious sources without validation.
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 · scopedUpdate PyTorch Lightning to the version containing commit d710d68 or later. Validate checkpoint file sources before loading and consider disabling _instantiator functionality if not required.
PyTorch Lightning 2.6.6 or latest stable release
- Upgrade PyTorch Lightning to version 2.6.6 or later, which contains the fix for commit d710d68 that addresses the RCE in _load_state function
- If immediate upgrade is not possible, avoid calling LightningModule.load_from_checkpoint with untrusted checkpoint files
- Verify that any checkpoint files loaded via load_from_checkpoint are from trusted sources
- If using checkpoints from external users or untrusted sources, review the checkpoint contents manually before loading
Generated from the published advisory — verify against the referenced sources before acting.
Fix this in Pytorch Lightning
Scoped from the published advisory
- Consultation2.0 h
- Implementation1.0 h
- Testing3.0 h
- Review / QA1.0 h
An estimate, not a bill — we confirm scope with you before any work starts. Need it this week? Rush from $1,936.
Scan for this in your stack
Free · runs locallyCheck whether your project pulls in CVE-2026-58659 — 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-2026-58659 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