CVE-2026-58659: Use of Externally-Controlled Input to Select Classes or Code ('Unsafe Reflection') in Lightning-AI pytorch-lightning
PyTorch 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.
AI Analysis
Technical Summary
CVE-2026-58659 is a remote code execution vulnerability in PyTorch Lightning through version 2.6.5. The issue arises from the _load_state function importing and executing attacker-controlled module names derived from checkpoint _instantiator hyperparameters. This unsafe reflection allows crafted malicious checkpoint files to bypass protections intended to restrict code execution (weights_only=True) and execute arbitrary code during the loading of a LightningModule checkpoint. The vulnerability has a CVSS 4.0 score of 8.4, indicating high severity. There is no explicit vendor advisory or patch information provided, and the remediation level is not specified.
Potential Impact
An attacker who can supply a malicious checkpoint file can achieve remote code execution on systems loading that checkpoint with PyTorch Lightning versions up to 2.6.5. This bypasses intended protections and allows arbitrary code execution in the context of the application using LightningModule.load_from_checkpoint.
Mitigation Recommendations
Patch status is not yet confirmed — check the vendor advisory for current remediation guidance. Until an official fix is available, avoid loading untrusted checkpoint files with LightningModule.load_from_checkpoint to prevent exploitation.
CVE-2026-58659: Use of Externally-Controlled Input to Select Classes or Code ('Unsafe Reflection') in Lightning-AI pytorch-lightning
Description
PyTorch 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.
CVSS v4.0
Score 8.4high
Affected software
pkg:github/pytorch-lightningRun on your own infrastructure? Check whether these packages are installed with threat-finder — our free open-source scanner.
AI-Powered Analysis
Machine-generated threat intelligence
Technical Analysis
CVE-2026-58659 is a remote code execution vulnerability in PyTorch Lightning through version 2.6.5. The issue arises from the _load_state function importing and executing attacker-controlled module names derived from checkpoint _instantiator hyperparameters. This unsafe reflection allows crafted malicious checkpoint files to bypass protections intended to restrict code execution (weights_only=True) and execute arbitrary code during the loading of a LightningModule checkpoint. The vulnerability has a CVSS 4.0 score of 8.4, indicating high severity. There is no explicit vendor advisory or patch information provided, and the remediation level is not specified.
Potential Impact
An attacker who can supply a malicious checkpoint file can achieve remote code execution on systems loading that checkpoint with PyTorch Lightning versions up to 2.6.5. This bypasses intended protections and allows arbitrary code execution in the context of the application using LightningModule.load_from_checkpoint.
Mitigation Recommendations
Patch status is not yet confirmed — check the vendor advisory for current remediation guidance. Until an official fix is available, avoid loading untrusted checkpoint files with LightningModule.load_from_checkpoint to prevent exploitation.
Technical Details
- Data Version
- 5.2
- Assigner Short Name
- VulnCheck
- Date Reserved
- 2026-07-01T21:54:37.946Z
- Cvss Version
- 4.0
- State
- PUBLISHED
- Remediation Level
- null
Threat ID: 6a57c45368715ace431fb2a0
Added to database: 07/15/2026, 17:33:07 UTC
Last enriched: 07/22/2026, 22:48:40 UTC
Last updated: 08/29/2026, 10:52:09 UTC
Views: 75
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.