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.
Attack Parameters
Attack Vector
Local
Requires local access
Attack Complexity
Low
Easy to exploit
Attack Requirements
None
No additional conditions
Privileges Required
None
No privileges needed
User Interaction
Active
User action required
Impact Assessment
Confidentiality
High
Complete data leak
Integrity
High
Complete data modification
Availability
High
Complete denial of service
CVSS Vector v4.0
Weakness Type (CWE)
Vulnerable Products 1
| Configuration | From (including) | Up to (excluding) |
|---|---|---|
|
Lightningai Pytorch_Lightning
cpe:2.3:a:lightningai:pytorch_lightning:*:*:*:*:*:python:*:*
|
— |
<= 2.6.5
|
References 4
https://github.com/Lightning-AI/pytorch-lightning/commit/d710d689510d50e800f53b…
disclosure@vulncheck.com
https://github.com/Lightning-AI/pytorch-lightning/issues/21822
disclosure@vulncheck.com
https://github.com/Lightning-AI/pytorch-lightning/pull/21832
disclosure@vulncheck.com
https://www.vulncheck.com/advisories/pytorch-lightning-arbitrary-code-execution…
disclosure@vulncheck.com