ESPnet before 202609 deserializes pretrained model checkpoints using torch.load with weights_only=False, allowing arbitrary code execution from attacker-supplied files. Attackers can craft malicious checkpoint files that execute code during deserialization when loaded through the initialization or fine-tuning path.
Attack Parameters
Attack Vector
Network
Can be exploited remotely
Attack Complexity
Low
Easy to exploit
Attack Requirements
None
No additional conditions
Privileges Required
None
No privileges needed
User Interaction
Passive
Minimal interaction
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
espnet:espnet
References 6
https://github.com/espnet/espnet/security/advisories/GHSA-64f6-3gqc-r926
GitHub Security Advisory (GHSA-64f6-3gqc-r926)
https://github.com/espnet/espnet/commit/91ca045fc179f29bc7b7fe05cd852aea65310003
Patch Commit
https://github.com/espnet/espnet/blob/v.202511/espnet2/torch_utils/load_pretrai…
Unsafe torch.load call at v.202511
https://github.com/espnet/espnet/releases/tag/v.202609
ESPnet v.202609 Release Notes
https://github.com/espnet/espnet
NVD
https://www.vulncheck.com/advisories/espnet-before-202609-remote-code-execution…
VulnCheck Advisory: ESPnet before 202609 Remote Code Execution via Unsafe Deserialization