speechbrain is vulnerable to Deserialization of Untrusted Data
88
CVE Score
High
Deserialization of Untrusted Data
The checkpoint recovery logic in the Checkpointer parses CKPT.yaml metadata files with PyYAML's unsafe Loader. During candidate enumeration in recover_if_possible(), any checkpoint metadata file found in the configured checkpoint path is deserialized, so metadata containing Python object construction tags instantiates arbitrary objects. Loading checkpoints from a path that contains untrusted metadata therefore leads to arbitrary code execution, even when the malicious checkpoint is never selected for recovery. The fix switches checkpoint metadata parsing to SafeLoader.
Upgrade the speechbrain library to the patch version.
You are affected if you are using a version that falls within the vulnerable range and you recover checkpoints from a path that can contain untrusted checkpoint metadata files.
0.0.1 - 1.1.0
Fixed in 1.1.1speechbrain is vulnerable to Deserialization of Untrusted Data in versions 0.0.1 - 1.1.0.
Secure your code, cloud, and runtime in one central system. Connect a repo to discover what the reasoning agents find in your codebase.
No credit card required | Scan results in 32 seconds.
I consent to receiving marketing communications based on Aikido’s Privacy Policy.
SOC 2Compliant
ISO 27001Compliant
ISO 42001Compliant© All Intel data is openly available and commercially licensed.