Deserialization of untrusted data in the LanguageModel class of Flair from versions 0.4.1 to latest are vulnerable to arbitrary code execution when loading a malicious model.
Queued for AI Analysis
This CVE will be auto-analyzed on the next cron run.