CVE-2026-71281 peft Unsafe Deserialization via torch.load() Without weights_only in LoRA-GA and CorDA Modules
Attackers can execute arbitrary code by providing malicious cache files to peft's LoRA-GA and CorDA modules due to unsafe torch.load usage.
- Exploited in the wild
- Not known
- Login needed
- Not needed
- Attackable
- Over the internet
- Victim action needed
- A user has to click or open something
- Fix
- None listed
What is CVE-2026-71281?
CVE-2026-71281 is a high-severity unsafe deserialization vulnerability in huggingface peft, published on 5 Aug 2026. An attacker can exploit it over the internet without logging in, and only if a user opens a file or follows a link.
Official description
Hugging Face peft's LoRA-GA and CorDA initialization modules (src/peft/tuners/lora/corda.py lines ~102 and ~163, and src/peft/tuners/lora/loraga.py line ~101) call torch.load on config-specified cache/covariance files without weights_only=True, bypassing peft's own safe-loading wrapper used elsewhere in the codebase.
Weakness: CWE-502
Which versions are affected by CVE-2026-71281?
How do I fix CVE-2026-71281?
No fix has been published for CVE-2026-71281 yet. Limit exposure of the affected software until one is available.
Upgrade, and only deserialise data from trusted sources or with an allow-list of types. More about Unsafe deserialization.
Is CVE-2026-71281 being exploited?
There is no known exploitation of CVE-2026-71281 in the wild and no public exploit code.
Sources
Also known as GHSA-g7pc-47rc-wvwf.