CVE-2025-71345 picklescan - Arbitrary Code Execution via torch.utils.bottleneck.__main__.run_autograd_prof
Attackers can execute arbitrary code by bypassing picklescan's detection of malicious pickle files invoking specific PyTorch functions.
- 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-2025-71345?
CVE-2025-71345 is a high-severity unsafe deserialization vulnerability in picklescan, published on 4 Jul 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
picklescan before 0.0.30 fails to detect malicious pickle files that invoke torch.utils.bottleneck.__main__.run_autograd_prof function. Attackers can embed undetected code in pickle files that executes during deserialization, enabling remote code execution.
Weakness: CWE-502
Which versions are affected by CVE-2025-71345?
2 open source packages ship an affected version, including picklescan.
| Package | Affected | Fixed in |
|---|---|---|
picklescan PyPI | All versions | 0.0.30 |
picklescan PyPI | All versions | 0.0.30 |
How do I fix CVE-2025-71345?
Upgrade picklescan to 0.0.30 or later.
Upgrade, and only deserialise data from trusted sources or with an allow-list of types. More about Unsafe deserialization.
Is CVE-2025-71345 being exploited?
It is not confirmed as exploited in the wild, but public exploit code for CVE-2025-71345 exists.
Sources
Also known as GHSA-4whj-rm5r-c2v8, GHSA-vwrw-h95r-ch46.