CVE-2026-112304 docTR through 1.1.0 Unsafe Deserialization RCE via load_pretrained_params()
Attackers can execute arbitrary Python code by supplying malicious checkpoints to docTR's parameter loading function.
- 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-2026-112304?
CVE-2026-112304 is a high-severity unsafe deserialization vulnerability in Mindee docTR, published on 19 Oct 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
docTR through 1.1.0 contains an unsafe deserialization vulnerability in load_pretrained_params() that loads checkpoints with torch.load without weights_only=True. Attackers who control a local or URL-supplied checkpoint can embed a malicious pickle payload to execute Python code with the loading process's privileges on torch releases before 2.6.
Weakness: CWE-502
Which versions are affected by CVE-2026-112304?
How do I fix CVE-2026-112304?
No fix has been published for CVE-2026-112304 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-2026-112304 being exploited?
It is not confirmed as exploited in the wild, but public exploit code for CVE-2026-112304 exists.
Sources
Also known as GHSA-rh68-f334-2256.