torch is vulnerable to Server-Side Request Forgery (SSRF)
54
Medium Risk
PyTorch bundles the third-party tensorpipe library, which statically links an outdated copy of libuv into distributed C++ binaries such as libtorch_cpu.so. That libuv copy truncates hostnames longer than 256 characters before passing them to getaddrinfo, so a crafted long hostname can be made to resolve to an unintended address. Code paths that resolve externally influenced hostnames can therefore be steered toward internal destinations, enabling server-side request forgery. The fix updates the bundled libuv dependency to a version that no longer truncates hostnames unsafely.
You are affected if you are using a version that falls within the vulnerable range and your application resolves externally influenced hostnames through the affected libuv code path (for example via distributed/RPC features).
torch is vulnerable to Server-Side Request Forgery (SSRF) in versions 0.0.1 - 2.8.0.
Upgrade the torch library to the patch version.
Connect your repositories to instantly see whether vulnerable or malicious packages exist in your codebase.
Free. No credit card required.

I consent to receiving marketing communications based on Aikido’s Privacy Policy.
SOC 2Compliant
ISO 27001Compliant