Vulnerability feed

Executive summary

A code injection flaw affects Sony NNabla Neural Network Libraries up to version 1.39.0. Converting a malicious model file can run attacker-chosen commands on the machine doing the conversion, and the advisory rates it High with a CVSS v3.1 score of 8.8. The advisory states that the project is end of life and no upstream fix is expected, so it recommends not converting .nnp files from untrusted sources until a patched fork exists.

Technical summary

The flaw is CWE-94 code injection: the NNP converter passes the attacker-controlled function.type field into Python eval() calls in func_set_nnabla_version_decorate (utils.py lines 111 and 116). A duplicate network name lets a malicious second network bypass the version gate, which only checks the first same-name network. Exploitation requires the victim to run nnabla_cli convert on a crafted .nnp file (network attack vector, user interaction required, no privileges). The advisory states that the affected version is 1.39.0 and that the earliest affected version is unknown, and it gives no fixed version because the project is EOL. Recommended fixes are replacing eval() with getattr/hasattr, validating function.type against known types, and enforcing unique network names. The advisory does not state any detection guidance, and no vendor advisory is listed.

Summaries are generated by AI from the linked advisory text. Confirm versions and fixes at the source.

Summarized from github.com/3em0/cve_repo/blob/main/2026…n-in-NNP-Converter-('function.type'-eval()).md

Impact if exploited

How it is exploited

Attack vector

Network

Attack complexity

Low

Attack requirements

None

Privileges required

None

User interaction

Passive

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:P/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

Affects

SonySony Neural Network LibrariesSony NNabla