The vulnerability identified as CVE-2025-71356 affects picklescan versions prior to 0.0.28. The tool fails to detect malicious calls to the function torch.fx.experimental.symbolic_shapes.ShapeEnv.evaluate_guards_expression embedded in pickle files. Because of this detection failure, attackers can craft pickle files containing code that executes remotely upon deserialization by a victim, leading to remote code execution. This vulnerability is categorized under CWE-502, which relates to deserialization of untrusted data. There is no patch or remediation information provided, and no known exploits in the wild have been reported.

Successful exploitation allows remote code execution on the victim's system when a malicious pickle file is loaded. This can lead to full compromise of the affected environment depending on the privileges of the process loading the pickle. The vulnerability arises from inadequate detection of malicious function calls in pickle files by picklescan before version 0.0.28.

Patch status is not yet confirmed — check the vendor advisory for current remediation guidance. Users should upgrade picklescan to version 0.0.28 or later once available, or apply any official fixes provided by the vendor. Until a fix is confirmed, avoid loading pickle files from untrusted sources to prevent potential remote code execution.