CVE-2025-71370

picklescan - Remote Code Execution via torch.jit.unsupported_tensor_ops.execWrapper

Published Jun 23, 2026 CVSS 4.0 · 7.6 High EPSS < 1% KEV · Not listed
Description
picklescan before 0.0.28 fails to detect malicious torch.jit.unsupported_tensor_ops.execWrapper function calls embedded in pickle files. Attackers can craft malicious pickle files that bypass picklescan detection and execute arbitrary code when loaded via pickle.load().

Analysis

AI-generated

Technical impact

The vulnerability in picklescan allows attackers to embed torch.jit.unsupported_tensor_ops.execWrapper calls within a pickle file. When a victim loads such a file using pickle.load(), the op executes arbitrary code in the context of the process. This flaw directly compromises the confidentiality, integrity, and availability of the affected system, enabling full remote code execution on the target machine.

Affected Systems

The vulnerability affects all installations of the picklescan tool before version 0.0.28. No specific sub‑versions are listed; any build of picklescan lower than 0.0.28 inherits the flaw.

Risk and Exploitability

The CVSS score of 7.6 indicates high severity. While the EPSS score is not available, the lack of listing in the CISA KEV catalog does not diminish the risk; the flaw can be exploited by crafting a malicious pickle and supplying it to any process that uses picklescan to load it. Attackers do not require special privileges to initiate the exploit; merely delivering a well‑formed pickle payload is sufficient. The attack vector is inferred to be local or remote depending on how the victim loads the pickle data, but the lack of required network permissions suggests that local compromise can further leverage the vulnerability for broader impact.

Remediation

No solution or workaround provided in the CVE record.

OpenCVE Recommended Actions

  • Upgrade picklescan to version 0.0.28 or later to receive the security fix
  • Avoid using pickle.load() on data from untrusted sources; switch to safer serialization mechanisms such as JSON or Protocol Buffers
  • If an immediate upgrade is not possible, mitigate by disabling torch.jit.unsupported_tensor_ops.execWrapper during pickle deserialization or implement a custom whitelist that explicitly blocks this operation
  • Enforce strict input validation and review the code path that processes pickle content to ensure no execWrapper calls are executed

Generated by OpenCVE AI on June 23, 2026 at 13:27 UTC.

Tracking

Sign in to view the affected projects.

Advisories

Source ID Title
Github GHSA Github GHSA GHSA-vr7h-p6mm-wpmh Picklescan missing detection when calling pytorch function torch.jit.unsupported_tensor_ops.execWrapper
History
Type Values Removed Values Added
Metrics ssvc

{'options': {'Automatable': 'no', 'Exploitation': 'poc', 'Technical Impact': 'total'}, 'version': '2.0.3'}

Type Values Removed Values Added
Description picklescan before 0.0.28 fails to detect malicious torch.jit.unsupported_tensor_ops.execWrapper function calls embedded in pickle files. Attackers can craft malicious pickle files that bypass picklescan detection and execute arbitrary code when loaded via pickle.load().
Title picklescan - Remote Code Execution via torch.jit.unsupported_tensor_ops.execWrapper
First Time appeared Mmaitre314
Mmaitre314 picklescan
Weaknesses CWE-502
CPEs cpe:2.3:a:mmaitre314:picklescan:*:*:*:*:*:*:*:*
Vendors & Products Mmaitre314
Mmaitre314 picklescan
References
Metrics cvssV3_1

{'score': 8.1, 'vector': 'CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:N'}

cvssV4_0

{'score': 7.6, 'vector': 'CVSS:4.0/AV:N/AC:L/AT:P/PR:N/UI:P/VC:H/VI:H/VA:N/SC:N/SI:N/SA:N'}

Weaknesses
  • CWE-502

    Deserialization of Untrusted Data

Sources

Status
PUBLISHED
Assigner
VulnCheck
Published
Updated
Jun 23, 2026, 1:37 PM
Reserved
Jun 20, 2026, 1:01 PM
Status
Deferred
Published
Jun 23, 2026, 1:16 PM
Modified
Oct 12, 2026, 9:10 PM