CVE-2026-24233
NVIDIA TensorRT-LLM for Linux contains a vulnerability in the restricted unpickler used for model weight deserialization, where a local, una
CVSS
8.4
High
EPSS
0.4%
p33
KEV
—
Exploit Today
10
0-100
Published: Jul 14, 2026 · Last modified: Jul 15, 2026 · CWE-502
0.4%EPSS · 30 days0.4%
2026-08-242026-09-21
NVIDIA TensorRT-LLM for Linux contains a vulnerability in the restricted unpickler used for model weight deserialization, where a local, unauthenticated attacker could cause deserialization of untrusted data. A successful exploit of this vulnerability might lead to code execution, escalation of privileges, data tampering, and information disclosure.
CVECVSSEPSSKEVRExploitTitleMod.
CVE-2026-918277.5 HIG—
———The Ninja Forms WordPress plugin 3.15.3 does not prevent user-submitted form field values from being deserialised when an administrator later exports form submissions to CSV, allowing unauthenticated attackers to perform PHP Object Injection; if a suitable POP chain is present via another installed plugin or theme, this can lead to actions such as arbitrary file operations or remote code execution.6hCVE-2026-196589.8 CRI—
———The Give Tributes plugin for WordPress is vulnerable to PHP Object Injection in all versions up to, and including, 2.3.1 via deserialization of untrusted input . This makes it possible for unauthenticated attackers to inject a PHP Object. No known POP chain is present in the vulnerable software, which means this vulnerability has no impact unless another plugin or theme containing a POP chain is installed on the site. If a POP chain is present via an additional plugin or theme installed on the target system, it may allow the attacker to perform actions like delete arbitrary files, retrieve sensitive data, or execute code depending on the POP chain present. This vulnerability is only reachable when the "Allow Multiple Recipients" option is enabled for the donation form, as the single-recipient code path applies sanitize_textarea_field() which would neutralize the payload. Exploitation additionally requires the eCard "Custom Message" option to be disabled, which is the plugin default: when it is enabled the personalized message becomes a required field and GiveWP's give_clean() blanks serialized input during validation, causing the donation to be rejected before it is stored.6hCVE-2026-943019.8 CRI—
——0The fix for CVE-2026-47065/ZDRES-232 ("resolveProxyClass Not Overridden - acceptMatchers Filter Bypass via java.lang.reflect.Proxy"), released on 2026-06-02 and announced as "Fully addressed" in MINA 2.2.8, 2.1.13 and 2.0.29, was committed to the
2.2.X branch only. The 2.0.X and 2.1.X maintenance branches never received the resolveProxyClass() override, so the 2.0.29 and 2.1.13 artifacts listed as fixed -- and every later release on those lines, up to and including the current 2.0.30 and 2.1.14 -- remain vulnerable to the exact allow-list bypass that CVE-2026-47065 was meant to close.13hCVE-2026-940936.3 MED17.8%
——5A security vulnerability has been detected in DLR-RM stable-baselines3 up to 2.9.0. This affects the function PPO.load/load_replay_buffer/VecNormalize.load of the file save_util.py. Such manipulation leads to deserialization. It is possible to launch the attack remotely. The exploit has been disclosed publicly and may be used. In v2.9.0 the PyTorch tensor load path is hardened (weights_only=True), but that hardening was later reverted on master via PR #1913 "Hotfix: revert loading with weights_only=True" [blocked] to fix PyTorch 1.13 compat - so even the one "safe" path is inconsistent across versions. #2281 was closed as a duplicate of #1831 since both are unsafe pickle deserialization - but #1831's fix (PR #41) only gated the Hugging Face Hub loader in the separate huggingface_sb3 package. This finding covers the core stable_baselines3 load APIs (PPO.load, load_replay_buffer, VecNormalize.load), which have no safe mode or gate and remained exploitable in v2.9.0 until the outstanding hardening (PR #2264) ships.1dCVE-2026-940925.5 MED8.9%
——3A vulnerability was detected in dmlc dgl up to 2.1.0. This impacts the function load_info/_read_torch_data of the file utils.py. Performing a manipulation of the argument path results in deserialization. The attack can be initiated remotely. The exploit is now public and may be used. The project was informed of the problem early through an issue report but has not responded yet.1dCVE-2026-940915.5 MED9.3%
——3A weakness has been identified in piskvorky gensim up to 4.4.0. The impacted element is the function Load of the file gensim/utils.py of the component Model Loader. This manipulation of the argument fname causes deserialization. It is possible to initiate the attack remotely. The exploit has been made available to the public and could be used for attacks. Maintainer closed #3663 same-day with no comment, PR, or fix; repo's last push (2025-11-01) predates the report, so the unsafe pickle.load in SaveLoad.load remains unguarded at develop HEAD.1d