Deserialization of Untrusted Data in GitHub repository huggingface/transformers prior to 4.36.

Applio is a voice conversion tool. Versions 3.2.8-bugfix and prior are vulnerable to unsafe deserialization in model_information.py. `model_name` in model_information.py takes user-supplied input (e.g. a path to a model) and pass that value to the `run_model_information_script` and later to `model_information` function, which loads that model with `torch.load` in rvc/train/process/model_information.py (on line 16 in 3.2.8-bugfix), which is vulnerable to unsafe deserialization. The issue can lead to remote code execution. A patch is available in the `main` branch of the repository.

In Progress Telerik UI for WinForms versions prior to 2024 Q4 (2024.4.1113), a code execution attack is possible through an insecure deserialization vulnerability.

OpenClaw before 2026.4.20 fails to properly reserve the OPENCLAW_ runtime-control environment namespace in workspace dotenv files, allowing attackers to override critical runtime variables. Malicious workspaces can set variables like OPENCLAW_GIT_DIR to manipulate trusted OpenClaw runtime behavior during source-update or installer flows.

Zed is a code editor. Prior to 0.229.0, Zed's terminal tool permission system can be bypassed by prepending environment variable assignments to allowlisted commands, hijacking program behavior (e.g., PAGER) to execute arbitrary code. This vulnerability is fixed in 0.229.0.

In Progress Telerik UI for WPF versions prior to 2024 Q3 (2024.3.924), a code execution attack is possible through an insecure deserialization vulnerability.

PySpector is a static analysis security testing (SAST) Framework engineered for modern Python development workflows. The plugin security validator in PySpector uses AST-based static analysis to prevent dangerous code from being loaded as plugins. Prior to version 0.1.8, the blocklist implemented in `PluginSecurity.validate_plugin_code` is incomplete and can be bypassed using several Python constructs that are not checked. An attacker who can supply a plugin file can achieve arbitrary code execution within the PySpector process when that plugin is installed and executed. Version 0.1.8 fixes the issue.

A CWE-502: Deserialization of Untrusted Data vulnerability exists in the Dashboard module that could cause an interpretation of malicious payload data, potentially leading to remote code execution when an attacker gets the user to open a malicious file.

Connected Components Workbench (v13.00.00 and prior), ISaGRAF Workbench (v6.0 though v6.6.9), and Safety Instrumented System Workstation (v1.2 and prior (for Trusted Controllers)) do not limit the objects that can be deserialized. This allows attackers to craft a malicious serialized object that, if opened by a local user in Connected Components Workbench, may result in arbitrary code execution. This vulnerability requires user interaction to be successfully exploited

PySpector is a static analysis security testing (SAST) Framework engineered for modern Python development workflows. PySpector versions 0.1.6 and prior are affected by a security validation bypass in the plugin system. The validate_plugin_code() function in plugin_system.py, performs static AST analysis to block dangerous API calls before a plugin is trusted and executed. However, the internal resolve_name() helper only handles ast.Name and ast.Attribute node types, returning None for all others. When a plugin uses indirect function calls via getattr() (such as getattr(os, 'system')) the outer call's func node is of type ast.Call, causing resolve_name() to return None, and the security check to be silently skipped. The plugin incorrectly passes the trust workflow, and executes arbitrary system commands on the user's machine when loaded. This issue has been patched in version 0.1.7.

A vulnerability in the Snake YAML parser of Magnolia CMS v6.2.3 and below allows attackers to execute arbitrary code via a crafted YAML file.

Project files are stored memory objects in the form of binary serialized data that can later be read and deserialized again to instantiate the original objects in memory. Malicious manipulation of these files may allow an attacker to corrupt memory.

PyTorch-Lightning versions 2.6.0 and earlier contain an insecure deserialization vulnerability (CWE-502) in the checkpoint loading mechanism. The LightningModule.load_from_checkpoint() method, which is commonly used to load saved model states, internally calls torch.load() without setting the security-restrictive weights_only=True parameter. This default behavior allows the deserialization of arbitrary Python objects via the Pickle module. A remote attacker can exploit this by providing a maliciously crafted checkpoint file, leading to arbitrary code execution on the victim's system when the file is loaded.

pytorch-lightning is vulnerable to Deserialization of Untrusted Data

PyDrive2 is a wrapper library of google-api-python-client that simplifies many common Google Drive API V2 tasks. Unsafe YAML deserilization will result in arbitrary code execution. A maliciously crafted YAML file can cause arbitrary code execution if PyDrive2 is run in the same directory as it, or if it is loaded in via `LoadSettingsFile`. This is a deserilization attack that will affect any user who initializes GoogleAuth from this package while a malicious yaml file is present in the same directory. This vulnerability does not require the file to be directly loaded through the code, only present. This issue has been addressed in commit `c57355dc` which is included in release version `1.16.2`. Users are advised to upgrade. There are no known workarounds for this vulnerability.

ADB Explorer is a fluent UI for ADB on Windows. Prior to Beta 0.9.26020, ADB Explorer is vulnerable to Insecure Deserialization leading to Remote Code Execution. The application attempts to deserialize the App.txt settings file using Newtonsoft.Json with TypeNameHandling set to Objects. This allows an attacker to supply a crafted JSON file containing a gadget chain (e.g., ObjectDataProvider) to execute arbitrary code when the application launches and subsequently saves its settings. This vulnerability is fixed in Beta 0.9.26020.

PowerDocu contains a Windows GUI executable to perform technical documentations. Prior to 2.4.0, PowerDocu contains a critical security vulnerability in how it parses JSON files within Flow or App packages. The application blindly trusts the $type property in JSON files, allowing an attacker to instantiate arbitrary .NET objects and execute code. This vulnerability is fixed in 2.4.0.

In Progress Telerik UI for WPF versions prior to 2024 Q4 (2024.4.1111), a code execution attack is possible through an insecure deserialization vulnerability.

mrdoc is vulnerable to Deserialization of Untrusted Data

NVIDIA Model Optimizer for Windows and Linux contains a vulnerability in the ONNX quantization feature, where a user could cause unsafe deserialization by providing a specially crafted input file. A successful exploit of this vulnerability might lead to code execution, escalation of privileges, data tampering, and information disclosure.

NVIDIA Megatron-LM contains a vulnerability in checkpoint loading where an Attacker may cause an RCE by convincing a user to load a maliciously crafted file. A successful exploit of this vulnerability may lead to code execution, escalation of privileges, information disclosure, and data tampering.

NVIDIA Megatron-LM contains a vulnerability in inferencing where an Attacker may cause an RCE by convincing a user to load a maliciously crafted input. A successful exploit of this vulnerability may lead to code execution, escalation of privileges, information disclosure, and data tampering.

NVIDIA Megatron-LM contains a vulnerability in checkpoint loading where an Attacker may cause an RCE by convincing a user to load a maliciously crafted file. A successful exploit of this vulnerability may lead to code execution, escalation of privileges, information disclosure, and data tampering.

NVIDIA BioNeMo contains a vulnerability where a user could cause a deserialization of untrusted data. A successful exploit of this vulnerability might lead to code execution, denial of service, information disclosure, and data tampering.

There exists a use after free vulnerability in Reverb. Reverb supports the VARIANT datatype, which is supposed to represent an arbitrary object in C++. When a tensor proto of type VARIANT is unpacked, memory is first allocated to store the entire tensor, and a ctor is called on each instance. Afterwards, Reverb copies the content in tensor_content to the previously mentioned pre-allocated memory, which results in the bytes in tensor_content overwriting the vtable pointers of all the objects which were previously allocated. Reverb exposes 2 relevant gRPC endpoints: InsertStream and SampleStream. The attacker can insert this stream into the server’s database, then when the client next calls SampleStream they will unpack the tensor into RAM, and when any method on that object is called (including its destructor) the attacker gains control of the Program Counter. We recommend upgrading past git commit  https://github.com/google-deepmind/reverb/commit/6a0dcf4c9e842b7f999912f792aaa6f6bd261a25

NVIDIA Transformers4Rec for Linux contains a vulnerability where an attacker could cause improper deserialization of untrusted data. A successful exploit of this vulnerability might lead to code execution, data tampering, and information disclosure.

NVIDIA Megatron Bridge for Linux contains a vulnerability where an 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.

NVIDIA Megatron Bridge for Linux contains a vulnerability where an 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.

NVIDIA Megatron Bridge for Linux contains a vulnerability where an 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.

NVIDIA Megatron Bridge for Linux contains a vulnerability where an 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.

NVIDIA Megatron Bridge for Linux contains a vulnerability where an 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.

Bio-Formats versions up to and including 8.3.0 perform unsafe Java deserialization of attacker-controlled memoization cache files (.bfmemo) during image processing. The loci.formats.Memoizer class automatically loads and deserializes memo files associated with images without validation, integrity checks, or trust enforcement. An attacker who can supply a crafted .bfmemo file alongside an image can trigger deserialization of untrusted data, which may result in denial of service, logic manipulation, or potentially remote code execution in environments where suitable gadget chains are present on the classpath.

A CWE-502: Deserialization of Untrusted Data vulnerability exists in the Dashboard module that could cause an interpretation of malicious payload data, potentially leading to remote code execution when an attacker gets the user to open a malicious file. Affected Products: IGSS Data Server(IGSSdataServer.exe)(V16.0.0.23040 and prior), IGSS Dashboard(DashBoard.exe)(V16.0.0.23040 and prior), Custom Reports(RMS16.dll)(V16.0.0.23040 and prior).

Manuskript through 0.12.0 allows remote attackers to execute arbitrary code via a crafted settings.pickle file in a project file, because there is insecure deserialization via the pickle.load() function in settings.py. NOTE: the vendor's position is that the product is not intended for opening an untrusted project file

An issue was discovered in Esoteric YamlBeans through 1.15. It allows untrusted deserialisation to Java classes by default, where the data and class are controlled by the author of the YAML document being processed.

A vulnerability in the HuggingFace Transformers library, specifically in the `Trainer` class, allows for arbitrary code execution. The `_load_rng_state()` method in `src/transformers/trainer.py` at line 3059 calls `torch.load()` without the `weights_only=True` parameter. This issue affects all versions of the library supporting `torch>=2.2` when used with PyTorch versions below 2.6, as the `safe_globals()` context manager provides no protection in these versions. An attacker can exploit this vulnerability by supplying a malicious checkpoint file, such as `rng_state.pth`, which can execute arbitrary code when loaded. The issue is resolved in version v5.0.0rc3.

A vulnerability in the `TFSMLayer` class of the `keras` package, version 3.13.0, allows attacker-controlled TensorFlow SavedModels to be loaded during deserialization of `.keras` models, even when `safe_mode=True`. This bypasses the security guarantees of `safe_mode` and enables arbitrary attacker-controlled code execution during model inference under the victim's privileges. The issue arises due to the unconditional loading of external SavedModels, serialization of attacker-controlled file paths, and the lack of validation in the `from_config()` method.

A safe mode bypass vulnerability in the `Model.load_model` method in Keras versions 3.0.0 through 3.10.0 allows an attacker to achieve arbitrary code execution by convincing a user to load a specially crafted `.keras` model archive.

There is a deserialization of untrusted data vulnerability in Digilent DASYLab. This vulnerability may result in arbitrary code execution. Successful exploitation requires an attacker to get a user to open a specially crafted DSB file. The vulnerability affects all versions of DASYLab.

Fuji Electric FRENIC-Loader 4 is vulnerable to a deserialization of untrusted data when importing a file through a specified window, which may allow an attacker to execute arbitrary code.

picklescan before 0.0.28 fails to detect malicious pickle files that invoke torch.utils._config_module.load_config function within reduce methods. Attackers can craft pickle files embedding arbitrary code that evades detection but executes during pickle.load, enabling remote code execution in supply chain attacks.

picklescan before 0.0.30 fails to detect malicious pickle files using idlelib.pyshell.ModifiedInterpreter.runcommand in reduce methods. Attackers can embed undetected code in pickle files that executes remote commands when loaded by victims.

picklescan before 0.0.30 fails to detect cProfile.runctx function calls in pickle file reduce methods, allowing attackers to execute arbitrary code. Malicious pickle files bypass picklescan detection and execute remote code when loaded via pickle.load().

Pdfminer.six is a community maintained fork of the original PDFMiner, a tool for extracting information from PDF documents. Prior to version 20251107, pdfminer.six will execute arbitrary code from a malicious pickle file if provided with a malicious PDF file. The `CMapDB._load_data()` function in pdfminer.six uses `pickle.loads()` to deserialize pickle files. These pickle files are supposed to be part of the pdfminer.six distribution stored in the `cmap/` directory, but a malicious PDF can specify an alternative directory and filename as long as the filename ends in `.pickle.gz`. A malicious, zipped pickle file can then contain code which will automatically execute when the PDF is processed. Version 20251107 fixes the issue.

Dataease is an open source data visualization analysis tool. In versions 2.10.14 and below, DataEase did not properly filter when establishing JDBC connections to Oracle, resulting in a risk of JNDI injection (Java Naming and Directory Interface injection). This issue is fixed in version 2.10.15.

A CWE-502 Deserialization of Untrusted Data vulnerability exists in SCADAPack 7x Remote Connect (V3.6.3.574 and prior) which could allow arbitrary code execution when an attacker builds a custom .PRJ file containing a malicious serialized buffer.

A vulnerability has been identified in the UA.Testclient utility, which is included in Rexroth IndraWorks. All versions prior to 15V24 are affected. This flaw allows an attacker to execute arbitrary code on the user's system by parsing a manipulated file containing malicious serialized data. Exploitation requires user interaction, specifically opening a specially crafted file, which then causes the application to deserialize the malicious data, enabling Remote Code Execution (RCE). This can lead to a complete compromise of the system running the UA.Testclient.

A vulnerability has been identified in Rexroth IndraWorks. This flaw allows an attacker to execute arbitrary code on the user's system by parsing a manipulated file containing malicious serialized data. Exploitation requires user interaction, specifically opening a specially crafted file, which then causes the application to deserialize the malicious data, enabling Remote Code Execution (RCE). This can lead to a complete compromise of the system running Rexroth IndraWorks.

A vulnerability has been identified in Rexroth IndraWorks. This flaw allows an attacker to execute arbitrary code on the user's system by parsing a manipulated file containing malicious serialized data. Exploitation requires user interaction, specifically opening a specially crafted file, which then causes the application to deserialize the malicious data, enabling Remote Code Execution (RCE). This can lead to a complete compromise of the system running Rexroth IndraWorks.

Visual Studio Code Remote Code Execution Vulnerability