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.

The Milestone XProtect Video Management Software (Corporate, Expert, Professional+, Express+, Essential+) 2016 R1 (10.0.a) to 2018 R1 (12.1a) contains .NET Remoting endpoints that are vulnerable to deserialization attacks resulting in remote code execution.

Remote Cross-site Request forgery (CSRF) potential has been identified in UCMBD Server version DDM Content Pack V 10.20, 10.21, 10.22, 10.22 CUP7, 10.30, 10.31, 10.32, 10.33, 10.33 CUP2, 11.0 and CMS Server version 2018.05 BACKGROUND which could allow for remote unsafe deserialization and cross-site request forgery (CSRF).

Lexmark Markvision Enterprise before 2.3.0 misuses the Apache Commons Collections Library, leading to remote code execution because of Java deserialization.

ToolboxST prior to version 7.10 is affected by a deserialization vulnerability. An attacker with local access to an HMI or who has conducted a social engineering attack on an authorized operator could execute code in a Toolbox user's context through the deserialization of an untrusted configuration file. Two CVSS scores have been provided to capture the differences between the two aforementioned attack vectors.  Customers are advised to update to ToolboxST 7.10 which can be found in ControlST 7.10. If unable to update at this time customers should ensure they are following the guidance laid out in GE Gas Power's Secure Deployment Guide (GEH-6839). Customers should ensure they are not running ToolboxST as an Administrative user. 

Remote Cross-site Request forgery (CSRF) potential has been identified in UCMBD Browser version 4.10, 4.11, 4.12, 4.13, 4.14, 4.15, 4.15.1 which could allow for remote unsafe deserialization and cross-site request forgery (CSRF).

In Hazelcast before 3.11, the cluster join procedure is vulnerable to remote code execution via Java deserialization. If an attacker can reach a listening Hazelcast instance with a crafted JoinRequest, and vulnerable classes exist in the classpath, the attacker can run arbitrary code.

gui2/viewer/bookmarkmanager.py in Calibre 3.18 calls cPickle.load on imported bookmark data, which allows remote attackers to execute arbitrary code via a crafted .pickle file, as demonstrated by Python code that contains an os.system call.

Improper deserialization in WebGL in Google Chrome on Mac prior to 68.0.3440.75 allowed a remote attacker to potentially exploit heap corruption via a crafted HTML page.

FasterXML jackson-databind through 2.8.11 and 2.9.x through 2.9.3 allows unauthenticated remote code execution because of an incomplete fix for the CVE-2017-7525 and CVE-2017-17485 deserialization flaws. This is exploitable via two different gadgets that bypass a blacklist.

Spring Framework 3.0.0 through 3.0.5, Spring Security 3.0.0 through 3.0.5 and 2.0.0 through 2.0.6, and possibly other versions deserialize objects from untrusted sources, which allows remote attackers to bypass intended security restrictions and execute untrusted code by (1) serializing a java.lang.Proxy instance and using InvocationHandler, or (2) accessing internal AOP interfaces, as demonstrated using deserialization of a DefaultListableBeanFactory instance to execute arbitrary commands via the java.lang.Runtime class.

A CWE-502 Deserialization of Untrusted Data vulnerability exists in SCADAPack x70 Security Administrator (V1.2.0 and prior) which could allow arbitrary code execution when an attacker builds a custom .SDB file containing a malicious serialized buffer.

CWE-502: Deserialization of Untrusted Data vulnerability exists that could cause remote code execution when a malicious project file is loaded into the application by a valid user.

Deserialization of Untrusted Data vulnerability in Apache Camel SQL ComponentThis issue affects Apache Camel: from 3.0.0 before 3.21.4, from 3.22.0 before 3.22.1, from 4.0.0 before 4.0.4, from 4.1.0 before 4.4.0. Users are recommended to upgrade to version 4.4.0, which fixes the issue. If users are on the 4.0.x LTS releases stream, then they are suggested to upgrade to 4.0.4. If users are on 3.x, they are suggested to move to 3.21.4 or 3.22.1

A deserialization of untrusted data vulnerability exists in NI DAQExpress that may result in remote code execution. Successful exploitation requires an attacker to get a user to open a specially crafted project file. This vulnerability affects DAQExpress 5.1 and prior versions.  Please note that DAQExpress is an EOL product and will not receive any updates.

CWE-502: Deserialization of untrusted data vulnerability exists that could lead to loss of confidentiality, integrity and potential remote code execution on workstation when a non-admin authenticated user opens a malicious project file.

A deserialization of untrusted data vulnerability exists in NI G Web Development Software that may result in arbitrary code execution.  Successful exploitation requires an attacker to get a user to open a specially crafted project file.  This vulnerability affects G Web Development Software 2022 Q3 and prior versions.

Delta Electronics DTM Soft deserializes objects, which could allow an attacker to execute arbitrary code.

Delta Electronics DTM Soft Project File Parsing Deserialization of Untrusted Data Remote Code Execution

Delta Electronics DTN Soft Project File Parsing Deserialization of Untrusted Data Remote Code Execution

Apache jUDDI uses several classes related to Java's Remote Method Invocation (RMI) which (as an extension to UDDI) provides an alternate transport for accessing UDDI services. RMI uses the default Java serialization mechanism to pass parameters in RMI invocations. A remote attacker can send a malicious serialized object to the above RMI entries. The objects get deserialized without any check on the incoming data. In the worst case, it may let the attacker run arbitrary code remotely. For both jUDDI web service applications and jUDDI clients, the usage of RMI is disabled by default. Since this is an optional feature and an extension to the UDDI protocol, the likelihood of impact is low. Starting with 3.3.10, all RMI related code was removed.

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.

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.

PHPMailer before 5.2.27 and 6.x before 6.0.6 is vulnerable to an object injection attack.

Deserialization of untrusted data in Microsoft Office allows an unauthorized attacker to elevate privileges locally.

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

In Invoice Ninja before 4.4.0, there is an unsafe call to unserialize() in app/Ninja/Repositories/AccountRepository.php that may allow an attacker to deserialize arbitrary PHP classes. In certain contexts, this can result in remote code execution. The attacker's input must be hosted at http://www.geoplugin.net (cleartext HTTP), and thus a successful attack requires spoofing that site or obtaining control of it.

LLama Factory enables fine-tuning of large language models. Prior to version 1.0.0, a critical vulnerability exists in the `llamafy_baichuan2.py` script of the LLaMA-Factory project. The script performs insecure deserialization using `torch.load()` on user-supplied `.bin` files from an input directory. An attacker can exploit this behavior by crafting a malicious `.bin` file that executes arbitrary commands during deserialization. This issue has been patched in version 1.0.0.

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.

In Laravel Framework through 5.5.40 and 5.6.x through 5.6.29, remote code execution might occur as a result of an unserialize call on a potentially untrusted X-XSRF-TOKEN value. This involves the decrypt method in Illuminate/Encryption/Encrypter.php and PendingBroadcast in gadgetchains/Laravel/RCE/3/chain.php in phpggc. The attacker must know the application key, which normally would never occur, but could happen if the attacker previously had privileged access or successfully accomplished a previous attack.

mrdoc is vulnerable to Deserialization of Untrusted Data

ZStack is open source IaaS(infrastructure as a service) software. In ZStack before versions 3.10.12 and 4.1.6 there is a pre-auth unsafe deserialization vulnerability in the REST API. An attacker in control of the request body will be able to provide both the class name and the data to be deserialized and therefore will be able to instantiate an arbitrary type and assign arbitrary values to its fields. This issue may lead to a Denial Of Service. If a suitable gadget is available, then an attacker may also be able to exploit this vulnerability to gain pre-auth remote code execution. For additional details see the referenced GHSL-2021-087.

Deserialization of Untrusted Data vulnerability in yolo 5 allows attackers to execute arbitrary code via crafted yaml file.

In conference-scheduler-cli, a pickle.load call on imported data allows remote attackers to execute arbitrary code via a crafted .pickle file, as demonstrated by Python code that contains an os.system call.

VT-Designer Version 2.1.7.31 is vulnerable by the program populating objects with user supplied input via a file without first checking for validity, allowing attacker supplied input to be written to known memory locations. This may cause the program to crash or allow remote code execution.

Deserialization of Untrusted Data vulnerability in yolo 3 allows attackers to execute arbitrary code via crafted yaml file.

JetBrains dotPeek before 2018.2 and ReSharper Ultimate before 2018.1.4 allow attackers to execute code by decompiling a compiled .NET object (such as a DLL or EXE file) with a specific file, because of Deserialization of Untrusted Data.

M&M Software fdtCONTAINER Component in versions below 3.5.20304.x and between 3.6 and 3.6.20304.x is vulnerable to deserialization of untrusted data in its project storage.

Rockwell Automation Connected Components Workbench v12.00.00 and prior does not limit the objects that can be deserialized. This vulnerability allows attackers to craft a malicious serialized object that, if opened by a local user in Connected Components Workbench, may result in remote code execution. This vulnerability requires user interaction to be successfully exploited.

Microsoft Exchange Server Remote Code Execution Vulnerability

Deserialization of untrusted data in Microsoft Office SharePoint allows an unauthorized attacker to execute code locally.

Deserialization of untrusted data in Ivanti EPM before the 2024 January-2025 Security Update and 2022 SU6 January-2025 Security Update allows a remote unauthenticated attacker to achieve remote code execution. Local user interaction is required.

It was found that the fix for CVE-2016-9606 in versions 3.0.22 and 3.1.2 was incomplete and Yaml unmarshalling in Resteasy is still possible via `Yaml.load()` in YamlProvider.

There is a Hazelcast Library Java Deserialization Vulnerability in Citrix XenMobile Server 10.8 before RP2 and 10.7 before RP3.

An issue was discovered in EasyLogin Pro through 1.3.0. Encryptor.php contains an unserialize call that can be exploited for remote code execution in the decrypt function, if the attacker knows the key.

FasterXML jackson-databind 2.x before 2.9.10.4 mishandles the interaction between serialization gadgets and typing, related to org.apache.commons.jelly.impl.Embedded (aka commons-jelly).

PEAR Archive_Tar version 1.4.3 and earlier contains a CWE-502, CWE-915 vulnerability in the Archive_Tar class. There are several file operations with `$v_header['filename']` as parameter (such as file_exists, is_file, is_dir, etc). When extract is called without a specific prefix path, we can trigger unserialization by crafting a tar file with `phar://[path_to_malicious_phar_file]` as path. Object injection can be used to trigger destruct in the loaded PHP classes, e.g. the Archive_Tar class itself. With Archive_Tar object injection, arbitrary file deletion can occur because `@unlink($this->_temp_tarname)` is called. If another class with useful gadget is loaded, it may possible to cause remote code execution that can result in files being deleted or possibly modified. This vulnerability appears to have been fixed in 1.4.4.

NASA Singledop version v1.0 contains a CWE-502 vulnerability in NASA Singledop library (Weather data) that can result in remote code execution. This attack appear to be exploitable via Victim opening a specially crafted radar data file. This vulnerability appears to have been fixed in v1.1.

NASA Kodiak version v1.0 contains a CWE-502 vulnerability in Kodiak library's data processing function that can result in remote code execution. This attack appear to be exploitable via Victim opens an untrusted file for optimization using Kodiak library.

YamlDotNet version 4.3.2 and earlier contains a Insecure Direct Object Reference vulnerability in The default behavior of Deserializer.Deserialize() will deserialize user-controlled types in the line "currentType = Type.GetType(nodeEvent.Tag.Substring(1), throwOnError: false);" and blindly instantiates them. that can result in Code execution in the context of the running process. This attack appear to be exploitable via Victim must parse a specially-crafted YAML file. This vulnerability appears to have been fixed in 5.0.0.