The Keras.Model.load_model method, including when executed with the intended security mitigation safe_mode=True, is vulnerable to arbitrary local file loading and Server-Side Request Forgery (SSRF). This vulnerability stems from the way the StringLookup layer is handled during model loading from a specially crafted .keras archive. The constructor for the StringLookup layer accepts a vocabulary argument that can specify a local file path or a remote file path. * Arbitrary Local File Read: An attacker can create a malicious .keras file that embeds a local path in the StringLookup layer's configuration. When the model is loaded, Keras will attempt to read the content of the specified local file and incorporate it into the model state (e.g., retrievable via get_vocabulary()), allowing an attacker to read arbitrary local files on the hosting system. * Server-Side Request Forgery (SSRF): Keras utilizes tf.io.gfile for file operations. Since tf.io.gfile supports remote filesystem handlers (such as GCS and HDFS) and HTTP/HTTPS protocols, the same mechanism can be leveraged to fetch content from arbitrary network endpoints on the server's behalf, resulting in an SSRF condition. The security issue is that the feature allowing external path loading was not properly restricted by the safe_mode=True flag, which was intended to prevent such unintended data access.

nGrinder before 3.5.9 allows connection to malicious JMX/RMI server by default, which could be the cause of executing arbitrary code via RMI registry by remote attacker.

Jenkins DotCi Plugin 2.40.00 and earlier does not configure its YAML parser to prevent the instantiation of arbitrary types, resulting in a remote code execution vulnerability.

The Stop Spammers Security | Block Spam Users, Comments, Forms WordPress plugin before 2022.6 passes base64 encoded user input to the unserialize() PHP function when CAPTCHA are used as second challenge, which could lead to PHP Object injection if a plugin installed on the blog has a suitable gadget chain

IBM SDK, Java Technology Edition 7.1.5.18 and 8.0.8.0 could allow a remote attacker to execute arbitrary code on the system, caused by an unsafe deserialization flaw. By sending specially-crafted data, an attacker could exploit this vulnerability to execute arbitrary code on the system. IBM X-Force ID: 236069.

A remote code execution risk when restoring backup files originating from Moodle 1.9 was identified.

Dataease is an open source data visualization analysis tool. Dataease prior to 1.15.2 has a deserialization vulnerability. In Dataease, the Mysql data source in the data source function can customize the JDBC connection parameters and the Mysql server target to be connected. In `backend/src/main/java/io/dataease/provider/datasource/JdbcProvider.java`, the `MysqlConfiguration` class does not filter any parameters. If an attacker adds some parameters to a JDBC url and connects to a malicious mysql server, the attacker can trigger the mysql jdbc deserialization vulnerability. Through the deserialization vulnerability, the attacker can execute system commands and obtain server privileges. Version 1.15.2 contains a patch for this issue.

Deserialization of Untrusted Data vulnerability in QuanticaLabs MediCenter - Health Medical Clinic medicenter allows Object Injection.This issue affects MediCenter - Health Medical Clinic: from n/a through <= 15.1.

The Cooked Pro WordPress plugin before 1.7.5.7 does not properly validate or sanitize the recipe_args parameter before unserializing it in the cooked_loadmore action, allowing an unauthenticated attacker to trigger a PHP Object injection vulnerability.

Deserialization of untrusted data in on-premises Microsoft SharePoint Server allows an unauthorized attacker to execute code over a network. Microsoft is aware that an exploit for CVE-2025-53770 exists in the wild. Microsoft is preparing and fully testing a comprehensive update to address this vulnerability. In the meantime, please make sure that the mitigation provided in this CVE documentation is in place so that you are protected from exploitation.

MelisCms provides a full CMS for Melis Platform, including templating system, drag'n'drop of plugins, SEO and many administration tools. Attackers can deserialize arbitrary data on affected versions of `melisplatform/melis-cms`, and ultimately leads to the execution of arbitrary PHP code on the system. Conducting this attack does not require authentication. Users should immediately upgrade to `melisplatform/melis-cms` >= 5.0.1. This issue was addressed by restricting allowed classes when deserializing user-controlled data.

A deserialization vulnerability existed in dubbo hessian-lite 3.2.12 and its earlier versions, which could lead to malicious code execution. This issue affects Apache Dubbo 2.7.x version 2.7.17 and prior versions; Apache Dubbo 3.0.x version 3.0.11 and prior versions; Apache Dubbo 3.1.x version 3.1.0 and prior versions.

Deserialization of Untrusted Data vulnerability in VictorThemes Seil seil allows Object Injection.This issue affects Seil: from n/a through <= 1.7.1.

Delta Electronics InfraSuite Device Master versions 00.00.01a and prior deserialize user-supplied data provided through the Device-Gateway service port without proper verification. An attacker could provide malicious serialized objects to execute arbitrary code upon deserialization.

Deserialization of Untrusted Data vulnerability in ThemeMakers ThemeMakers Visual Content Composer tmm_content_composer allows Object Injection.This issue affects ThemeMakers Visual Content Composer: from n/a through <= 1.5.8.

Unauthenticated Java deserialization vulnerability in Serviceguard Manager

ThinkPHP v6.0.13 was discovered to contain a deserialization vulnerability via the component League\Flysystem\Cached\Storage\Psr6Cache. This vulnerability allows attackers to execute arbitrary code via a crafted payload.

This vulnerability allows remote attackers to execute arbitrary code on affected installations of Ivanti Avalanche 6.3.2.3490. Although authentication is required to exploit this vulnerability, the existing authentication mechanism can be bypassed. The specific flaw exists within the Certificate Management Server service. The issue results from the lack of proper validation of user-supplied data, which can result in deserialization of untrusted data. An attacker can leverage this vulnerability to execute code in the context of the service account. Was ZDI-CAN-15449.

Apache Geode versions up to 1.12.5, 1.13.4 and 1.14.0 are vulnerable to a deserialization of untrusted data flaw when using JMX over RMI on Java 8. Any user still on Java 8 who wishes to protect against deserialization attacks involving JMX or RMI should upgrade to Apache Geode 1.15 and Java 11. If upgrading to Java 11 is not possible, then upgrade to Apache Geode 1.15 and specify "--J=-Dgeode.enableGlobalSerialFilter=true" when starting any Locators or Servers. Follow the documentation for details on specifying any user classes that may be serialized/deserialized with the "serializable-object-filter" configuration option. Using a global serial filter will impact performance.

This vulnerability allows remote attackers to execute arbitrary code on affected installations of Ivanti Avalanche 6.3.2.3490. Although authentication is required to exploit this vulnerability, the existing authentication mechanism can be bypassed. The specific flaw exists within the Web File Server service. The issue results from the lack of proper validation of user-supplied data, which can result in deserialization of untrusted data. An attacker can leverage this vulnerability to execute code in the context of the service account. Was ZDI-CAN-15330.

This vulnerability allows remote attackers to execute arbitrary code on affected installations of Ivanti Avalanche 6.3.2.3490. Although authentication is required to exploit this vulnerability, the existing authentication mechanism can be bypassed. The specific flaw exists within the Notification Server service. The issue results from the lack of proper validation of user-supplied data, which can result in deserialization of untrusted data. An attacker can leverage this vulnerability to execute code in the context of the service account. Was ZDI-CAN-15448.

nGrinder before 3.5.9 allows to accept serialized Java objects from unauthenticated users, which could allow remote attacker to execute arbitrary code via unsafe Java objects deserialization.

EasyUse MailHunter Ultimate’s cookie deserialization function has an inadequate validation vulnerability. Deserializing a cookie containing malicious payload will trigger this insecure deserialization vulnerability, allowing an unauthenticated remote attacker to execute arbitrary code, manipulate system command or interrupt service.

An issue was discovered in RWS WorldServer before 11.7.3. /clientLogin deserializes Java objects without authentication, leading to command execution on the host.

It was discovered that a previous vulnerability was not completely fixed with SolarWinds Access Rights Manager. While some controls were implemented the researcher was able to bypass these and use a different method to exploit the vulnerability.

NVFLARE, versions prior to 2.1.2, contains a vulnerability in its PKI implementation module, where The CA credentials are transported via pickle and no safe deserialization. The deserialization of Untrusted Data may allow an unprivileged network attacker to cause Remote Code Execution, Denial Of Service, and Impact to both Confidentiality and Integrity.

NVFLARE, versions prior to 2.1.2, contains a vulnerability in its utils module, where YAML files are loaded via yaml.load() instead of yaml.safe_load(). The deserialization of Untrusted Data, may allow an unprivileged network attacker to cause Remote Code Execution, Denial Of Service, and Impact to both Confidentiality and Integrity.

Deserialization of Untrusted Data vulnerability in ThemeREX Classter classter allows Object Injection.This issue affects Classter: from n/a through <= 2.5.

Remote code execution vulnerabilities exist in the Netwrix Auditor User Activity Video Recording component affecting both the Netwrix Auditor server and agents installed on monitored systems. The remote code execution vulnerabilities exist within the underlying protocol used by the component, and potentially allow an unauthenticated remote attacker to execute arbitrary code as the NT AUTHORITY\SYSTEM user on affected systems, including on systems Netwrix Auditor monitors.

A vulnerability has been identified in Biograph Horizon PET/CT Systems (All VJ30 versions < VJ30C-UD01), MAGNETOM Family (NUMARIS X: VA12M, VA12S, VA10B, VA20A, VA30A, VA31A), MAMMOMAT Revelation (All VC20 versions < VC20D), NAEOTOM Alpha (All VA40 versions < VA40 SP2), SOMATOM X.cite (All versions < VA30 SP5 or VA40 SP2), SOMATOM X.creed (All versions < VA30 SP5 or VA40 SP2), SOMATOM go.All (All versions < VA30 SP5 or VA40 SP2), SOMATOM go.Now (All versions < VA30 SP5 or VA40 SP2), SOMATOM go.Open Pro (All versions < VA30 SP5 or VA40 SP2), SOMATOM go.Sim (All versions < VA30 SP5 or VA40 SP2), SOMATOM go.Top (All versions < VA30 SP5 or VA40 SP2), SOMATOM go.Up (All versions < VA30 SP5 or VA40 SP2), Symbia E/S (All VB22 versions < VB22A-UD03), Symbia Evo (All VB22 versions < VB22A-UD03), Symbia Intevo (All VB22 versions < VB22A-UD03), Symbia T (All VB22 versions < VB22A-UD03), Symbia.net (All VB22 versions < VB22A-UD03), syngo.via VB10 (All versions), syngo.via VB20 (All versions), syngo.via VB30 (All versions), syngo.via VB40 (All versions < VB40B HF06), syngo.via VB50 (All versions), syngo.via VB60 (All versions < VB60B HF02). The application deserialises untrusted data without sufficient validations that could result in an arbitrary deserialization. This could allow an unauthenticated attacker to execute code in the affected system if ports 32912/tcp or 32914/tcp are reachable.

Deserialization of Untrusted Data in Liferay Portal prior to 7.2.1 CE GA2 allows remote attackers to execute arbitrary code via JSON web services (JSONWS).

A Java Deserialization vulnerability in the Fishbowl Server in Fishbowl Inventory before 2022.4.1 allows remote attackers to execute arbitrary code via a crafted XML payload.

The Boost plugin for WordPress is vulnerable to PHP Object Injection in versions up to, and including, 2.0.3 via deserialization of untrusted input in the STYXKEY-BOOST_USER_LOCATION cookie. 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.

A vulnerability was found in laravel 5.1 and classified as problematic. This issue affects some unknown processing. The manipulation leads to deserialization. The attack may be initiated remotely. The exploit has been disclosed to the public and may be used. The identifier VDB-206501 was assigned to this vulnerability.

Deserialization of Untrusted Data vulnerability in Apache InLong.This issue affects Apache InLong: from 1.7.0 through 1.11.0,  the attackers can bypass using malicious parameters. Users are advised to upgrade to Apache InLong's 1.12.0 or cherry-pick [1], [2] to solve it. [1] https://github.com/apache/inlong/pull/9694 [2]  https://github.com/apache/inlong/pull/9707

SGLangs multimodal generation runtime scheduler's ROUTER socket binds to 0.0.0.0 by default and contains a sink that calls pickle.loads() on incoming messages, enabling RCE when exposed to the internet.

An issue was discovered in MISP before 2.4.158. PHAR deserialization can occur.

The Solr plugin of Apache OFBiz is configured by default to automatically make a RMI request on localhost, port 1099. In version 18.12.05 and earlier, by hosting a malicious RMI server on localhost, an attacker may exploit this behavior, at server start-up or on a server restart, in order to run arbitrary code. Upgrade to at least 18.12.06 or apply patches at https://issues.apache.org/jira/browse/OFBIZ-12646.

Phpok v6.1 was discovered to contain a deserialization vulnerability via the update_f() function in login_control.php. This vulnerability allows attackers to getshell via writing arbitrary files.

Deserialization of Untrusted Data vulnerability in scriptsbundle Exertio exertio allows Object Injection.This issue affects Exertio: from n/a through <= 1.3.2.

Deserialization of Untrusted Data vulnerability in the message processing component of Bitdefender GravityZone Console allows an attacker to pass unsafe commands to the environment. This issue affects: Bitdefender GravityZone Console On-Premise versions prior to 6.29.2-1. Bitdefender GravityZone Cloud Console versions prior to 6.27.2-2.

An insecure deserialization operation in Trend Micro Apex Central below version 8.0.7007 could lead to a pre-authentication remote code execution on affected installations. Note that this vulnerability is similar to CVE-2025-49219 but is in a different method.

In Progress® Telerik® UI for AJAX versions 2024.4.1114 through 2026.1.421, the RadFilter control is vulnerable to insecure deserialization when restoring filter state if the state is exposed to the client. If an attacker tampers with this state, a server-side remote code execution is possible.

All versions of bson before 1.1.4 are vulnerable to Deserialization of Untrusted Data. The package will ignore an unknown value for an object's _bsotype, leading to cases where an object is serialized as a document rather than the intended BSON type.

Phpok 6.1 has a deserialization vulnerability via framework/phpok_call.php.

A deserialization of untrusted data vulnernerability exists in rails < 5.2.4.3, rails < 6.0.3.1 that can allow an attacker to unmarshal user-provided objects in MemCacheStore and RedisCacheStore potentially resulting in an RCE.

Fluentd collects events from various data sources and writes them to files, RDBMS, NoSQL, IaaS, SaaS, Hadoop and so on. A remote code execution (RCE) vulnerability in non-default configurations of Fluentd allows unauthenticated attackers to execute arbitrary code via specially crafted JSON payloads. Fluentd setups are only affected if the environment variable `FLUENT_OJ_OPTION_MODE` is explicitly set to `object`. Please note: The option FLUENT_OJ_OPTION_MODE was introduced in Fluentd version 1.13.2. Earlier versions of Fluentd are not affected by this vulnerability. This issue was patched in version 1.15.3. As a workaround do not use `FLUENT_OJ_OPTION_MODE=object`.

OpenStack Object Storage (swift) before 1.7.0 uses the loads function in the pickle Python module unsafely when storing and loading metadata in memcached, which allows remote attackers to execute arbitrary code via a crafted pickle object.

MelisFront is the engine that displays website hosted on Melis Platform. It deals with showing pages, plugins, URL rewritting, search optimization and SEO, etc. Attackers can deserialize arbitrary data on affected versions of `melisplatform/melis-front`, and ultimately leads to the execution of arbitrary PHP code on the system. Conducting this attack does not require authentication. Users should immediately upgrade to `melisplatform/melis-front` >= 5.0.1. This issue was addressed by restricting allowed classes when deserializing user-controlled data.

Deserialization of Untrusted Data in the BVMS Mobile Video Service (BVMS MVS) allows an unauthenticated remote attacker to execute arbitrary code on the system. This affects Bosch BVMS versions 10.0 <= 10.0.0.1225, 9.0 <= 9.0.0.827, 8.0 <= 8.0.0.329 and 7.5 and older. This affects Bosch DIVAR IP 3000 and DIVAR IP 7000 if a vulnerable BVMS version is installed.