Just published.
The newest CVEs, straight from CVE.org and NVD.
-
CVE-2026-115886
MLflow is an open source AI engineering platform for agents, large language models, and machine learning models. From 3.12.0rc0 until 3.17.0, the scorer invoke, register, and GenAI evaluate endpoints, including unauthenticated POST /ajax-api/3.0/mlflow/scorer/invoke, accept serialized third-party scorer data that Scorer.model_validate_json resolves using a module-path allowlist without constraining the selected class or its constructor arguments. When optional trulens and trulens-providers-litellm packages are installed, an attacker can select mlflow.genai.scorers.trulens, choose a model URI outside MLflow's native provider registry so execution uses the LiteLLM fallback, and supply api_base and api_key values that cause the server to send an outbound request to an attacker-chosen or internal host with attacker-controlled headers and body. The vulnerable deserialization occurs before a tracking-store lookup or job submission. This issue is fixed in version 3.17.0.
-
CVE-2026-91093
A potential security vulnerability has been identified in HP One Agent for certain HP PC products, which might allow for escalation of privilege and/or arbitrary code execution. HP is releasing software updates to mitigate the potential vulnerability.
-
CVE-2026-115845
An incomplete list of disallowed inputs in the Python code validator of the run_script tool in AWS MCP Server might allow an authenticated remote user to execute arbitrary code within their own isolated session environment via a crafted script. AWS fixed this issue service-side. No customer action is required.
-
CVE-2026-57845
Hugging Face Datasets before 5.2.0 contains an insecure deserialization vulnerability that allows attackers to execute arbitrary Python code by crafting a malicious pickle file loaded through the deprecated pandas packaged builder. Attackers can supply a pickle-backed dataset that is processed by Pandas._generate_tables via pandas.read_pickle() without validation, so when a victim calls datasets.load_dataset("pandas", data_files=...), the embedded __reduce__ payload executes code with the victim process's privileges.
-
CVE-2026-50056
A local privilege escalation vulnerability in the zmmailboxdmgr launcher of Zimbra Collaboration Suite (ZCS) allows the Unix zimbra account to execute arbitrary commands as root. The passwordless sudo interface accepts JVM arguments that are not adequately restricted. An attacker can postpone Jetty's privilege drop, register an OnOutOfMemoryError command and force heap exhaustion during startup, causing the command to execute while the JVM still runs as root. Exploitation requires local command execution as zimbra on a mailbox server and interrupts mailboxd during the malicious restart. ZCS 10.1.16 and 10.1.21 are affected.
-
CVE-2026-50057
A local privilege escalation vulnerability in the Zimbra Collaboration Suite (ZCS) MTA package allows the Unix zimbra account to execute arbitrary commands as root. The installed sudo policy permits passwordless root execution of Postfix administration commands without restricting their arguments. An attacker can select a zimbra-controlled daemon_directory in the Postfix configuration and invoke postfix set-permissions, causing Postfix to execute the attacker's post-install program as root. The Postfix configuration directory is also writable by zimbra, so restricting postconf alone does not remove the execution path. Exploitation requires local command execution as zimbra on an MTA node. ZCS 10.1.16 and 10.1.21 are affected.
-
CVE-2026-115481
baserCMS is a website development framework. Starting in version 5.0.0-beta1 and prior to versions 5.3.2 and 5.4.2, `PermissionGroupsService::build()` reads a plugin's `config/permission.php` to decide what each user group may reach. When that file is absent it writes a wildcard allow rule instead: `url` `/baser/<prefix>/<plugin>/*`, `method` `*`, `auth` true, for every non-admin group. BcCustomContent has no such file in any release. After an administrator installs it, the default `operators` group can store and execute arbitrary PHP under the document root. Versions 5.3.2 and 5.4.2 patch the issue.
-
CVE-2026-114624
baserCMS is a website development framework. In versions 5.0.0 before 5.3.2 and 5.4.0 before 5.4.2, the widget-area entity serializes widgets only when the submitted value is an array, allowing a string submitted through the admin JSON API to be stored unchanged. Entity reads, including front-end rendering and API responses, deserialize the stored value with unserialize() without an allowed_classes restriction, accepting both raw and base64-encoded serialized data. An administrator, or an attacker who obtains widget-write capability through a separate vulnerability such as chained XSS, can plant an object graph that subsequently triggers on unauthenticated page requests. Two demonstrated gadget chains in the default dependency set use composer/composer ProcessExecutor::startJob() with Process::fromShellCommandline(), or logiscape/mcp-sdk-php StdioTransport::connect() with StdioServerParameters, to reach proc_open() and execute commands as the web-server user. Deserialization bypasses constructor-based readonly safeguards, and the stored payload can execute repeatedly until the database row is repaired. Unauthenticated page access is a trigger after privileged payload storage, not an unauthenticated means of initially planting the payload. This issue has been patched in versions 5.3.2 and 5.4.2.
-
CVE-2026-114353
Hidden functionality (undocumented backdoor) in the root-privileged 'assistant' daemon in DJI CrystalSky ZS600C firmware v0.3.9.7 through v0.6.6.9 on Android allows a physically-present attacker with USB access to execute arbitrary OS commands as root by sending the hard-coded magic string 'dji26656677' to the USB-gadget serial interface (/dev/ttyGS0), after which each subsequent line is passed to system().
-
CVE-2026-115749
hydra-ray-launcher 1.1.0.dev1 before 1.2.2 and 1.4.0.dev4 before 1.4.0rc2 contains an arbitrary module import vulnerability via the ray_pkg_version OmegaConf resolver passing input to importlib.import_module(). Attackers controlling configuration values or command-line overrides can supply a ray_pkg_version interpolation to import any installed module, triggering its import-time side effects in the application process.
-
CVE-2026-112854
Deserialization of Untrusted Data vulnerability in axiomthemes Holistic Center holistic-center allows Object Injection.This issue affects Holistic Center: from n/a through 1.2.6.
-
CVE-2026-112855
Deserialization of Untrusted Data vulnerability in axiomthemes Redy redy allows Object Injection.This issue affects Redy: from n/a through 1.3.5.
-
CVE-2026-112856
Deserialization of Untrusted Data vulnerability in ThemeREX Group Vagabonds vagabonds allows Object Injection.This issue affects Vagabonds: from n/a through 1.3.10.
-
CVE-2026-112857
Deserialization of Untrusted Data vulnerability in axiomthemes Reprizo reprizo allows Object Injection.This issue affects Reprizo: from n/a through 1.0.8.
-
CVE-2026-112858
Deserialization of Untrusted Data vulnerability in axiomthemes Rhodos rhodos allows Object Injection.This issue affects Rhodos: from n/a through 1.3.3.
-
CVE-2026-112859
Deserialization of Untrusted Data vulnerability in ThemeREX Group Little Birdies little-birdies allows Object Injection.This issue affects Little Birdies: from n/a through 1.3.16.
-
CVE-2026-112846
Deserialization of Untrusted Data vulnerability in ThemeREX Group Save Life save-life allows Object Injection.This issue affects Save Life: from n/a through 1.2.14.
-
CVE-2026-112847
Deserialization of Untrusted Data vulnerability in ThemeREX ZenSoul zensoul allows Object Injection.This issue affects ZenSoul: from n/a through 1.13.
-
CVE-2026-112848
Deserialization of Untrusted Data vulnerability in AivahThemes Hostme v2 hostmev2 allows Object Injection.This issue affects Hostme v2: from n/a through 7.0.
-
CVE-2026-112851
Deserialization of Untrusted Data vulnerability in ThemeREX Group HotLock hotlock allows Object Injection.This issue affects HotLock: from n/a through 1.3.9.
-
CVE-2026-112852
Deserialization of Untrusted Data vulnerability in ThemeREX Group Helion helion allows Object Injection.This issue affects Helion: from n/a through 1.1.12.
-
CVE-2026-112853
Deserialization of Untrusted Data vulnerability in axiomthemes Hallelujah hallelujah allows Object Injection.This issue affects Hallelujah: from n/a through 1.1.1.
-
CVE-2026-115727
AgenticSeek before commit 409840c contains a remote code execution vulnerability because llm_server/app.py hardcodes Flask debug mode while binding to all interfaces. Attackers who reach the server port and obtain the derivable Werkzeug debugger PIN can execute Python code and operating system commands with llm_server privileges.
-
CVE-2026-112862
Deserialization of Untrusted Data vulnerability in xnau webdesign Participants Database participants-database allows Object Injection.This issue affects Participants Database: from n/a through 2.7.8.5.
-
CVE-2025-30054
CGM CLININET is vulnerable to local privilege escalation. The service that processes incoming HL7 messages runs with root privileges, but due to incorrect permissions, its scripts and configuration files can be modified by the uhcapache user. An attacker with uhcapache privileges can modify these files to execute arbitrary code as root. The vulnerability was confirmed to affect versions up to 2026.MS3 but later versions may also be vulnerable.
-
CVE-2025-30046
In the CGM CLININET application server, multiple processes responsible for HL7 communication with external systems run with root privileges while processing untrusted data received from those external systems. An attacker in adjacent network able to send data to these HL7 communication processes can achieve remote code execution with root privileges. The vulnerability was confirmed to affect versions up to 2026.MS3 but later versions may also be vulnerable.
-
CVE-2025-30047
CGM CLININET is vulnerable to Remote Code Execution caused by deserialization of untrusted data. The server exposes an HTTP-based API, used by the administrative module, that exchanges data using Java object serialization. An attacker within the same local network is able to send requests to this API and transmit a crafted serialized object. When it is deserialized, it allows arbitrary code execution with root privileges on the server. This issue was fixed in version 2026.MS3
-
CVE-2026-113472
The Depicter — Popup & Slider Builder WordPress plugin before 4.8.2 does not properly validate uploaded files in its editor media-upload endpoint, checking only the client-supplied MIME type and a case-sensitive block of the .php extension, which allows users with the Editor role and above to upload files with other active extensions into the public uploads directory, leading to stored script execution in the site origin and, depending on the server's handler mapping, to code execution.
-
CVE-2026-54630
Part-DB is an open source inventory management system for electronic components. Prior to version 2.12.1, the part attachment upload handler stores uploaded files under the web-served media directory using the client-supplied file extension, and its extension blocklist does not include phar. AttachmentSubmitHandler::renameBlacklistedExtensions() rewrites dangerous extensions (php, phtml, sh, cgi, py, js, html, htaccess, ...) to .txt, but phar is absent from the list, so a .phar file keeps its extension. Public attachments (the default) are written into public/media/, which is the Apache document root. The shipped Apache configuration routes .phar to PHP-FPM, so the stored file is executed as PHP on request. The per-attachment-type extension filter is empty by default and is not consulted by the upload handler in a way that blocks this, and a non-admin user in the default users group can create an attachment type with no filter and attach files to parts. Any authenticated user in the default users group uploads a .phar attachment to a part and fetches it under /media/, executing arbitrary PHP as the www-data web user. Ability to exploit depends on the webserver config, docker image contained the vulnerability by default. This issue has been patched in version 2.12.1.
-
CVE-2026-79850
Vulnerabilities in the affected interface of AOS-8 could allow an unauthenticated remote attacker to bypass existing authentication controls and run arbitrary commands on the underlying host. Successful exploitation could allow an attacker to execute arbitrary commands as root on the underlying operating system leading to complete system compromise.
-
CVE-2026-114608
Pylint is a static code analyser for Python. From 0.15.2 until 4.0.9, Pylint stores module statistics in predictable files under PYLINT_HOME and loads those files with pickle.load on a later run. An attacker who can write to a shared, group-writable, or otherwise untrusted cache directory can replace a stats cache with crafted serialized data, causing arbitrary code to execute with the privileges of the user who next runs Pylint. The default per-user cache is not normally writable by another user. This issue is fixed in version 4.0.9.
-
CVE-2026-51954
An issue in Webasyst Framework 4.0.3 and before allows a remote attacker to execute arbitrary code via a crafted OAuth authorization URL with an arbitrary redirect_uri.
-
CVE-2026-16477
CleanShot X for macOS insecurely deserializes untrusted data embedded in a .cleanshot project file (editable screenshot annotations). Exploitation requires an attacker to craft a malicious .cleanshot file, deliver it to the target, and have the target open it; the payload then executes automatically during parsing, resulting in arbitrary code execution in the context of the current user. Users who have never opened a .cleanshot file from an untrusted source are not affected by this vulnerability. This vulnerability was fixed in versions: 4.5.2, 4.6.3, 4.7.8 and 4.8.10
-
CVE-2026-102785
Joomla Extension - yootheme.com - Unauthenticated remote code execution in YOOtheme Pro 5.0.0-5.0.52 - YOOtheme Pro’s publicly accessible image transformation handler incompletely binds normalized image operations and the output destination to its request signature. An unauthenticated remote attacker can reuse a signature disclosed in a legitimate public image URL, inject additional image operations, traverse out of the intended cache directory, and write an attacker-controlled PNG/PHP polyglot to an executable public media directory.
-
CVE-2026-114724
Roundcube Webmail before 1.6.20 and 1.7.x before 1.7.5 contains an object injection primitive in rcube_session::unserialize(). A session key prefixed with “!” was treated as having no value, and the reconstructed serialized string was emitted with a key but no corresponding value, desynchronizing the remainder of the input so that following session data was interpreted as a serialized value. This yields a PHP object injection primitive for an attacker able to influence stored session data. Upstream fixed this in the commit “Fix object injection primitive in rcube_session::unserialize() via “!”-prefixed session keys”, which rejects such keys.
-
CVE-2026-29001
TastyIgniter through 4.4.5 contains a PHP object injection vulnerability that allows authenticated administrators to achieve arbitrary code execution by submitting a serialized PHP object as a settings value, which is stored unsanitized and later unserialized by getValueAttribute() when the setting is read. Attackers can submit a serialized gadget chain, such as one generated via phpggc, through the admin settings form (for example the maintenance_message setting loaded by CheckMaintenance.php) to achieve arbitrary operating-system command execution as the PHP application process user, with the impact persisting until the value is overwritten.
-
CVE-2026-114145
GitPython is a python library used to interact with Git repositories. Prior to 3.2.1, create_submodule() in git/objects/submodule/base.py accepts an attacker-influenced submodule path such as .git/hooks and invokes _clone_repo() before safely rejecting the destination. In 3.1.x, _validate_repo_path() does not exist and the call succeeds silently, while in 3.2.0 and later affected revisions the check in git/index/base.py raises ValueError only after the clone has written to disk and does not remove the planted file. An attacker-controlled submodule can therefore place an executable post-checkout hook that a later Git operation executes. Exploitation requires an application to accept an untrusted submodule path and a clone destination such as .git/hooks that is empty enough for Git to clone into it, which commonly means that default sample hooks are absent. This issue is fixed in version 3.2.1.
-
CVE-2026-113870
MongoDB Compass improperly escapes certain BSON values when converting a stored view’s aggregation pipeline into editable source. When a user connects to a MongoDB deployment containing a specially crafted view and opens that view for editing, the values can cause arbitrary code to execute with the privileges of the user running Compass.
-
CVE-2026-113461
Local privilege escalation in Checkmk 2.5.0p14 allows a local user with write access to the Oracle client library directory of an Oracle home listed in the 'oratab' file or to the 'oratab' file itself to execute code as root if an agent has the 'mk-oracle' plugin enabled.
-
CVE-2026-92206
CrewAI crewAI Framework Agent Loading Unsafe Reflection Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of CrewAI crewAI. User interaction is required to exploit this vulnerability in that the target must load a malicious agent configuration from the repository. The specific flaw exists within the load_agent_from_repository function. The process does not properly restrict a user-supplied argument before using it to import a module. An attacker can leverage this vulnerability to execute code in the context of the service account. Was ZDI-CAN-28387.
-
CVE-2026-92202
Luxion KeyShot BIP File Parsing Uncontrolled Search Path Element Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Luxion KeyShot. User interaction is required to exploit this vulnerability in that the target must visit a malicious page or open a malicious file. The specific flaw exists within the parsing of BIP files. The issue results from loading a library from an unsecured location. An attacker can leverage this vulnerability to execute code in the context of the current process. Was ZDI-CAN-29268.
-
CVE-2026-111184
NoMachine Named Pipe Incorrect Permission Assignment Local Privilege Escalation Vulnerability. This vulnerability allows local attackers to escalate privileges on affected installations of NoMachine. An attacker must first obtain the ability to execute low-privileged code on the target system in order to exploit this vulnerability. The specific flaw exists within the NoMachine Device Service. The issue results from incorrect permissions on a named pipe. An attacker can leverage this vulnerability to escalate privileges and execute arbitrary code in the context of SYSTEM. Was ZDI-CAN-31213.
-
CVE-2026-113779
The pfSense REST API package is an unofficial, open-source REST and GraphQL API for pfSense CE and pfSense Plus firewalls. A privilege escalation vulnerability in pfSense-pkg-RESTAPI prior to version 2.9.0 allows authenticated accounts holding the dedicated settings sync permission to modify REST API configuration secrets. This enables a non-administrative user to elevate their privileges to a full administrator. pfSense-pkg-RESTAPI 2.9.0 patches the issue. Access to the HA settings sync endpoint now strictly requires full administrative privileges (`page-all`). The sub-administrative sync privilege has been removed in versions v2.9.0 and later. In addition to correcting the access control boundary, version 2.9.0 includes the following hardening measures: opt-in default and format and validation modernization. As a workaround, one may mitigate exposure by taking either or all of the following steps. Revoke Scoped Permissions: Remove the `api-v2-system-restapi-settings-sync-post` privilege from all non-administrative users. Apply ACLs: Configure REST API Access Control Lists (ACLs) and/or firewall rules to restrict access to API endpoints exclusively to trusted administrator and HA peer IP addresses.
-
CVE-2026-111832
Unrestricted Upload of File with Dangerous Type vulnerability in WP Mailster WP Mailster wp-mailster allows Upload a Web Shell to a Web Server.This issue affects WP Mailster: from n/a through 1.9.1.0.
-
CVE-2026-111834
Unrestricted Upload of File with Dangerous Type vulnerability in Kitae Park Mang Board WP mangboard allows Upload a Web Shell to a Web Server.This issue affects Mang Board WP: from n/a through 2.4.4.
-
CVE-2026-94300
Replaying GSS-TSIG signed UDP messages can result in memory corruption in "`named`". An attacker who can capture a previously signed UDP message can replay it while the negotiated TSIG key is still valid. By also sending crafted queries to a zone the attacker controls, the attacker may potentially achieve remote code execution. This issue affects BIND 9 versions 9.11.0 through 9.18.50, 9.20.0 through 9.20.29, 9.21.0 through 9.21.26, 9.11.3-S1 through 9.18.50-S1, and 9.20.9-S1 through 9.20.29-S1.
-
CVE-2026-114403
PHPOffice PHPPresentation through 1.2.0 contains an insecure deserialization vulnerability that allows attackers to instantiate arbitrary PHP objects by supplying a crafted presentation file. Attackers can upload a ZIP containing PhpPresentation.xml that IOFactory::load passes to unserialize(), enabling gadget-chain code execution, file writes, or denial of service.
-
CVE-2026-114398
Trudesk through 1.2.11 contains an unrestricted file upload vulnerability that allows any authenticated user to place HTML or SVG files in the public webroot via the uploadlogo, uploadpagelogo and uploadfavicon endpoints. Attackers can spoof an image MIME type to upload script-bearing files served unauthenticated from /assets/, enabling stored cross-site scripting and replacement of instance branding.
-
CVE-2026-110369
Deserialization of Untrusted Data vulnerability in GS Plugins WordPress Team Members – GS Plugins gs-team-members allows Object Injection.This issue affects WordPress Team Members – GS Plugins: from n/a through 2.8.2.
-
CVE-2026-110370
Deserialization of Untrusted Data vulnerability in GS Plugins GS Pins for Pinterest gs-pinterest-portfolio allows Object Injection.This issue affects GS Pins for Pinterest: from n/a through 1.9.2.
The newest published CVE records from CVE.org and NVD, refreshed every few minutes, with severity, exploit odds and fix status.