CVE-20XX-?
agent?

The internet is broken right now.

65 critical CVEs without a fix in the last 48 hours. Developers are crying quietly into their IDEs. 149 more already have a fix that someone still has to apply.

1,101 CVEs · updated 2026-11-19 05:57 UTC

Reset all
  1. CVE-2026-65949

    Improper neutralization of special elements used in an OS command in Lenovo XClarity Integrator for Microsoft Windows Admin Center 5.1.1 and 5.2.0 allows an authenticated remote attacker to execute arbitrary PowerShell commands by submitting a crafted hostname.

    Injection OS & kernel

    8.6 HIGH 0.6% exploit odds No fix listed
  2. CVE-2026-102815

    An operating-system command injection vulnerability exists in the tmpServer service of TP-Link RE305 V3. The service runs with root privileges and listens on the local loopback interface. A local attacker with access to a low-privileged process on the device may submit a crafted request to the TDDPv1 service and execute operating-system commands with root privileges. Successful exploitation may allow a local attacker with low privileges to obtain root privileges and execute arbitrary operating-system commands on the affected device.

    Injection Network gear

    6.9 MEDIUM 0.6% exploit odds No fix listed
  3. CVE-2026-115878

    Stanza is a Stanford NLP Python library for tokenization, sentence segmentation, NER, and parsing of many human languages. From 1.14.0 until 1.15.0, loading an attacker-supplied lemmatizer checkpoint can execute arbitrary code because stanza/models/lemma/trainer.py stores pos_dict as gzip-compressed pickle bytes and Trainer.load passes the decompressed stream to unrestricted pickle.loads. The outer torch.load call uses weights_only=True, but the nested bytes are treated as a primitive and are not inspected by that allowlist. The path is reachable during pipeline construction through a caller-influenced lemma_model_path, before inference begins. This issue is fixed in version 1.15.0.

    Injection Library

    7.8 HIGH 0.6% exploit odds Fix listed
  4. CVE-2026-65874

    This issue affects Apache Fineract versions up to and including 1.15.0. The Center read endpoint (GET /api/v1/centers) and the datatable entry read endpoints (GET /api/v1/datatables/{datatable}/{apptableId} and GET /api/v1/datatables/{datatable}/{apptableId}/{datatableId}) concatenate caller-supplied ordering parameters (orderBy, sortOrder, order) directly into the SQL ORDER BY clause. The existing regex-based input validation can be bypassed, allowing time-based blind SQL injection by any authenticated user with read access to centers or to the targeted datatable. Users are recommended to upgrade to 1.16.1 which fixes the issue.

    Injection Web app Apache Fineract

    8.8 HIGH 0.6% exploit odds Fix listed
  5. CVE-2026-117568

    mcp-ffmpeg through 1.0.5 contains an OS command injection vulnerability that allows attackers to run shell commands by supplying crafted videoPath or outputDir values to the resize-video, extract-audio and get-video-info tools. Attackers can embed $(), backticks or a closing double quote in an existing file or directory name to break out of the child_process.exec command string and execute commands as the MCP server user.

    Injection Server

    7.3 HIGH 0.6% exploit odds No fix listed
  6. CVE-2026-116852

    Author Remote Code Execution (RCE) in Widget Options <= 4.2.5 versions.

    Injection Web app

    9.1 CRITICAL 0.6% exploit odds No fix listed
  7. CVE-2026-113673

    The GeoDirectory WordPress plugin before 2.8.194 does not properly escape administrator-defined text before writing it into a generated PHP file in its own folder, allowing high privilege users such as a subsite administrator on a multisite network, who cannot otherwise run code, to achieve Remote Code Execution that any unauthenticated visitor can trigger.

    Injection Web app

    7.2 HIGH 0.6% exploit odds No fix listed
  8. CVE-2026-115189

    A server-side template injection vulnerability exists in the Twig theme rendering sandbox of Mautic. The sandbox introduced to restrict user-uploaded theme templates (the fix for GHSA-9fx4-7cmj-47vg) does not fully enforce its restrictions, allowing a crafted theme template to escape the sandbox. An authenticated user with the core:themes:create permission can upload a malicious theme whose template is later rendered, resulting in execution of arbitrary code on the server with the privileges of the web server process. This allows the attacker to execute operating system commands, read server-side files including configuration files and database credentials, write files, and reach internal services. In default installations the core:themes:create permission is granted to the Administrator role, and in multi-operator setups it may be delegated to other roles. This is an incomplete fix for GHSA-9fx4-7cmj-47vg.

    Injection Web app

    8.8 HIGH 0.6% exploit odds Fix listed
  9. CVE-2026-79887

    A vulnerability in the web interface of AOS-8 could allow an unauthenticated remote attacker to achieve remote code execution. Successful exploitation could lead to complete device control, enabling interception of network traffic and potential compromise of connected systems.

    Injection Web app

    8.1 HIGH 0.6% exploit odds No fix listed
  10. CVE-2026-119989

    Improper neutralization of special elements used in a command ('command injection') in Azure Pipelines allows an authorized attacker to execute code over a network.

    Injection Microsoft Azure Web Apps

    8.2 HIGH 0.6% exploit odds Fix listed
  11. CVE-2026-122988

    Improper authorization in Windows Admin Center allows an authorized attacker to elevate privileges over a network.

    Injection OS & kernel Microsoft Windows Admin Center

    8.0 HIGH 0.6% exploit odds Fix listed
  12. CVE-2026-124430

    Craft CMS from 5.10.0 before 5.11.2 contains a remote code execution vulnerability that allows non-administrator control panel users to obtain HMAC-signed Twig templates via the app/render-elements endpoint. Attackers can submit the signed templated returnUrl as the login redirect parameter, triggering unsandboxed renderObjectTemplate() evaluation and executing OS commands as the PHP worker.

    Injection Web app

    8.7 HIGH 0.6% exploit odds Fix listed
  13. CVE-2026-118692

    CherryTree through 1.7.2 contains an OS command injection vulnerability in CtActions::link_clicked that allows attackers to execute commands by embedding shell metacharacters in web link URLs. When the non-default Custom Web Link Clicked Action option is enabled, attackers can supply crafted documents whose links pass $(...), backticks, or semicolons unquoted to std::system, running commands as the victim on click.

    Injection Web app

    7.1 HIGH 0.6% exploit odds No fix listed
  14. CVE-2026-57117

    PraisonAI is a multi-agent teams system. From version 4.6.10 to before version 4.6.59, LocalManagedAgent / SandboxedAgent compute bridging wraps read_file, list_files, and write_file when a compute provider is attached. The bridge converts those file operations into shell command strings using raw path arguments, then sends those strings to shell-backed compute providers. An attacker who can influence a file-tool path argument can break out of the quoted path and execute arbitrary shell commands in the compute environment. With compute="local", commands execute through the local subprocess compute provider on the host. With Docker, commands execute in the container. This issue has been patched in version 4.6.59.

    Injection Server

    8.8 HIGH 0.6% exploit odds Fix listed
  15. CVE-2026-122979

    Improper control of generation of code ('code injection') in Microsoft Teams for Android allows an unauthorized attacker to perform spoofing over a network.

    Injection OS & kernel Microsoft Teams For Android

    8.1 HIGH 0.6% exploit odds Fix listed
  16. CVE-2026-96376

    Improper Neutralization of Special Elements used in an OS Command ("OS Command Injection") vulnerability in Drupal Cloud allows OS Command Injection. This issue affects Cloud versions: from 0.0.0 to 7.0.1.

    Injection Web app

    6.4 MEDIUM 0.6% exploit odds No fix listed
  17. CVE-2026-110309

    IBM Langflow OSS 1.0.0 through 1.12.4 could allow a remote authenticated attacker to execute arbitrary commands due to improper control of code generation.

    Injection Web app Langflow

    8.8 HIGH 0.6% exploit odds Fix listed
  18. CVE-2026-58256

    GRR is a resource management and reservation application. Prior to the advisory-reported fixed version 4.5.0, an administrator can inject malicious PHP through LDAP connector configuration parameters such as `ldap_filter`, poisoning `/personnalisation/config_ldap.inc.php` and permitting remote code execution. The administrator must first successfully connect to an LDAP server, which can be attacker-controlled, and then submit malicious input on the subsequent connector settings page. The advisory reports this issue as fixed in version 4.5.0.

    Injection Web app

    9.1 CRITICAL 0.6% exploit odds Fix listed
  19. CVE-2026-104351

    Adobe Campaign Classic (ACC) is affected by an Improper Control of Generation of Code ('Code Injection') vulnerability that could result in arbitrary code execution in the context of the current user. A low-privileged attacker could exploit this vulnerability to execute arbitrary code. Exploitation of this issue does not require user interaction. Scope is changed.

    Injection Web app

    9.9 CRITICAL 0.6% exploit odds No fix listed
  20. CVE-2026-121265

    SQL::Abstract versions before 2.000003 for Perl allow SQL injection via an arrayref column value in insert and update. With the array_datatypes option off, which is the default, insert(), update() and values() treat an arrayref column value as literal SQL. Its first element goes into the SQL text unescaped and the rest become bind values. A JSON array, or a repeated form field decoded as an arrayref, is such a value. An application that passes request data as column values to insert or update with the default options allows SQL injection. Releases before 2.000003 also treat an arrayref in generate() as literal SQL, including through AUTOLOAD when allow_autoload is on, whatever array_datatypes is set to.

    Injection

    9.8 CRITICAL 0.6% exploit odds No fix listed
  21. CVE-2026-108097

    Use of Externally-Controlled Input to Select Classes or Code ('Unsafe Reflection') vulnerability in Drupal Drupal core allows Reflection Injection. This issue affects Drupal core versions: from 8.9.0 to 10.6.19, from 11.0.0 to 11.3.19, from 11.4.0 to 11.4.10, from 12.0.0-alpha1 to 12.0.0-beta2.

    Injection Web app

    8.1 HIGH 0.6% exploit odds No fix listed
  22. CVE-2026-110273

    IBM TRIRIGA Application Platform 5.0.0 through 5.0.9 IBM TRIRIGA could allow a remote authenticated attacker to execute arbitrary code due to improper neutralization of special elements used in an OS command.

    Injection Web app

    8.8 HIGH 0.6% exploit odds No fix listed
  23. CVE-2026-88711

    Improper neutralization of special elements used in an OS command in certain ASUS Router firmware models allows an remote attacker to to execute arbitrary commands with root privileges by inducing a logged-in administrator to visit a malicious website. Refer to the ' Security Update for ASUS Router Firmware ' section on the ASUS Security Advisory for more information.

    Injection Network gear ASUS Router Firmware

    8.7 HIGH 0.6% exploit odds No fix listed
  24. CVE-2026-125120

    Mongoose is a MongoDB object modeling tool designed to work in an asynchronous environment. Prior to 6.13.11, 7.8.12, 8.24.3, and 9.9.2, Mongoose countDocuments() does not apply the configured sanitizeFilter query selector injection defense. An application that enables sanitizeFilter and passes attacker-controlled filter values to countDocuments() can therefore process query operator objects as operators rather than literal values, causing the count to include documents that should not match and potentially undermining authorization or application logic based on that count. This issue is fixed in versions 6.13.11, 7.8.12, 8.24.3, and 9.9.2.

    Injection

    9.8 CRITICAL 0.6% exploit odds Fix listed
  25. CVE-2026-126180

    A improper neutralization of special elements used in a template engine vulnerability in Fortinet FortiManager 8.0.0 through 8.0.1, FortiManager 7.6.0 through 7.6.7, FortiManager 7.4.0 through 7.4.11, FortiManager 7.2 all versions, FortiManager Cloud 8.0.0 through 8.0.1, FortiManager Cloud 7.6.0 through 7.6.7, FortiManager Cloud 7.4.1 through 7.4.11, FortiManager Cloud 7.2 all versions may allow a low privileged Device Manager admin to obtain super admin privilege on managed FortiGates via injecting malicious commands.

    Injection Network gear

    8.4 HIGH 0.6% exploit odds No fix listed
  26. CVE-2026-81584

    An improper neutralization of special elements used in an SQL command vulnerability in the User API in Synology DiskStation Manager (DSM) before 7.3.2-86009-5 and 7.4.1-90080-1 allows remote attackers with control over a compromised LDAP or Active Directory server to read or write arbitrary files and conduct denial-of-service attacks.

    Injection Server

    8.1 HIGH 0.6% exploit odds No fix listed
  27. CVE-2026-129273

    N-able N-central before 2026.4.1 HF3 does not neutralise control characters in an Audit Export setting before writing it to the system logging configuration. An authenticated attacker with product administrator privileges can inject configuration and execute commands as root on the N-central server.

    Injection Server N-able N-central

    8.6 HIGH 0.6% exploit odds No fix listed
  28. CVE-2026-54140

    OpenHuman is an open-source agent harness available as a desktop application and Rust library. Before 0.57.13, OpenHuman places names, descriptions, titles, and parameter definitions returned by configured external MCP servers into the agent model context without neutralizing instructions or bounding the descriptive fields. The McpRemoteTool records produced by list_tools preserve this remote content, while tool filtering considers the tool name and the prompt-injection detector screens user messages rather than MCP definitions. A malicious or compromised MCP server can therefore return a poisoned description that is evaluated on every agent turn and may direct the model to invoke other available tools, disclose accessible data, or perform unauthorized actions without a separate malicious user message. The issue affects users who configure an attacker-controlled or compromised MCP server. This issue is fixed in version 0.57.13.

    Injection Library

    8.3 HIGH 0.6% exploit odds Fix listed
  29. CVE-2026-119085

    Improper neutralization of special elements used in a template engine in the publishing component of Amazon Virtual Engineering Workbench (VEW) could allow a remote authenticated third party to run arbitrary code with the permissions of the publishing Lambda's IAM role. This could allow access to data and resources across all tenant projects. The code is reachable through a crafted CloudFormation template body containing template engine directives submitted through the product versioning API. To remediate this issue, users should redeploy from the latest version of awslabs/virtual-engineering-workbench.

    Injection Server

    9.4 CRITICAL 0.6% exploit odds Fix listed
  30. CVE-2026-65964

    Pi-hole FTL is the core engine of the Pi-hole network-level advertisement and tracker blocker. Prior to 6.7, log_http_access in src/webserver/webserver.c writes attacker-controlled User-Agent content to the access log without output neutralization when debug.api is enabled. If an administrator configures files.log.webserver inside webserver.paths.webroot with a filename matching lua_server_page_pattern, an unauthenticated request can plant Lua markup in that log and a later request can cause CivetWeb to evaluate the log as a Lua server page. The resulting commands run as the pihole operating-system user and can access or modify pihole-owned configuration, credentials, databases, logs, and rendered interface content. When the log is not served through the Lua handler, the same unneutralized writer still permits log forgery and downstream log-parser injection. This issue is fixed in version 6.7.

    Injection Library

    6.6 MEDIUM 0.6% exploit odds Fix listed
  31. CVE-2026-120153

    SQL injection in PostgreSQL extension-script substitutions allows an attacker with database-level CREATE privilege to execute arbitrary code as the bootstrap superuser, via a schema or owner name containing newlines or other special substrings. This is similar to CVE-2023-39417, but it requires a non-bundled extension with a vulnerable script pattern. Versions before PostgreSQL 18.7, 17.12, 16.16, 15.20, and 14.25 are affected.

    Injection Server

    7.2 HIGH 0.6% exploit odds No fix listed
  32. CVE-2026-59106

    A security vulnerability has been identified in the WolfProt interface implementation. Insufficient neutralization of user-controlled input could allow an unauthenticated attacker with network access to a vulnerable service to execute operating-system commands with the privileges of the affected application service.

    Injection

    9.3 CRITICAL 0.5% exploit odds No fix listed
  33. CVE-2026-15723

    GitLab has remediated an issue in GitLab CE/EE affecting all versions from 10.1 before 19.3.6, 19.4 before 19.4.3, and 19.5 before 19.5.1 that, when OmniAuth SAML and LDAP auto-linking were both enabled, an authenticated user could have taken over an existing LDAP-backed account without any interaction from the targeted user, due to improper input sanitization in the LDAP email lookup path used during OmniAuth identity linking.

    Injection Web app GitLab

    8.2 HIGH 0.5% exploit odds Fix listed
  34. CVE-2026-125119

    Mongoose is a MongoDB object modeling tool designed to work in an asynchronous environment. From 6.0.0 until 6.13.12, 7.8.13, 8.24.7, and 9.11.4, Mongoose stores the connection-level sanitizeFilter option in Connection.config, but Query.prototype._castConditions() and query casting check other option locations and do not read that configuration. An application that enables sanitizeFilter through mongoose.connect() or createConnection() and passes an attacker-controlled object into a query filter therefore receives a connection that reports the defense as enabled while query selector objects remain unsanitized. A crafted password value using a query operator can match an account record without the expected literal password, while global configuration through mongoose.set('sanitizeFilter') and per-query Model.find().sanitizeFilter() remain unaffected. This issue is fixed in versions 6.13.12, 7.8.13, 8.24.7, and 9.11.4.

    Injection

    7.5 HIGH 0.5% exploit odds Fix listed
  35. CVE-2026-65213

    This issue may allow an IPMI authenticated administrator to execute commands as the operating system root user on the BMC

    Injection

    6.6 MEDIUM 0.5% exploit odds No fix listed
  36. CVE-2026-126162

    A xml injection (aka blind xpath injection) vulnerability in Fortinet FortiMail 8.0.0 through 8.0.1, FortiMail 7.6.0 through 7.6.6 may allow a remote unauthenticated attacker to execute unauthorized code or commands via a specially crafted request.

    Injection Server

    8.7 HIGH 0.5% exploit odds No fix listed
  37. CVE-2026-61564

    AVideo is a video-sharing Platform. Prior to 8.1.0, the unauthenticated installer endpoint `install/checkConfiguration.php` assembles `videos/configuration.php` by interpolating request parameters, including `webSiteRootURL`, into single-quoted PHP string literals without escaping. An attacker can terminate a literal and inject PHP that executes as the web-server user when the generated, directly requestable file is loaded. Exploitation requires `videos/configuration.php` to be absent, as on every fresh deployment or after the file is removed or moved; an existing file blocks this endpoint. The supplied `systemRootPath` must contain index.php and a connection using `databaseHost`, `databaseUser`, and `databasePass` must succeed, but the attacker can point those parameters to their own database server and needs no target credentials. The write location is also derived from `systemRootPath` without a base-directory constraint. This permits arbitrary command execution, reading and overwriting files owned by the web-server user, persistent PHP code, and control over stored database connection settings. The endpoint additionally interpolates request values into streamer and configuration INSERT statements, creating a secondary SQL-injection sink. This issue is fixed in version 8.1.0.

    Injection Web app

    10.0 CRITICAL 0.5% exploit odds Fix listed
  38. CVE-2026-17140

    IBM API Connect 10.0.8.0 through 10.0.8.12 and 12.1.0.0 through 12.1.1.2 could allow a remote authenticated attacker to execute arbitrary code due to improper control of generation of code. Input supplied by a user can influence code generated by the product, which could lead to code execution.

    Injection IBM Api Connect

    3.3 LOW 0.5% exploit odds Fix listed
  39. CVE-2026-115547

    An input validation issue in the Okta Access Gateway (OAG) Admin Console allows the processing of specially formatted configuration values. An authenticated administrator can leverage this to execute system commands within the context of the OAG service account.

    Injection Web app

    9.1 CRITICAL 0.5% exploit odds No fix listed
  40. CVE-2026-113615

    The GS JWT Authentication for WP REST API WordPress plugin through 1.0.0 does not sanitize or parameterize a request value before using it in SQL queries on several unauthenticated REST routes, allowing unauthenticated attackers to perform SQL injection and obtain an administrator session.

    Injection Web app

    9.8 CRITICAL 0.5% exploit odds No fix listed
  41. CVE-2026-21596

    This High severity SSTI (Server-side Template Injection) vulnerability known as CVE-2026-21596 was introduced in version 12.0.0, 12.1.0. This SSTI (Server-side Template Injection) vulnerability, with a CVSS Score of 7.7 and a CVSS Vector of CVSS:4.0/AV:N/AC:L/AT:P/PR:L/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N allows an authenticated attacker to execute arbitrary commands, which has high impact to confidentiality, high impact to integrity, high impact to availability, and requires no user interaction. Atlassian recommends that Bamboo Data Center customers upgrade to the latest version. If you are unable to do so, upgrade your instance to the specified supported fixed version: Bamboo Data Center 12.1: Upgrade to a release greater than or equal to 12.1.13 See the release notes ([https://confluence.atlassian.com/bambooreleases/bamboo-release-notes-1189793869.html]). You can download the latest version of Bamboo Data Center from the download center ([https://www.atlassian.com/software/bamboo/download-archives]).

    Injection Server Atlassian Bamboo Data Center

    7.7 HIGH 0.5% exploit odds Fix listed
  42. CVE-2026-123837

    In JetBrains IntelliJ IDEA before 2026.3, 2026.2.1 command injection was possible in untrusted projects via Tomcat run configurations on WSL

    Injection Jetbrains Intellij Idea

    7.8 HIGH 0.5% exploit odds Fix listed
  43. CVE-2026-79858

    An unauthenticated SQL injection vulnerability exists in the PAPI protocol in AOS-8. Successful exploitation could allow an unauthenticated remote attacker to obtain and modify sensitive information in the underlying database and to interrupt the normal operation of the affected service.

    Injection Server

    9.8 CRITICAL 0.5% exploit odds No fix listed
  44. CVE-2026-121662

    Improper neutralization of special elements used in a command ('command injection') in Visual Studio Code allows an unauthorized attacker to execute code locally.

    Injection Microsoft Visual Studio Code

    8.6 HIGH 0.5% exploit odds Fix listed
  45. CVE-2026-116533

    The Static Snapshot WordPress plugin through 0.2.1 does not sanitise shell metacharacters in a user-supplied value before concatenating it into operating-system commands run by the server, allowing unauthenticated attackers to execute arbitrary commands on the host.

    Injection Web app

    9.8 CRITICAL 0.5% exploit odds No fix listed
  46. CVE-2026-61567

    AVideo is a video-sharing Platform. Prior to 8.1.0, the live-streaming endpoint `view/videosListToLive.php` accepts an attacker-named streamer in `webSiteRootURL`, calls `Login::run` in `objects/Login.php`, and trusts the returned JSON as session state. An unauthenticated remote attacker who can host an HTTP service reachable by the encoder can self-grant `isLogged` and `canStream` without legitimate credentials. The endpoint requires only nonempty `playlists_id` and `APISecret` values; `APISecret` is forwarded rather than checked against a stored secret. This path never calls `Streamer::isURLAllowed` and therefore works regardless of the operator's streamer allowlist. It fetches a second attacker-controlled video-list response and interpolates `videos[].path` without shell argument escaping into `ffprobe` and `ffmpeg` commands passed to `exec`; `isAudio` calls `isVideo` to reach the `ffprobe` command. Paths containing shell delimiters execute arbitrary operating-system commands as the web-server user. The request filter does not inspect the video-list JSON. This can disclose database and stored streamer credentials, compromise the upstream platform using those credentials, modify files owned by the web-server user, establish persistence, and deny encoder service. This issue is fixed in version 8.1.0.

    Injection Web app

    10.0 CRITICAL 0.5% exploit odds Fix listed
  47. CVE-2026-17134

    IBM API Connect 12.1.0.0 through 12.1.1.2 could allow a remote authenticated attacker to execute arbitrary SQL commands due to SQL injection.

    Injection Server IBM Api Connect

    9.9 CRITICAL 0.5% exploit odds Fix listed
  48. CVE-2026-54142

    OpenHuman is an open-source agent harness available as a desktop application and Rust library. Before 0.57.13, background synchronization for connected services such as email, Slack, and Notion stores raw external content and later places it in the model context of autonomous background agents without neutralizing embedded instructions. ApprovalGate treats tool calls without an interactive chat context as pre-authorized automation, so attacker-authored content processed during a background tick can influence the model to invoke Composio actions without presenting an approval request. An attacker who can deliver content to a connected account can use this path to send or forward data, post messages, create or modify third-party content, or disclose information from OpenHuman memory and integrations. The issue requires an active integration and successful model interpretation of the injected instructions. This issue is fixed in version 0.57.13.

    Injection Library

    8.7 HIGH 0.5% exploit odds Fix listed
  49. CVE-2026-106763

    An issue in @justice8096/sagemath-mcp-server 1.0.1 allows a remote attacker to execute arbitrary code via the escapeShellString and sanitizeCode functions.

    Injection Server

    7.5 HIGH 0.5% exploit odds No fix listed
  50. CVE-2026-106764

    An issue in @rog0x/mcp-git-tools 1.0.1 allows a remote attacker to execute arbitrary code via the resolveBaseBranch function.

    Injection

    7.5 HIGH 0.5% exploit odds No fix listed