Início / CVEs
Base de CVEs
Vulnerabilidades publicadas, com severidade, exploração conhecida e fontes oficiais.
CVE-2026-26682
Alta · 7.8
CWE-94 26/02/2026
Code Injection em Xjd2020 Fastcms
An issue in fastCMS before v.0.1.6 allows a local attacker to execute arbitrary code via the PluginController.java component
CVE-2026-27966
Crítica · 9.8
Alta · 45 Exploit público CWE-94 25/02/2026
Langflow, agente CSV: injeção de código e execução remota (RCE)
Langflow é uma ferramenta para construir e implantar agentes e fluxos de trabalho baseados em IA. Antes da versão 1.8.0, o nó CSV Agent do Langflow fixa no código o valor allow_dangerous_code=True, o que expõe automaticamente a ferramenta REPL do Python do LangChain (python_repl_ast). Como resultado, um atacante pode executar Python arbitrário e comandos do sistema operacional no servidor por meio de injeção de prompt, levando a uma execução remota de código (RCE) completa. A versão 1.8.0 corrige o problema.
CVE-2026-27952
Crítica · 9.9
CWE-94 25/02/2026
Agenta has Python Sandbox Escape, Leading to Remote Code Execution (RCE)
Agenta is an open-source LLMOps platform. In Agenta-API prior to version 0.48.1, a Python sandbox escape vulnerability existed in Agenta's custom code evaluator. Agenta used RestrictedPython as a sandboxing mechanism for user-supplied evaluator code, but incorrectly whitelisted the `numpy` package as safe within the sandbox. This allowed authenticated users to bypass the sandbox and achieve arbitrary code execution on the API server. The escape path was through `numpy.ma.core.inspect`, which exposes Python's introspection utilities — including `sys.modules` — thereby providing access to unfiltered system-level functionality like `os.system`. This vulnerability affects the Agenta self-hosted platform (API server), not the SDK when used as a standalone Python library. The custom code evaluator runs server-side within the API process. The issue is fixed in v0.48.1 by removing `numpy` from the sandbox allowlist. In later versions (v0.60+), the RestrictedPython sandbox was removed entirely and replaced with a different execution model.
CVE-2026-27830
Alta · 8.0
CWE-94 25/02/2026
c3p0 vulnerable to Remote Code Execution via unsafe deserialization of userOverridesAsString property
c3p0, a JDBC Connection pooling library, is vulnerable to attack via maliciously crafted Java-serialized objects and `javax.naming.Reference` instances. Several c3p0 `ConnectionPoolDataSource` implementations have a property called `userOverridesAsString` which conceptually represents a `Map<String,Map<String,String>>`. Prior to v0.12.0, that property was maintained as a hex-encoded serialized object. Any attacker able to reset this property, on an existing `ConnectionPoolDataSource` or via maliciously crafted serialized objects or `javax.naming.Reference` instances could be tailored execute unexpected code on the application's `CLASSPATH`. The danger of this vulnerability was strongly magnified by vulnerabilities in c3p0's main dependency, mchange-commons-java. This library includes code that mirrors early implementations of JNDI functionality, including ungated support for remote `factoryClassLocation` values. Attackers could set c3p0's `userOverridesAsString` hex-encoded serialized objects that include objects "indirectly serialized" via JNDI references. Deserialization of those objects and dereferencing of the embedded `javax.naming.Reference` objects could provoke download and execution of malicious code from a remote `factoryClassLocation`. Although hazard presented by c3p0's vulnerabilites are exarcerbated by vulnerabilities in mchange-commons-java, use of Java-serialized-object hex as the format for a writable Java-Bean property, of objects that may be exposed across JNDI interfaces, represents a serious independent fragility. The `userOverridesAsString` property of c3p0 `ConnectionPoolDataSource` classes has been reimplemented to use a safe CSV-based format, rather than rely upon potentially dangerous Java object deserialization. c3p0-0.12.0+ and above depend upon mchange-commons-java 0.4.0+, which gates support for remote `factoryClassLocation` values by configuration parameters that default to restrictive values. c3p0 additionally enforces the new mchange-commons-java `com.mchange.v2.naming.nameGuardClassName` to prevent injection of unexpected, potentially remote JNDI names. There is no supported workaround for versions of c3p0 prior to 0.12.0.
CVE-2026-27577
Crítica · 9.9
CWE-94 25/02/2026
n8n: Expression Sandbox Escape Leads to RCE
n8n is an open source workflow automation platform. Prior to versions 2.10.1, 2.9.3, and 1.123.22, additional exploits in the expression evaluation of n8n have been identified and patched following CVE-2025-68613. An authenticated user with permission to create or modify workflows could abuse crafted expressions in workflow parameters to trigger unintended system command execution on the host running n8n. The issues have been fixed in n8n versions 2.10.1, 2.9.3, and 1.123.22. Users should upgrade to one of these versions or later to remediate all known vulnerabilities. If upgrading is not immediately possible, administrators should consider the following temporary mitigations. Limit workflow creation and editing permissions to fully trusted users only, and/or deploy n8n in a hardened environment with restricted operating system privileges and network access to reduce the impact of potential exploitation. These workarounds do not fully remediate the risk and should only be used as short-term mitigation measures.
CVE-2026-27498
Alta · 8.8
CWE-94 25/02/2026
n8n has Arbitrary Command Execution via File Write and Git Operations
n8n is an open source workflow automation platform. Prior to versions 2.2.0 and 1.123.8, an authenticated user with permission to create or modify workflows could chain the Read/Write Files from Disk node with git operations to achieve remote code execution. By writing to specific configuration files and then triggering a git operation, the attacker could execute arbitrary shell commands on the n8n host. The issue has been fixed in n8n versions 2.2.0 and 1.123.8. Users should upgrade to one of these versions or later to remediate the vulnerability. If upgrading is not immediately possible, administrators should consider the following temporary mitigations. Limit workflow creation and editing permissions to fully trusted users only, and/or disable the Read/Write Files from Disk node by adding `n8n-nodes-base.readWriteFile` to the `NODES_EXCLUDE` environment variable. These workarounds do not fully remediate the risk and should only be used as short-term mitigation measures.
CVE-2026-27495
Crítica · 9.9
CWE-94 25/02/2026
Escape de sandbox no JavaScript Task Runner do n8n (CWE-94)
O n8n é uma plataforma de automação de workflows de código aberto. Antes das versões 2.10.1, 2.9.3 e 1.123.22, um usuário autenticado com permissão para criar ou modificar workflows poderia explorar uma vulnerabilidade no sandbox do JavaScript Task Runner para executar código arbitrário fora da fronteira do sandbox. Em instâncias que usam Task Runners internos (modo de execução padrão), isso poderia resultar no comprometimento total do host do n8n. Em instâncias que usam Task Runners externos, o atacante poderia obter acesso a outras tarefas executadas no Task Runner ou afetá-las. Os Task Runners precisam ser habilitados com N8N_RUNNERS_ENABLED=true. O problema foi corrigido nas versões 2.10.1, 2.9.3 e 1.123.22 do n8n. Usuários devem atualizar para uma dessas versões ou posterior para remediar a vulnerabilidade. Se a atualização não for possível imediatamente, administradores devem considerar as seguintes mitigações temporárias: limitar as permissões de criação e edição de workflows apenas a usuários totalmente confiáveis e/ou usar o modo de runner externo (N8N_RUNNERS_MODE=external) para reduzir o raio de impacto. Essas medidas não remediam totalmente o risco e devem ser usadas apenas como soluções de curto prazo.
CVE-2026-27493
Crítica · 9.0
CWE-94 25/02/2026
n8n has Unauthenticated Expression Evaluation via Form Node
n8n is an open source workflow automation platform. Prior to versions 2.10.1, 2.9.3, and 1.123.22, a second-order expression injection vulnerability existed in n8n's Form nodes that could allow an unauthenticated attacker to inject and evaluate arbitrary n8n expressions by submitting crafted form data. When chained with an expression sandbox escape, this could escalate to remote code execution on the n8n host. The vulnerability requires a specific workflow configuration to be exploitable. First, a form node with a field interpolating a value provided by an unauthenticated user, e.g. a form submitted value. Second, the field value must begin with an `=` character, which caused n8n to treat it as an expression and triggered a double-evaluation of the field content. There is no practical reason for a workflow designer to prefix a field with `=` intentionally — the character is not rendered in the output, so the result would not match the designer's expectations. If added accidentally, it would be noticeable and very unlikely to persist. An unauthenticated attacker would need to either know about this specific circumstance on a target instance or discover a matching form by chance. Even when the preconditions are met, the expression injection alone is limited to data accessible within the n8n expression context. Escalation to remote code execution requires chaining with a separate sandbox escape vulnerability. The issue has been fixed in n8n versions 2.10.1, 2.9.3, and 1.123.22. Users should upgrade to one of these versions or later to remediate the vulnerability. If upgrading is not immediately possible, administrators should consider the following temporary mitigations. Review usage of form nodes manually for above mentioned preconditions, disable the Form node by adding `n8n-nodes-base.form` to the `NODES_EXCLUDE` environment variable, and/or disable the Form Trigger node by adding `n8n-nodes-base.formTrigger` to the `NODES_EXCLUDE` environment variable. These workarounds do not fully remediate the risk and should only be used as short-term mitigation measures.
CVE-2026-27702
Crítica · 9.0
CWE-94 25/02/2026
Injeção de código no Budibase Cloud via eval() inseguro em filtro de views
O Budibase é uma plataforma low code para criação de ferramentas internas, fluxos de trabalho e painéis administrativos. Antes da versão 3.30.4, uma vulnerabilidade de eval() inseguro na implementação de filtragem de views do Budibase permite que qualquer usuário autenticado (incluindo contas do plano gratuito) execute JavaScript arbitrário no servidor. Esta vulnerabilidade afeta SOMENTE o Budibase Cloud (SaaS); implantações self-hosted usam views nativas do CouchDB e não são vulneráveis. A vulnerabilidade existe em packages/server/src/db/inMemoryView.ts, onde funções de mapeamento de views controladas pelo usuário são avaliadas diretamente sem sanitização. O principal impacto vem do que está no ambiente do pod: o pod app-service executa com segredos incorporados em suas variáveis de ambiente, incluindo INTERNAL_API_KEY, JWT_SECRET, credenciais de administrador do CouchDB, chaves AWS e outros. Usando as credenciais do CouchDB extraídas, os pesquisadores verificaram acesso direto ao banco de dados, enumeraram todos os bancos de dados de tenants e confirmaram que registros de usuários (endereços de e-mail) são legíveis. A versão 3.30.4 contém a correção.
CVE-2026-27701
Alta · 8.8
CWE-94 25/02/2026
LiveCodes vulnerable to JavaScript Injection via untrusted PR title in i18n-update-pull workflow
LiveCode is an open-source, client-side code playground. Prior to commit e151c64c2bd80d2d53ac1333f1df9429fe6a1a11, LiveCode's `i18n-update-pull` GitHub Actions workflow is vulnerable to JavaScript injection. The title of the Pull Request associated with the triggering issue comment is interpolated directly into a `actions/github-script` JavaScript block using a GitHub Actions template expression. An attacker who opens a PR with a crafted title can inject arbitrary JavaScript that executes with the privileges of the CI bot token (`CI_APP_ID` / `CI_APP_PRIVATE_KEY`), enabling exfiltration of repository secrets and unauthorized GitHub API operations. Commit e151c64c2bd80d2d53ac1333f1df9429fe6a1a11 fixes the issue.
CVE-2026-1929
Alta · 8.8
CWE-94 25/02/2026
Advanced Woo Labels <= 2.37 - Authenticated (Contributor+) Remote Code Execution via 'callback' Parameter
The Advanced Woo Labels plugin for WordPress is vulnerable to Remote Code Execution in all versions up to, and including, 2.37. This is due to the use of `call_user_func_array()` with user-controlled callback and parameters in the `get_select_option_values()` AJAX handler without an allowlist of permitted callbacks or a capability check. This makes it possible for authenticated attackers, with Contributor-level access and above, to execute arbitrary PHP functions and operating system commands on the server via the 'callback' parameter.
CVE-2026-27745
Alta · 8.8
CWE-94 25/02/2026
SPIP interface_traduction_objets < 2.2.2 Authenticated RCE
The SPIP interface_traduction_objets plugin versions prior to 2.2.2 contain an authenticated remote code execution vulnerability in the translation interface workflow. The plugin incorporates untrusted request data into a hidden form field that is rendered without SPIP output filtering. Because fields prefixed with an underscore bypass protection mechanisms and the hidden content is rendered with filtering disabled, an authenticated attacker with editor-level privileges can inject crafted content that is evaluated through SPIP's template processing chain, resulting in execution of code in the context of the web server.
CVE-2026-27744
Crítica · 9.8
CWE-94 25/02/2026
Plugin SPIP Tickets: injeção de código e execução remota sem autenticação
A versão do plugin SPIP tickets anteriores a 4.3.3 contém uma vulnerabilidade de execução remota de código não autenticada no tratamento da pré-visualização de fórum para páginas públicas de tickets. O plugin anexa parâmetros de requisição não confiáveis ao HTML que é posteriormente renderizado por um template que usa renderização de ambiente sem filtro (#ENV**), o que desativa a filtragem de saída do SPIP. Como resultado, um atacante não autenticado pode injetar conteúdo elaborado que é avaliado pela cadeia de processamento de templates do SPIP, levando à execução de código no contexto do servidor web.
CVE-2026-27597
Crítica · 10.0
CWE-94 25/02/2026
Fuga de Sandbox no Agentfront Enclave (@enclave-vm/core) com Potencial de Execução Remota de Código
Enclave é um sandbox JavaScript seguro projetado para a execução segura de código de agentes de IA. Antes da versão 2.11.1, é possível escapar dos limites de segurança definidos pelo `@enclave-vm/core`, o que pode ser usado para obter execução remota de código (RCE). O problema foi corrigido na versão 2.11.1.
CVE-2024-56373
Alta · 8.4
CWE-94 24/02/2026
Apache Airflow: SSTI to Code Execution in Airflow through Shared DB Information
DAG Author (who already has quite a lot of permissions) could manipulate database of Airflow 2 in the way to execute arbitrary code in the web-server context, which they should normally not be able to do, leading to potentially remote code execution in the context of web-server (server-side) as a result of a user viewing historical task information.
The functionality responsible for that (log template history) has been disabled by default in 2.11.1 and users should upgrade to Airflow 3 if they want to continue to use log template history. They can also manually modify historical log file names if they want to see historical logs that were generated before the last log template change.
CVE-2026-25797
Média · 5.3
CWE-94 23/02/2026
ImageMagick vulnerable to Code injection via PostScript header in ps coders
ImageMagick is free and open-source software used for editing and manipulating digital images. Prior to versions 7.1.2-15 and 6.9.13-40, the ps coders, responsible for writing PostScript files, fails to sanitize the input before writing it into the PostScript header. An attacker can provide a malicous file and inject arbitrary PostScript code. When the resulting file is processed by a printer or a viewer (like Ghostscript), the injected code is interpreted and executed. The html encoder does not properly escape strings that are written to in the html document. An attacker can provide a malicious file and injection arbitrary html code. Versions 7.1.2-15 and 6.9.13-40 contain a patch.
CVE-2025-9120
Alta · 8.6
CWE-94 23/02/2026
RCE vulnerability has been discovered in OpenText™ Carbonite Safe Server Backup.
Improper Control of Generation of Code ('Code Injection') vulnerability in OpenText™ Carbonite Safe Server Backup allows Code Injection.
The vulnerability could be exploited through an open port, potentially allowing unauthorized access.
This issue affects Carbonite Safe Server Backup: through 6.8.3.
CVE-2026-2964
Crítica · 9.8
CWE-94 22/02/2026
higuma web-audio-recorder-js: poluição de protótipo na função extend de WebAudioRecorder.js
Foi identificada uma vulnerabilidade no higuma web-audio-recorder-js 0.1/0.1.1. É afetada a função extend na biblioteca lib/WebAudioRecorder.js do componente Dynamic Config Handling. Essa manipulação leva a modificação inadequadamente controlada de atributos do protótipo de objetos. É possível realizar o ataque remotamente. Ataques desse tipo são de alta complexidade. A explorabilidade é considerada difícil. O exploit está publicamente disponível e pode ser utilizado. O fornecedor foi contatado cedo sobre esta divulgação, mas não respondeu de nenhuma forma.
CVE-2026-27574
Crítica · 9.9
CWE-94 21/02/2026
OneUptime: escape do sandbox node:vm no probe leva a RCE
O OneUptime é uma solução para monitoramento e gerenciamento de serviços online. Nas versões 9.5.13 e anteriores, o recurso de monitor JavaScript personalizado usa o módulo node:vm do Node.js (explicitamente documentado como não sendo um mecanismo de segurança) para executar código fornecido pelo usuário, permitindo um escape trivial do sandbox por meio de um one-liner bem conhecido que concede acesso total ao processo subjacente. Como o probe é executado com rede do host e mantém todas as credenciais do cluster (ONEUPTIME_SECRET, DATABASE_PASSWORD, REDIS_PASSWORD, CLICKHOUSE_PASSWORD) nas variáveis de ambiente, e a criação de monitores está disponível para o papel mais baixo (ProjectMember) com registro aberto habilitado por padrão, qualquer usuário anônimo pode alcançar o comprometimento total do cluster em cerca de 30 segundos. Este problema foi corrigido na versão 10.0.5.
CVE-2026-27464
Média · 6.5
CWE-94 21/02/2026
Metabase: Server-Side Template Injection via Notifications Endpoint Leads to RCE
Metabase is an open-source data analytics platform. In versions prior to 0.57.13 and versions 0.58.x through 0.58.6, authenticated users are able to retrieve sensitive information from a Metabase instance, including database access credentials. During testing, it was confirmed that a low-privileged user can extract sensitive information including database credentials, into the email body via template evaluation. This issue has been fixed in versions 0.57.13 and 0.58.7. To workaround this issue, users can disable notifications in their Metabase instance to disallow access to the vulnerable endpoints.
CVE-2026-26045
Alta · 7.2
CWE-94 21/02/2026
Moodle: moodle: improper validation in file restore functionality leading to remote code execution
A flaw was identified in Moodle’s backup restore functionality where specially crafted backup files were not properly validated during processing. If a malicious backup file is restored, it could lead to unintended execution of server-side code. Since restore capabilities are typically available to privileged users, exploitation requires authenticated access. Successful exploitation could result in full compromise of the Moodle server.
CVE-2025-67979
Crítica · 9.9
CWE-94 20/02/2026
WordPress WPForms Google Sheet Connector plugin <= 4.0.1 - Remote Code Execution (RCE) vulnerability
Improper Control of Generation of Code ('Code Injection') vulnerability in WesternDeal WPForms Google Sheet Connector gsheetconnector-wpforms allows Code Injection.This issue affects WPForms Google Sheet Connector: from n/a through <= 4.0.1.
CVE-2025-52744
Alta · 7.7
CWE-94 20/02/2026
WordPress Inpersttion For Theme plugin <= 1.0 - Arbitrary Code Execution vulnerability
Improper Control of Generation of Code ('Code Injection') vulnerability in inpersttion Inpersttion For Theme err-our-team allows Code Injection.This issue affects Inpersttion For Theme: from n/a through <= 1.0.
CVE-2026-26030
Crítica · 9.9
CWE-94 19/02/2026
Vulnerabilidade de execução remota de código no filtro InMemoryVectorStore do Microsoft Semantic Kernel
O Semantic Kernel, o SDK Python de kernel semântico da Microsoft, possui uma vulnerabilidade de execução remota de código em versões anteriores à 1.39.4, especificamente na funcionalidade de filtro do `InMemoryVectorStore`. O problema foi corrigido na versão 1.39.4. Os usuários devem atualizar para esta versão ou superior. Como solução alternativa, evite usar o `InMemoryVectorStore` em cenários de produção.
CVE-2025-71243
Crítica · 9.8
Exploit público CWE-94 19/02/2026
SPIP Saisies Plugin < 5.11.1 Remote Code Execution
The 'Saisies pour formulaire' (Saisies) plugin for SPIP versions 5.4.0 through 5.11.0 contains a critical Remote Code Execution (RCE) vulnerability. An attacker can exploit this vulnerability to execute arbitrary code on the server. Users should immediately update to version 5.11.1 or later.