CVE Vulnerability Catalog

Wekan before version 9.31 has a vulnerability allowing arbitrary file read and DoS. A logged-in board member can insert an attachment with a controlled file path, and the server reads it without checking if it is within the allowed storage directory.

WWBN AVideo versions 29.0 and below contain a Stored XSS vulnerability in the MessageSQLite WebSocket handler. An attacker can bypass sanitization by placing the payload in the 'json' key instead of 'msg', allowing arbitrary JavaScript execution in other users' browsers.

The vulnerability in AdonisJS @adonisjs/bodyparser incompletely fixes CVE-2026-25754, allowing prototype pollution attacks via nested multipart fields like user.__proto__.polluted, which cause lodash _.set() via @poppinss/utils to create plain intermediate objects and pollute Object.prototype.

A vulnerability in Sandboxie-Plus before version 1.17.6 allows a sandboxed process to execute arbitrary code in an unsandboxed host process. The GuiServer::WndHookRegisterSlave and GuiServer::WndHookNotifySlave functions do not validate thread handles and function pointers, enabling privilege escalation.

A vulnerability in NanoMQ version 0.24.9 allows a remote attacker to cause a denial of service (DoS) via the nni_qos_db_set function in the broker_tcp.c component.

9Router prior to 0.5.2 allows a remote authenticated attacker to achieve arbitrary code execution on the host OS. The vulnerability combines a Host header bypass of localhost-only routes with unvalidated MCP plugin args passed to child_process.spawn(), enabling command execution via malicious plugins.

A vulnerability in the MCP Python SDK (mcp) prior to version 1.28.1 affects the deprecated WebSocket transport, which did not validate Host or Origin headers. This allowed arbitrary third parties to establish WebSocket connections to applications using that transport.

In 9Router prior to version 0.5.4, the PATCH /api/settings endpoint writes the entire request body to persistent settings without a field whitelist. This allows an authenticated user to set security-critical fields such as requireLogin and disable authentication for the whole application, exposing protected routes to unauthenticated access.

In 9Router version 0.4.45 and earlier, the isLocalRequest() function in src/dashboardGuard.js uses Host and Origin headers to protect /api/mcp/*, /api/tunnel/*, and /api/cli-tools/* endpoints. This allows bypassing security through header spoofing in reverse proxy or tunnel deployments, enabling access to MCP child process stdin paths.

A vulnerability in the MCP Python SDK (mcp) versions 1.23.0 through 1.27.2 allows any client to access other clients' tasks. Default handlers for task operations do not verify the session, enabling enumeration, reading results, consuming messages, or canceling tasks of other clients.

A vulnerability in MCP Python SDK (mcp) before version 1.27.2 allows an attacker to inject JSON-RPC messages into an existing session. The issue affects SSE and Streamable HTTP transports, which do not verify the authenticated client's identity when using the session_id parameter or Mcp-Session-Id header.

In the ncnn framework, a vulnerability allows an out-of-bounds heap write. The issue occurs in ParamDict::load_param() when loading a malicious .param model file, as a negative parameter ID can index before the params[NCNN_MAX_PARAM_COUNT] array.

A vulnerability in DataEase before version 2.10.23 allows remote code execution by uploading a crafted payload.zip via the Excel upload API /datasource/upload. The attacker creates an H2 datasource using the zip: protocol and executes an SQL query that triggers precompiled Java aliases in test.mv.db, leading to arbitrary code execution.

A vulnerability in DataEase before version 2.10.23 allows arbitrary SQL read queries on connected datasources. The SQL preview function does not enforce table access restrictions.

A vulnerability in Argo CD allows users with application write access to set link.argocd.argoproj.io/* annotations whose pipe-separated values are rendered as anchor href values in the Summary tab URLs section without URL validation, enabling JavaScript execution in a higher-privileged user's authenticated session. The issue is fixed in versions 3.2.12, 3.3.10, and 3.4.2.

In DataEase prior to version 2.10.23, SQL-type datasets store attacker-controlled SQL variable default values such as ${var}, and SqlparserUtils.handleVariableDefaultValue() inserts them using String.replace() without escaping or parameterization, causing stored SQL injection whenever a user with dataset read permission accesses the dataset.

In DataEase prior to version 2.10.23, a vulnerability was found in the bulk delete API endpoint where path traversal sequences like ../ are not sanitized. Attackers can delete arbitrary server directories through the export task cleanup mechanism.

In DataEase prior to version 2.10.23, a vulnerability was found due to insufficient path validation in template saving endpoints. An attacker can use attacker-controlled static resource names and Base64 content to perform path traversal and arbitrary file writes on the server.

In DataEase prior to version 2.10.23, a SQL injection vulnerability was found in the datasource connection status check function. An attacker can exploit the schema parameter in the SQL query, leading to unauthorized data access.

In DataEase prior to version 2.10.23, a SQL injection vulnerability was found. Dashboard SQL variables like ${deptId} are processed by SqlparserUtils.transFilter(), which returns raw user input for operators other than 'in' and 'between', and then spliced into the dashboard SQL without proper sanitization. This allows authenticated users with dashboard access to inject malicious SQL into integrated datasources.