CVE Vulnerability Catalog

An integer truncation issue in XkbResizeKeyType() that could lead to a heap buffer overflow exists in X.org X server before 21.1.25 and XWayland before 24.1.14. Authenticated X clients could exploit it for crashes or potential code execution in the X server.

OpenVPN ovpn-dco-win versions 0.6.5 through 2.8.13 allow unprivileged local users to cause a denial of service (system crash) via a race condition in the TCP connect path.

Uncontrolled recursion in MessagePack\BufferUnpacker::unpack() in rybakit/msgpack versions prior to 0.11.0. An attacker can cause memory exhaustion when decoding deeply nested MessagePack input.

An integer overflow in the MessagePack\BufferUnpacker component of rybakit/msgpack versions prior to 0.11.0. On 32-bit PHP, crafted MessagePack input can cause denial of service.

A heap buffer overflow in handling GLX RenderLarge requests where dataBytes exceeds cmdlen exists in X.org X server before 21.1.25 and XWayland before 24.1.14. It can be triggered by an authenticated X client and cause crashes or code execution in the X server context.

In the SAIL library prior to 1.0.2, the built-in WebP decoder can crash during loader initialization when processing a malformed animated WebP file with inconsistent container or ANMF frame metadata. A zero image width is used as a divisor when computing bytes_per_pixel, causing an arithmetic fault and process termination.

In identrail prior to 1.0.2, the GitHub connect/complete endpoint accepts installation_id without verifying the installation belongs to the caller's workspace. An authenticated tenant can supply another customer's installation_id to read the victim's private repositories or drive posture scans if identrail's GitHub App is installed there.

In @openhop/server prior to 0.3.7, unauthenticated flow route parameters from GET and DELETE /api/flows/:id are passed to FlowStore.filePath() without an identifier allow-list. After URI decoding and path.join() normalization, an attacker can read or permanently delete YAML files outside the flow directory, and CORS origin: true enables attacks from any browser origin.

In ESPHome Device Builder Dashboard prior to 1.0.10, the Home Assistant add-on's trusted ingress site binds to 0.0.0.0 while skipping dashboard authentication, and host networking required for mDNS exposes it on the LAN interface. Any device on the local network can connect directly to the ingress port, bypass the supervisor, and obtain the full dashboard without credentials.

In functype-mcp-server prior to 1.4.4, the set_functype_version MCP tool accepts an unconstrained version string and interpolates it into functype@<version> for pnpm add, allowing an arbitrary local or remote package to be installed as functype. initDocsData(true) then dynamically imports the newly installed module, executing attacker-controlled JavaScript with the MCP server process's privileges.

In Joker prior to 1.8.2, joker --lint <file> walks upward from the linted file to locate a .jokerd/ directory and executes matching linter.* files before linting. These files contain executable Joker/Clojure code, so linting a file inside an untrusted repository can execute repository-supplied code with the linter user's privileges.

In NORA (a self-hosted artifact registry) prior to 0.9.5, the token-management UI is authenticated but does not require administrator authorization. A read-only token can enumerate metadata of all users' tokens, and a write-capable token can revoke arbitrary admin and service tokens.

In Xibo prior to 4.4.3, missing access-control checks on PUT/DELETE /action/{id} allow an authenticated user to modify or delete actions on any Layout. Standard privileges granted to non-administrators are sufficient to exploit this.

A vulnerability in Xibo (versions prior to 4.4.3) involves missing access-control checks on the GET /dataset/export/csv/{id} endpoint. An authenticated user can export data from any dataset, including those created by administrators, by iterating over sequential dataset IDs. Exploitation requires the privilege allowing edits and deletion of all datasets, which is not granted to non-administrators by default.

The WordPress plugin Filter Everything — WordPress & WooCommerce Filters through 1.9.7 contains a deserialization of untrusted data vulnerability allowing Object Injection.

The WordPress plugin Phoenix Media Rename through 3.13.5 contains a deserialization of untrusted data vulnerability allowing Object Injection.

In the Elixir bandit library from 0.3.1 before 1.14.1, the HTTP/2 request body reader lacks resource limits, allowing an unauthenticated remote attacker to exhaust server memory via a request body sent as many 1-byte DATA frames. Routes whose Plug reads the request body, including those behind Plug.Parsers, are affected.

A deserialization of untrusted data vulnerability in the Ultimate Member WordPress plugin allows Object Injection. Affects versions up to and including 2.13.1.

The WordPress plugin Advanced Ads – Tracking before 4.0.0 contains an SQL injection vulnerability allowing blind SQL injection.

The WordPress plugin Dokan (dokan-lite) through 5.3.1 contains a deserialization of untrusted data vulnerability allowing Object Injection.