CVE Database

`@zereight/mcp-gitlab` is a Model Context Protocol server for GitLab. Prior to version 2.1.27, the SSE transport mode (`SSE=true`) exposes all MCP tools without any authentication. The `upload_markdown` tool reads arbitrary files from the server's local filesystem via an unsanitized `file_path` parameter and uploads them to a GitLab project. Combined, any unauthenticated network-reachable attacker can read `/proc/self/environ` to steal the server's `GITLAB_PERSONAL_ACCESS_TOKEN` and achieve full GitLab account takeover. This is the default configuration for Docker deployments. Version 2.1.27 contains a patch.

Rsbuild before 2.0.9 contains a command injection vulnerability that allows attackers to execute arbitrary OS commands by supplying a crafted URL containing shell metacharacters to the server.open configuration on macOS. The openBrowser() function in packages/core/src/server/open.ts passes the URL through encodeURI() before interpolating it into a shell command executed via child_process.exec(), but because encodeURI() does not encode dollar signs, parentheses, or semicolons, embedded shell metacharacters are evaluated by /bin/sh, enabling arbitrary command execution.

Insertion of sensitive information into log file in the slow query logging feature in Devolutions PowerShell Universal 2026.2.5 and earlier allows an authenticated user with log read permission to obtain application tokens, data protection key material and other stored credentials via SQL parameter values written to the system log on instances backed by Microsoft SQL Server.

QloApps through 1.7.0 reflects unescaped child feature names into back-office validation error messages in the Hotel Reservation System feature management page. Authenticated back-office users who follow a crafted link can execute injected JavaScript in their administrative session via the child_features parameter.

adm-zip versions 0.5.14 through 0.6.0 fail to apply zlib decompression output limits when ZIP entries declare zero uncompressed size. Attackers can craft malicious ZIP archives with highly compressible entries declaring zero size to exhaust memory and cause denial of service.

Cotonti 1.0.0 Comments plugin passes the ci GET parameter to unserialize() without allowed_classes restriction, allowing unauthenticated attackers to instantiate arbitrary PHP classes with attacker-controlled properties. Attackers can exploit PHP object injection through crafted serialized payloads to trigger gadget chains and achieve database manipulation or code execution.

Use after free in Workers in Google Chrome prior to 153.0.8010.47 allowed a remote attacker to potentially execute arbitrary code outside the sandbox via a crafted HTML page. (Chromium security severity: Critical)

Confused deputy in PriceTracking in Google Chrome on on iOS prior to 153.0.8010.47 allowed a remote attacker leveraging social engineering to bypass system access restrictions into a privileged page via crafted network traffic. (Chromium security severity: Medium)

Missing authorization in AppManifest in Google Chrome prior to 153.0.8010.47 allowed a remote attacker who had compromised the renderer process and leveraged social engineering to bypass web origin policy via a crafted HTML page. (Chromium security severity: Medium)

Race condition in WebAppInstalls in Google Chrome prior to 153.0.8010.47 allowed a remote attacker to spoof UI elements via a crafted HTML page. (Chromium security severity: Medium)

Code injection in XML in Google Chrome prior to 153.0.8010.47 allowed a remote attacker to bypass web origin policy via a crafted HTML page. (Chromium security severity: Low)

`tts-be` is a backend for a timetable selector that aims to help students better choose their class schedules. Versions prior to 2.1.0 have a Broken Access Control vulnerability across several API endpoints (such as `/api/student/{id}/photo` and `/api/course_unit/{id}/exchange/metadata`). By chaining these unauthenticated endpoints, a remote attacker can use the backend as an open proxy to bypass authorization checks, allowing for the enumeration and extraction of sensitive Personally Identifiable Information (PII) from upstream university systems. The exposed data includes full names, student IDs, class schedules, and photos. This issue was fixed in version 2.1.0.

Concrete CMS before 9.5.3 contained a stored cross-site scripting vulnerability in SVG file handling. When SVG processing was set to the non-default "Reject files containing potentially harmful elements" mode (concrete.file_manager.images.svg_sanitization.action = reject), uploaded SVGs were checked only against a small built-in blocklist covering the script element and on* event-handler attributes; the broader enshrined/svg-sanitize pass still ran, but its result was discarded, so vectors it would have stripped, such as a javascript: URI in an xlink:href, were stored unmodified and executed when the file was opened directly. A user able to upload files could thereby run arbitrary JavaScript in the browser of any user who viewed the file. The default sanitize mode was not affected. The Concrete CMS security team gave this vulnerability a CVSS v4.0 score of 1.8 with vector CVSS:4.0/AV:N/AC:L/AT:P/PR:H/UI:A/VC:L/VI:L/VA:N/SC:N/SI:N/SA:N. Thanks Oriol Ortiz for reporting.

Concrete CMS 9.4.0 through 9.5.2 did not escape colliding page paths before rendering them in the location panel's duplicate-path confirmation dialog. The panel's check endpoint returned the submitted path unmodified in its JSON response, and client-side JavaScript inserted each value into the dialog as raw HTML, so a crafted page path executed script in the editor's authenticated browser session. The Concrete CMS security team gave this vulnerability a CVSS v4.0 score of 2.0 with vector CVSS:4.0/AV:N/AC:L/AT:P/PR:H/UI:P/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N. Thanks Yonatan Drori from Tenzai for reporting.

The guest-to-host Unix-domain socket relay in Docker Sandboxes validates that a socket path is inside an authorized workspace, but later reconnects using the pathname. A malicious guest can replace an intermediate directory with a symlink between validation and connection, causing the host to connect to an arbitrary AF_UNIX socket outside the shared workspace. This can expose data or host-side capabilities provided by the targeted socket.

The affected products are vulnerable to an authentication bypass that allows unauthenticated remote attackers to disclose sensitive device information, including administrator credentials in plaintext, by sending crafted HTTP(S) requests.

The affected products use hard-coded credentials, which could allow an attacker to run the ftpd service as root, providing remote root file access where FTP is reachable.

An insufficient check allowed for the overwrite of arbitrary files via a symlink.

The affected products are missing authentication for a critical function, which could allow an attacker to run as root and pass received bytes directly to a system command.

The affected products use hard-coded credentials, which could allow remote access to files with root privileges where FTP is reachable.

The affected products are missing authorization on state-changing CGIs and session checks are not performed.

The affected products use insufficiently random values, which allows web session tokens to be predictable, bounding token entropy to the seed space.

`@zereight/mcp-gitlab` is a Model Context Protocol server for GitLab. Versions prior to 2.1.30 expose the Streamable HTTP MCP endpoint without an effective Host or Origin allowlist. A malicious web page can use DNS rebinding to route browser requests to a victim's local MCP listener while preserving an attacker-controlled `Host` and `Origin`. The server accepts those headers and reaches the MCP initialization path instead of rejecting the request at the HTTP boundary. Version 2.1.30 contains a patch.

`@zereight/mcp-gitlab` is a Model Context Protocol server for GitLab. Starting in version 0.0.1 and prior to version 2.1.27, when the environment variable `ENABLE_DYNAMIC_API_URL=true` is set, the server reads the `X-GitLab-API-URL` HTTP request header and uses it as the base URL for all outbound GitLab API calls made within that request. The server validates that the value is a well-formed URL (`new URL(dynamicApiUrl)`) but applies no allowlist or hostname restriction. The server then attaches the victim's `Private-Token` to every outbound fetch that uses the redirected URL. Any caller who can reach the HTTP transport can set `X-GitLab-API-URL` to an attacker-controlled host. The next GitLab API call the server makes delivers the victim's token to that host. Version 2.1.27 contains a patch.

emp3r0r is a C2 designed by Linux users for Linux environments. Prior to version 4.2.5, the `http_poll` C2 transport accepts attacker-controlled HTTP polling sessions before CBOR `MsgAuth` authentication is completed. A remote unauthenticated attacker can create arbitrary polling sessions and send request bodies that are forwarded into the C2 dispatch path. This can consume server resources and trigger pre-auth C2 processing. Version 4.2.5 patches the issue.