| CVE |
Vendors |
Products |
Updated |
CVSS v3.1 |
| Cursor is a code editor built for programming with AI. Prior to 2.3, hen the Cursor Agent is running in Auto-Run Mode with Allowlist mode enabled, certain shell built-ins can still be executed without appearing in the allowlist and without requiring user approval.
This allows an attacker via indirect or direct prompt injection to poison the shell environment by setting, modifying, or removing environment variables that influence trusted commands. This vulnerability is fixed in 2.3. |
| eBay API MCP Server is an open source local MCP server providing AI assistants with comprehensive access to eBay's Sell APIs. All versions are vulnerable to Environment Variable Injection through the updateEnvFile function. The ebay_set_user_tokens tool allows updating the .env file with new tokens. The updateEnvFile function in src/auth/oauth.ts blindly appends or replaces values without validating them for newlines or quotes. This allows an attacker to inject arbitrary environment variables into the configuration file. An attacker can inject arbitrary environment variables into the .env file. This could lead to configuration overwrites, Denial of Service, and potential RCE. There was no fix for this issue at the time of publication. |
| Out-of-bounds character read vulnerability in Bluetooth. Impact: Successful exploitation of this vulnerability may affect service confidentiality. |
| Improper neutralization of escape, meta, or control sequences in Copilot allows an unauthorized attacker to disclose information over a network. |
| An improper neutralization of escape, meta, or control sequences vulnerability has been reported to affect QHora. If a local attacker gains an administrator account, they can then exploit the vulnerability to cause unexpected behavior.
We have already fixed the vulnerability in the following version:
QuRouter 2.6.3.009 and later |
| Socket Firewall is an HTTP/HTTPS proxy server that intercepts package manager requests and enforces security policies by blocking dangerous packages. Socket Firewall binary versions (separate from installers) prior to 0.15.5 are vulnerable to arbitrary code execution when run in untrusted project directories. The vulnerability allows an attacker to execute arbitrary code by placing a malicious `.sfw.config` file in a project directory. When a developer runs Socket Firewall commands (e.g., `sfw npm install`) in that directory, the tool loads the `.sfw.config` file and populates environment variables directly into the Node.js process. An attacker can exploit this by setting `NODE_OPTIONS` with a `--require` directive to execute malicious JavaScript code before Socket Firewall's security controls are initialized, effectively bypassing the tool's malicious package detection. The attack vector is indirect and requires a developer to install dependencies for an untrusted project and execute a command within the context of the untrusted project. The vulnerability has been patched in Socket Firewall version 0.15.5. Users should upgrade to version 0.15.5 or later. The fix isolates configuration file values from subprocess environments. Look at `sfw --version` for version information. If users rely on the recommended installation mechanism (e.g. global installation via `npm install -g sfw`) then no workaround is necessary. This wrapper package automatically ensures that users are running the latest version of Socket Firewall. Users who have manually installed the binary and cannot immediately upgrade should avoid running Socket Firewall in untrusted project directories. Before running Socket Firewall in any new project, inspect `.sfw.config` and `.env.local` files for suspicious `NODE_OPTIONS` or other environment variable definitions that reference local files. |
| The Cloud MQTT service of the affected products supports wildcard topic
subscription which could allow an attacker to obtain sensitive
information from tapping the service communications. |
| Sandro Poppi, member of the AXIS OS Bug Bounty Program, has found that the VAPIX APIs local_list.cgi, create_overlay.cgi and irissetup.cgi was vulnerable for file globbing which could lead to a resource exhaustion attack. Axis has released patched AXIS OS
versions for the highlighted flaw. Please refer to the Axis security advisory
for more information and solution. |
| PWAsForFirefox is a tool to install, manage and use Progressive Web Apps (PWAs) in Mozilla Firefox. Due to improper sanitization of web app properties (such as name, description, shortcuts), web apps were able to inject additional lines into XDG Desktop Entries (on Linux) and `AppInfo.ini` (on PortableApps.com). This allowed malicious web apps to introduce keys like `Exec`, which could run arbitrary code when the affected web app was launched. This vulnerability affects all Linux and PortableApps.com users of all PWAsForFirefox versions up to (excluding) 2.12.0. Windows and macOS users are not affected. This vulnerability has been fixed in commit `9932d4b` which has been included in release in v2.12.0. The main fix is implemented in the native part, but the extension also contains additional fixes. All Linux and PortableApps.com users are advised to update to this version as soon as possible. It is also recommended for Windows and macOS users to update to this version, as it contains additional fixes related to properties sanitization. There are no known workarounds for this vulnerability. |
| Versions of the package github.com/gitpod-io/gitpod/components/server/go/pkg/lib before main-gha.27122; versions of the package github.com/gitpod-io/gitpod/components/ws-proxy/pkg/proxy before main-gha.27122; versions of the package github.com/gitpod-io/gitpod/install/installer/pkg/components/auth before main-gha.27122; versions of the package github.com/gitpod-io/gitpod/install/installer/pkg/components/public-api-server before main-gha.27122; versions of the package github.com/gitpod-io/gitpod/install/installer/pkg/components/server before main-gha.27122; versions of the package @gitpod/gitpod-protocol before 0.1.5-main-gha.27122 are vulnerable to Cookie Tossing due to a missing __Host- prefix on the _gitpod_io_jwt2_ session cookie. This allows an adversary who controls a subdomain to set the value of the cookie on the Gitpod control plane, which can be assigned to an attacker’s own JWT so that specific actions taken by the victim (such as connecting a new Github organization) are actioned by the attackers session. |
| kubewarden-controller is a Kubernetes controller that allows you to dynamically register Kubewarden admission policies. By design, AdmissionPolicy and AdmissionPolicyGroup can evaluate only namespaced resources. The resources to be evaluated are determined by the rules provided by the user when defining the policy. There might be Kubernetes namespaced resources that should not be validated by AdmissionPolicy and by the AdmissionPolicyGroup policies because of their sensitive nature. For example, PolicyReport are namespaced resources that contain the list of non compliant objects found inside of a namespace. An attacker can use either an AdmissionPolicy or an AdmissionPolicyGroup to prevent the creation and update of PolicyReport objects to hide non-compliant resources. Moreover, the same attacker might use a mutating AdmissionPolicy to alter the contents of the PolicyReport created inside of the namespace. Starting from the 1.21.0 release, the validation rules applied to AdmissionPolicy and AdmissionPolicyGroup have been tightened to prevent them from validating sensitive types of namespaced resources. |
| jte (Java Template Engine) is a secure and lightweight template engine for Java and Kotlin. In affected versions Jte HTML templates with `script` tags or script attributes that include a Javascript template string (backticks) are subject to XSS. The `javaScriptBlock` and `javaScriptAttribute` methods in the `Escape` class do not escape backticks, which are used for Javascript template strings. Dollar signs in template strings should also be escaped as well to prevent undesired interpolation. HTML templates rendered by Jte's `OwaspHtmlTemplateOutput` in versions less than or equal to `3.1.15` with `script` tags or script attributes that contain Javascript template strings (backticks) are vulnerable. Users are advised to upgrade to version 3.1.16 or later to resolve this issue. There are no known workarounds for this vulnerability. |
| gurk (aka gurk-rs) through 0.6.3 mishandles ANSI escape sequences. |
| A CWE-15 "External Control of System or Configuration Setting" in GE Vernova UR IED family devices from version 7.0 up to 8.60 allows an attacker to provide input that establishes a TCP connection through a port forwarding. The lack of the IP address and port validation may allow the attacker to bypass firewall rules or to send malicious traffic in the network. |
| The CloudEdge Cloud does not sanitize the MQTT topic input, which could allow an attacker to leverage the MQTT wildcard to receive all the messages that should be delivered to other users by subscribing to the a MQTT topic. In these messages, the attacker can obtain the credentials and key information to connect to the cameras from peer to peer. |
| tracing is a framework for instrumenting Rust programs to collect structured, event-based diagnostic information. Prior to version 0.3.20, tracing-subscriber was vulnerable to ANSI escape sequence injection attacks. Untrusted user input containing ANSI escape sequences could be injected into terminal output when logged, potentially allowing attackers to manipulate terminal title bars, clear screens or modify terminal display, and potentially mislead users through terminal manipulation. tracing-subscriber version 0.3.20 fixes this vulnerability by escaping ANSI control characters when writing events to destinations that may be printed to the terminal. A workaround involves avoiding printing logs to terminal emulators without escaping ANSI control sequences. |
| For TCAS II systems using transponders compliant with MOPS earlier than RTCA DO-181F, an attacker can impersonate a ground station and issue a Comm-A Identity Request. This action can set the Sensitivity Level Control (SLC) to the lowest setting and disable the Resolution Advisory (RA), leading to a denial-of-service condition. |
| Via the GUI of the "bestinformed Infoclient", a low-privileged user is by default able to change the server address of the "bestinformed Server" to which this client connects. This is dangerous as the "bestinformed Infoclient" runs with elevated permissions ("nt authority\system"). By changing the server address to a malicious server, or a script simulating a server, the user is able to escalate his privileges by abusing certain features of the "bestinformed Web" server. Those features include:
* Pushing of malicious update packages
* Arbitrary Registry Read as "nt authority\system"
An attacker is able to escalate his privileges to "nt authority\system" on the Windows client running the "bestinformed Infoclient".
This attack is not possible if a custom configuration ("Infoclient.ini") containing the flags "ShowOnTaskbar=false" or "DisabledItems=stPort,stAddress" is deployed. |
| Crayfish is a collection of Islandora 8 microservices, one of which, Homarus, provides FFmpeg as a microservice. Prior to Crayfish version 4.1.0, remote code execution may be possible in web-accessible installations of Homarus in certain configurations. The issue has been patched in `islandora/crayfish:4.1.0`. Some workarounds are available. The exploit requires making a request against the Homarus's `/convert` endpoint; therefore, the ability to exploit is much reduced if the microservice is not directly accessible from the Internet, so: Prevent general access from the Internet from hitting Homarus. Alternatively or additionally, configure auth in Crayfish to be more strongly required, such that requests with `Authorization` headers that do not validate are rejected before the problematic CLI interpolation occurs. |
| A vulnerability was found in python-sql where unary operators do not escape non-Expression. |
