🚨 CVE-2026-50099
During WiFi association, Naxclow device firmware prints the host network’s SSID, PSK, and negotiated WPA keys in cleartext to an exposed UART console on production hardware. The UART pads are labeled, run with default serial settings, and drop to an interactive RT-Thread shell that permits arbitrary memory reads, enabling full firmware extraction. An attacker with brief physical access, common for outdoor-mounted devices, can therefore recover WiFi credentials and bootstrap firmware-side attacks.

🎖@cveNotify

🚨 CVE-2026-50101
Naxclow devices use a server-side, per-device relay credential that never rotates and is re-issued to the device on each boot. Because this credential remains valid indefinitely and cannot be reset or revoked by the legitimate owner, any party that obtains it through any exposure path can maintain persistent access to the device’s relay channel. This enables long-term impersonation or interception, even after factory resets or re-onboarding.

🎖@cveNotify

🚨 CVE-2026-50108
The Naxclow platform API that returns device relay registration details exposes a persistent credential without verifying that the requester is the legitimate device or owner. An actor able to present a platform-valid request signature can retrieve credentials for arbitrary devices and register on the relay as that device, enabling interception and disruption of its communications.

🎖@cveNotify

🚨 CVE-2026-50244
The Naxclow platform exposes a registration endpoint that accepts signed requests containing a batch prefix and an arbitrary caller-supplied account identifier, without validating any ownership relationship. Each call mints a new sequential device identifier and returns the current high-water counter value for the batch, allowing callers to measure and enumerate the active device space. The endpoint’s behavior enables precise fleet enumeration.

🎖@cveNotify

🚨 CVE-2026-42604
Actual is a local-first personal finance tool. The `POST /openid/config` endpoint in Actual Budget's sync-server versions <= 26.4.0 exposes the full OpenID Connect configuration—including the OAuth2 `client_secret`—to any caller who knows the bootstrap password. The endpoint also lacks authentication and rate limiting, making the bootstrap password brute-forceable. Version 26.5.0 fixes the issue.

🎖@cveNotify

🚨 CVE-2026-42890
Actual is an open-source personal finance application. In the macOS desktop application version 25.x (built on Electron 39.2.7), the ELECTRON_RUN_AS_NODE fuse is not disabled, allowing an attacker who can place a file on disk or control command-line arguments to invoke the signed Actual.app binary with the ELECTRON_RUN_AS_NODE=1 environment variable set. This converts the application into a Node.js REPL capable of executing arbitrary code that inherits the application's entitlements and code signature, bypassing macOS Gatekeeper review. Version 26.5.0 patches the issue.

🎖@cveNotify

🚨 CVE-2026-43872
Actual is an open-source personal finance application. Prior to version 26.5.0, several endpoints are affected by a path traversal vulnerability. Version 26.5.0 fixes the issue.

🎖@cveNotify

🚨 CVE-2026-4870
IBM Qiskit SDK 0.43.0 through 2.5.0 could allow an attacker to trigger a segmentation fault leading to a denial of service due to uncontrolled recursion in the parser.

🎖@cveNotify

🚨 CVE-2026-54057
Kitty is a cross-platform GPU based terminal. In versions prior to 0.47.3, kitty's OSC 21 (color-control) query reply reflects attacker-controlled bytes, including newlines, into the shell's input without sanitization. Version 0.47.3 fixes the issue.

🎖@cveNotify

🚨 CVE-2026-34195
Software installed and run as a non-privileged user may conduct intentional GPU sparse memory API calls to cause out of bounds write in the kernel.



The product incorrectly indexes internal state when performing sparse allocation remapping.

🎖@cveNotify

🚨 CVE-2026-41155
An attacker could cooperatively pass data from one secure GPU process to another secure GPU process through shared secure memory allocations in the kernel module. Additionally, an attacker could disrupt the operation of another secure GPU process leading to image corruption / GPU hardware recovery.



Sharing secure memory allocations among various GPU secure processes allows an attacker to corrupt shared resource affecting other users.

🎖@cveNotify

🚨 CVE-2026-41157
A web page that contains unusual WebGPU content loaded into the GPU GLES render process and can trigger an out-of-bound write in the GPU user-space driver, leading to memory corruption and possible browser/GPU process crash.



The software computes a required memory size from untrusted input, but integer overflow can produce a value smaller than needed. Subsequent write operations may then occur past the intended memory boundary, corrupting adjacent memory and causing process instability or termination.

🎖@cveNotify

🚨 CVE-2026-41158
Software installed and run as a non-privileged user may conduct GPU system calls to write to arbitrary freed physical pages.



Physical memory allocated and freed, without the deferred free mechanism can lead to those resources being used for read/write by the GPU after the kernel module has freed the resource.

🎖@cveNotify

🚨 CVE-2026-11769
We have released version 5.24.0 of the Grafana Operator. This patch includes a CRITICAL severity security fix for a path traversal/privilege escalation vulnerability in the Grafana Operator.



### Summary



The Grafana Operator supports loading dashboards & library panels using the jsonnet data templating language. The jsonnet expression is evaluated in the context of the operator manager pod.



### Impact



It is possible for a malicious user who can create Dashboard or LibraryPanel resources for a Grafana instance to obtain the Kubernetes service account token of the Grafana Operator manager.



### Affected versions



All Grafana Operator versions <= 5.23



### Solutions and mitigations



All installations should be upgraded as soon as possible.



As a workaround, the following ValidatingAdmissionPolicy prevent the creation or modification of jsonnet based resources:



apiVersion: admissionregistration.k8s.io/v1



kind: ValidatingAdmissionPolicy



metadata:

name: "prevent-jsonnet-dashboards"


spec:

failurePolicy: Fail
matchConstraints:
resourceRules:
- apiGroups: ["grafana.integreatly.org"]
apiVersions: ["v1beta1"]
operations: ["CREATE", "UPDATE"]
resources: ["grafanadashboards", "grafanalibrarypanels"]
validations:
- expression: "!has(object.spec.jsonnetLib)"


---



apiVersion: admissionregistration.k8s.io/v1



kind: ValidatingAdmissionPolicyBinding



metadata:

name: "prevent-jsonnet-dashboards-clusterwide"


spec:

policyName: "prevent-jsonnet-dashboards"
validationActions: [Deny]


### Acknowledgement



We would like to thank Artem Cherezov for responsibly disclosing the vulnerability.

🎖@cveNotify

🚨 CVE-2026-11624
The Model Context Protocol has a security warning advising servers to validate the "Origin" header on all incoming connections to prevent DNS rebinding attacks. Prior to the v0.25.0 release, users had no way to validate the origin's host. In v0.25.0, a new "--allowed-hosts" flag was introduced alongside the existing "--allowed-origins" flag, enabling users to specify permitted hosts at server startup. Both flags default to "*", allowing users to implement strict access controls as needed without breaking existing setups. If either flag is set to "*", the server will output a startup warning about potential vulnerabilities. Documentation has also been updated to highlight these security considerations.

🎖@cveNotify

🚨 CVE-2026-12183
Nefteprodukttekhnika BUK TS-G Gas Station Automation System 2.9.1 through 2.10.2 on Linux contains an Improper Authentication vulnerability (CWE-287) in the system configuration module. The /php/ajax-login.php endpoint returns userid=1 (administrator) in response to any HTTP POST request that supplies arbitrary credentials (e.g., action=dologin&login=<any_value>&pwd=<any_value>), and subsequent privileged endpoints under /php/ajax-main.php and /modules/* do not validate a server-side session. A remote unauthenticated attacker can invoke any administrative action exposed by the configuration module, including reading and modifying user rules, fuel tank gauges, fuel dispensers, relays, cash registers, bank terminals, fuel cards, price and customer displays, cash collection, and pricing rules.

🎖@cveNotify

🚨 CVE-2026-11526
GD versions before 2.86 for Perl allow OS command injection and file overwrite via a 2-arg open() of filename arguments in _make_filehandle.

GD::Image::_make_filehandle opens a filename argument with Perl's 2-arg open(), so a filename that begins or ends with a pipe ("| cmd", "cmd |") or begins with a redirect ("> path", ">> path") is run as a command or redirect rather than opened as a file. _make_filehandle is the single open path behind every filename-accepting constructor (new, newFromPng, newFromJpeg, and the rest); the in-memory *Data variants do not open a path and are unaffected.

Any caller that forwards untrusted input to one of these constructors as a pathname can run an arbitrary command or truncate a file under the process UID.

🎖@cveNotify

🚨 CVE-2026-11527
Config::IniFiles versions before 3.001000 for Perl allow OS command injection and file overwrite via a 2-arg open() of the -file argument in _make_filehandle.

Config::IniFiles::_make_filehandle opens a filename argument with Perl's 2-arg open(), so a filename that begins or ends with a pipe ("| cmd", "cmd |") or begins with a redirect ("> path", ">> path") is run as a command or redirect rather than opened as a file. The helper is the open path behind the documented -file argument: new(-file => $thing) reaches it through ReadConfig. An in-memory scalar reference (-file => \$text) does not open a path and is unaffected.

Any caller that forwards untrusted input to the -file argument can run an arbitrary command or truncate a file under the process UID.

🎖@cveNotify

🚨 CVE-2026-54410
nanoMODBUS through v1.23.0 contains an off-by-one buffer overflow in the recv_msg_header() function of the Modbus/TCP server that allows remote unauthenticated attackers to write one attacker-controlled byte past the end of the 260-byte receive buffer by sending a crafted MBAP frame whose Length field is set to 255. The overflow corrupts the adjacent buffer-index field of the nanoMODBUS state structure, resulting in denial of service through invalid memory accesses and, on bare-metal and RTOS targets without memory protection, one-byte information disclosure and writes to unintended register addresses on the Write Multiple Registers (FC16) handler path.

🎖@cveNotify

🚨 CVE-2026-54411
Linux-PAM through 1.7.2 contains an observable timing discrepancy (CWE-208) in the pam_userdb module's plaintext-password comparison path in modules/pam_userdb/pam_userdb.c that allows a local or network-adjacent attacker able to repeatedly drive authentication through a calling service to recover the plaintext password of a target account by measuring response-timing differences. The comparison uses strncmp() (or strncasecmp() when PAM_ICASE_ARG is set) preceded by a length-equality check, so the time to reject a candidate depends on the index of the first differing byte and on whether the candidate's length matches the stored password, leaking the password length and individual prefix bytes. The vulnerable path is reached when the administrator configures pam_userdb with crypt=none, with an unrecognized crypt method, or without a crypt= argument, causing the module to store and compare credentials in plaintext.

🎖@cveNotify

🚨 CVE-2026-54412
LiamBindle MQTT-C through version 1.1.6 contains a heap-based out-of-bounds read and integer underflow in the mqtt_unpack_publish_response() function in src/mqtt.c that allows a remote unauthenticated attacker controlling an MQTT broker - or able to inject MQTT traffic into an unencrypted session - to crash a subscribed MQTT-C client and potentially disclose adjacent heap memory by sending a single crafted PUBLISH packet. The function validates only that the fixed-header remaining_length is at least 4, then reads the 16-bit topic_name_size field from the broker-controlled packet and advances the parse pointer by that value without verifying that topic_name_size plus the surrounding overhead fits within remaining_length; it subsequently computes application_message_size as remaining_length - topic_name_size - 2 (QoS 0) or - 4 (QoS greater than 0) in unsigned arithmetic, producing an integer underflow that is then passed to memmove(). A PUBLISH packet with topic_name_size = 0xFFFF and remaining_length = 7 advances the parse pointer 65535 bytes past the receive buffer (out-of-bounds read) and causes an application_message_size near 2^32, crashing the process when the resulting memmove() is executed.

🎖@cveNotify