| CVE |
Vendors |
Products |
Updated |
CVSS v3.1 |
| Typemill is a flat-file, Markdown-based CMS designed for informational documentation websites. A reflected Cross-Site Scripting (XSS) exists in the login error view template `login.twig` of versions 2.19.1 and below. The `username` value can be echoed back without proper contextual encoding when authentication fails. An attacker can execute script in the login page context. This issue has been fixed in version 2.19.2. |
| Discourse is an open source discussion platform. A vulnerability present in versions prior to 3.5.4, 2025.11.2, 2025.12.1, and 2026.1.0 affects anyone who uses S3 for uploads. While scripts may be executed, they will only be run in the context of the S3/CDN domain, with no site credentials. Versions 3.5.4, 2025.11.2, 2025.12.1, and 2026.1.0 fix the issue. As a workaround, disallow html or xml files for uploads in authorized_extensions. For existing html xml uploads, site owners can consider deleting them. |
| Envoy is a cloud-native high-performance edge/middle/service proxy. Envoy will crash when the http async client is handling `sendLocalReply` under some circumstance, e.g., websocket upgrade, and requests mirroring. The http async client will crash during the `sendLocalReply()` in http async client, one reason is http async client is duplicating the status code, another one is the destroy of router is called at the destructor of the async stream, while the stream is deferred deleted at first. There will be problems that the stream decoder is destroyed but its reference is called in `router.onDestroy()`, causing segment fault. This will impact ext_authz if the `upgrade` and `connection` header are allowed, and request mirrorring. This issue has been addressed in versions 1.31.2, 1.30.6, 1.29.9, and 1.28.7. Users are advised to upgrade. There are no known workarounds for this vulnerability. |
| Software installed and run as a non-privileged user may conduct improper GPU system calls to subvert GPU HW to write to arbitrary physical memory pages.
Under certain circumstances this exploit could be used to corrupt data pages not allocated by the GPU driver but memory pages in use by the kernel and drivers running on the platform altering their behaviour.
This attack can lead the GPU to perform write operations on restricted internal GPU buffers that can lead to a second order affect of corrupted arbitrary physical memory. |
| Invalid memory access in Sentencepiece versions less than 0.2.1 when using a vulnerable model file, which is not created in the normal training procedure. |
| A malicious actor with access to the adjacent network could overflow the UniFi Protect Application (Version 6.1.79 and earlier) discovery protocol causing it to restart.
Affected Products:
UniFi Protect Application (Version 6.1.79 and earlier).
Mitigation:
Update your UniFi Protect Application to Version 6.2.72 or later. |
| 5ire is a cross-platform desktop artificial intelligence assistant and model context protocol client. Prior to version 0.15.3, an unsafe HTML rendering permits untrusted HTML (including on* event attributes) to execute in the renderer context. An attacker can inject an `<img onerror=...>` payload to run arbitrary JavaScript in the renderer, which can call exposed bridge APIs such as `window.bridge.mcpServersManager.createServer`. This enables unauthorized creation of MCP servers and lead to remote command execution. Version 0.15.3 fixes the issue. |
| A stored cross-site scripting (XSS) vulnerability exists in the Altium Support Center AddComment endpoint due to missing server-side input sanitization. Although the client interface applies HTML escaping, the backend accepts and stores arbitrary HTML and JavaScript supplied via modified POST requests.
The injected content is rendered verbatim when support cases are viewed by other users, including support staff with elevated privileges, allowing execution of arbitrary JavaScript in the victim’s browser context. |
| When Conn.Handshake fails during ALPN negotiation the error contains attacker controlled information (the ALPN protocols sent by the client) which is not escaped. |
| Shenzhen Tenda W30E V2 firmware versions up to and including V16.01.0.19(5037) fail to include the X-Content-Type-Options: nosniff response header on web management interfaces. As a result, browsers that perform MIME sniffing may incorrectly interpret attacker-influenced responses as executable script. |
| A weakness has been identified in UGREEN DH2100+ up to 5.3.0.251125. This affects the function handler_file_backup_create of the file /v1/file/backup/create of the component nas_svr. Executing a manipulation of the argument path can lead to buffer overflow. The attack can be executed remotely. The exploit has been made available to the public and could be used for attacks. It is recommended to upgrade the affected component. |
| Improper Restriction of Operations within the Bounds of a Memory Buffer vulnerability in anyrtcIO-Community anyRTC-RTMP-OpenSource (third_party/faad2-2.7/libfaad modules). This vulnerability is associated with program files bits.C, syntax.C.
This issue affects anyRTC-RTMP-OpenSource: before 1.0. |
| Improper Restriction of Operations within the Bounds of a Memory Buffer vulnerability in CardboardPowered cardboard (src/main/java/org/cardboardpowered/impl/world modules). This vulnerability is associated with program files WorldImpl.Java.
This issue affects cardboard: before 1.21.4. |
| Improper Restriction of Operations within the Bounds of a Memory Buffer vulnerability in GaijinEntertainment DagorEngine (prog/3rdPartyLibs/miniupnpc modules). This vulnerability is associated with program files upnpreplyparse.C.
This issue affects DagorEngine: through dagor_2025_01_15. |
| The issue was addressed with improved bounds checks. This issue is fixed in iOS 18.1 and iPadOS 18.1. An app may be able to corrupt coprocessor memory. |
| The exos 9300 application can be used to configure Access Managers (e.g. 92xx, 9230 and 9290). The configuration is done in a graphical user interface on the dormakaba exos server. As soon as the save button is clicked in exos 9300, the whole configuration is sent to the selected Access Manager via SOAP. The SOAP request is sent without any prior authentication or authorization by default. Though authentication and authorization can be configured using IPsec for 92xx-K5 devices and mTLS for 92xx-K7 devices, it is not enabled by default and must therefore be activated with additional steps.
This insecure default allows an attacker with network level access to completely control the whole environment. An attacker is for example easily able to conduct the following tasks without prior authentication:
- Re-configure Access Managers (e.g. remove alarming system requirements)
- Freely re-configure the inputs and outputs
- Open all connected doors permanently
- Open all doors for a defined time interval
- Change the admin password
- and many more
Network level access can be gained due to an insufficient network segmentation as well as missing LAN firewalls. Devices with an insecure configuration have been identified to be directly exposed to the internet. |
| On the exos 9300 server, a SOAP API is reachable on port 8002. This API does not require any authentication prior to sending requests. Therefore, network access to the exos server allows e.g. the creation of arbitrary access log events as well as querying the 2FA PINs associated with the enrolled chip cards. |
| An API endpoint allows arbitrary log entries to be created via POST request. Without sufficient validation of the input data, an attacker can create manipulated log entries and thus falsify or dilute logs, for example. |
| Gakido is a Python HTTP client focused on browser impersonation and anti-bot evasion. A vulnerability was discovered in Gakido prior to version 0.1.1 that allowed HTTP header injection through CRLF (Carriage Return Line Feed) sequences in user-supplied header values and names. When making HTTP requests with user-controlled header values containing `\r\n` (CRLF), `\n` (LF), or `\x00` (null byte) characters, an attacker could inject arbitrary HTTP headers into the request. The fix in version 0.1.1 adds a `_sanitize_header()` function that strips `\r`, `\n`, and `\x00` characters from both header names and values before they are included in HTTP requests. |
| A flaw was found in Keycloak's SAML brokering functionality. When Keycloak is configured as a client in a Security Assertion Markup Language (SAML) setup, it fails to validate the `NotOnOrAfter` timestamp within the `SubjectConfirmationData`. This allows an attacker to delay the expiration of SAML responses, potentially extending the time a response is considered valid and leading to unexpected session durations or resource consumption. |
