| CVE |
Vendors |
Products |
Updated |
CVSS v3.1 |
| On affected platforms running Arista EOS with VPLS-BGP enabled, an attacker who has established a BGP session with the PE routers can send a specially crafted BGP packet to cause the BGP agent to restart unexpectedly. |
| This vulnerability affects customers using Arista Pathfinder for WAN path selection with Border Gateway Protocol-Link State (BGP-LS) configured for Dynamic Path Selection (DPS). On affected platforms running Arista EOS with BGP-LS configured for DPS, an attacker who has established an authenticated BGP-LS session with the device can send a specially crafted BGP packet to cause incorrect WAN path selection and disrupt traffic forwarding. |
| On affected Arista EOS platforms running Border Gateway Protocol with Link State ( BGP-LS) and configured with the Dynamic Path Selection (DPS) consumer role, an attacker who has established an authenticated BGP-LS session with the device can send a crafted message to cause the BGP agent to restart unexpectedly. An Arista Pathfinder deployment is not required. |
| On affected platforms running Arista EOS with 802.1X with Dynamic Host Configuration Protocol (DHCP) device profiling enabled, a network-adjacent attacker can send a crafted DHCP packet to force the device to record a spoofed device profile. This may potentially allow the authenticated supplicant to acquire unauthorized access to a network segment with elevated privileges. |
| In JetBrains IntelliJ IDEA before 2026.3 untrusted projects could trigger changes to GPG agent configuration |
| On affected platforms running Arista EOS with IEEE 802.1X authentication and MACsec using dynamically derived keys, sensitive authentication information may be recorded in diagnostic data during a Smart Switch Upgrade (SSU).
An attacker with administrative access on the device who obtains the affected diagnostic data could recover sensitive authentication information. This could potentially compromise the confidentiality and integrity of traffic protected by MACsec on the affected link.
The vulnerability is only exposed during and after an SSU reload where an active 802.1X session used for MACsec is present. |
| On affected platforms running Arista EOS with automatic certificate management configured and automatic Certificate Authority (CA) certificate retrieval enabled, a malicious packet may cause the AutoCertMgmt process to be terminated unexpectedly. Managed X.509 certificates will not be renewed automatically and may eventually expire, potentially disrupting TLS-dependent services. |
| On affected platforms running Arista EOS with IPsec AutoVPN Feature configured with dynamic paths, the configured IPsec replay window may not be applied as expected. As a result, IPsec anti-replay protection may be silently inactive even when it has been configured via an SA policy, contrary to the operator's expectation. This issue does not allow decryption of the affected IPsec traffic, but it may allow replay of previously valid protected traffic. |
| Arista C-260 and C-250 access points ship with fleet-wide static cryptographic keys used for firmware image verification and SSLP authentication. An attacker who is able to access these keys would be able to create modified firmware that will pass verification checks. This vulnerability is specific to the C-260 and C-250 models; other access point models enforce digitally signed upgrade images with secure boot and are not affected by this issue. |
| In JetBrains IntelliJ IDEA before 2026.3 arbitrary file write was possible via symlinks when extracting plugin archives |
| Winter CMS is a content management system built on the Laravel PHP framework. From 1.2.13 until 1.2.14, the Backend\Controllers\MyAccount controller inherits the FormController actions create, update, and preview while declaring an empty $requiredPermissions array, and formFindModelObject() resolves a caller-supplied record identifier through an unscoped Backend\Models\User query. Any authenticated backend user can request another sequential user identifier through /backend/backend/myaccount/preview/{other_user_id} or the inherited update action and read profile data including name, login, email, avatar, role, group membership, superuser status, and throttle state. Password material is not exposed, and existing model authorization guards block cross-user writes. This issue is fixed in version 1.2.14. |
| In JetBrains IntelliJ IDEA before 2026.3 code execution was possible via symlinks in an imported settings archive |
| In JetBrains IntelliJ IDEA before 2026.3 hTML injection was possible in Remote Development editor tabs |
| In JetBrains IntelliJ IDEA before 2026.3 code execution via the document import was possible in untrusted projects |
| An issue in MongoDB server’s aggregation framework allows an authenticated user with read privileges to cause a denial of service by sending specially crafted aggregate commands using an internal expression. Despite its internal naming, this expression is registered as accessible to any authenticated client. By supplying crafted input parameters across multiple concurrent connections, each evaluation triggers unbounded memory allocations with no per-expression memory limits enforced. This can exhaust available server memory and cause the mongod process to be terminated by the operating system’s out-of-memory killer. |
| In JetBrains IntelliJ IDEA before 2026.3,
2026.2.4 stored XSS was possible on the built-in web server via Markdown preview |
| In JetBrains IntelliJ IDEA before 2026.3 kotlin scripts could trigger dependency downloads from arbitrary repositories in untrusted projects |
| In JetBrains IntelliJ IDEA before 2026.3,
2026.2.4 command execution via a hidden run control in the Markdown preview was possible |
| In JetBrains IntelliJ IDEA before 2026.3,
2026.2.4 sSRF was possible via annotations.xml files during indexing |
| In JetBrains IntelliJ IDEA before 2026.3 improper checks could enable disabled plugins via external dependency settings |
