Vulnerability intelligence

CVE-2025-8045 is a "Use After Free" vulnerability affecting the Arm Ltd Valhall GPU Kernel Driver and Arm 5th Gen GPU Architecture Kernel Driver. This flaw allows a local non-privileged user process to execute improper GPU processing operations, thereby gaining access to memory that has already been freed. The vulnerability impacts driver versions r53p0 through r54p1 for both the Valhall GPU Kernel Driver and the Arm 5th Gen GPU Architecture Kernel Driver. Successful exploitation of this issue can result in the disclosure of kernel memory contents to an unprivileged process running on the device, requiring local access and no user interaction.

CVE-2026-18577 is an authentication bypass and account takeover vulnerability affecting N-able N-central versions through 2026.3.1. This flaw is a result of an incomplete patch for a previously identified vulnerability, CVE-2026-18556. Successful exploitation of CVE-2026-18577 allows remote attackers to gain administrative access to vulnerable N-central servers. Once administrative control is established, attackers can abuse the built-in "Take Control" feature to pivot into managed endpoints, deploy scripts, run tools, initiate remote-control sessions, and establish persistence within the compromised environment. The vulnerability impacts both hosted and on-premises deployments of N-able N-central.

CVE-2025-15039 describes a vulnerability found in the Conditional Authentication (Adaptive Authentication) script used across several WSO2 products, including API Control Plane, API Manager, Carbon Identity Application Authentication Framework, Identity Server, Identity Server as Key Manager, Open Banking AM, Open Banking IAM, Traffic Manager, and Universal Gateway. This flaw allows an attacker to bypass intermediate authentication challenges by exploiting how the script handles callbacks and re-execution of authentication steps. Successful exploitation of CVE-2025-15039 can lead to unauthorized access to a targeted user account. This vulnerability is exploitable under specific conditions: the application's login flow must include a particular secondary authenticator, the Conditional Authentication script needs to be configured with specific event callbacks and re-execute an authentication step, the targeted user must have one of the impacted authenticators enrolled, and the attacker must successfully complete any preceding authentication steps.

OS Command Injection Remote Code Execution Vulnerability in API in Progress ADC Products allows an un-authenticated attacker to execute arbitrary commands on the LoadMaster appliance by exploiting unsanitized input in multiple command endpoints

In JetBrains TeamCity before 2026.1.3, 2025.11.7 unauthenticated remote code execution was possible via the agent polling protocol

CVE-2026-9198 is an unauthenticated remote code execution vulnerability affecting IBM Langflow OSS versions 1.0.0 through 1.10.0. This flaw allows attackers to gain full control over default Langflow deployments by exploiting a chain of two API endpoints. Specifically, the vulnerability leverages the `/api/v1/auto_login` endpoint, which issues SUPERUSER tokens to any network caller without requiring authentication. Subsequently, attackers can utilize the `/api/v1/validate/code` endpoint to execute arbitrary Python code via `exec()`, leading to full remote code execution. This issue is classified under CWE-94, indicating improper control of code generation. The vulnerability has been addressed in Langflow version 1.10.1.