A vulnerability has been identified in the RDSEED instruction on AMD Zen 5 (Turin) processors. The 16- and 32-bit versions of the RDSEED instruction can silently fail under certain system load conditions, potentially compromising applications that rely on these instructions for cryptographic random number generation. The 64-bit version of RDSEED is not affected and remains safe. AMD is currently investigating this issue. The 64-bit Linux kernel uses the safe 64-bit RDSEED instruction for /dev/[u]random, so those random numbers are not impacted. Applications that synthesize random numbers using the 16- or 32-bit RDSEED instructions should be aware of this insecurity.
A vulnerability in the CPP RCTX instruction on select Arm processors used in Google Cloud Compute Engine Arm VMs (C4A, A4X) allows an attacker with privileged guest kernel access to inhibit TLB invalidations. This can potentially enable unauthorized reading of sensitive data. Google has already applied mitigations to the affected Arm server fleet, and no customer action is required.
A vulnerability identified as CVE-2026-23268, known as CrackArmor, exists in the Linux kernel's AppArmor component on Container-Optimized OS (COS) nodes. This flaw allows unprivileged users to bypass kernel protections, escalate privileges to root, and break local container isolation. The vulnerability is rated high severity. Google recommends upgrading COS nodes to version cos-125-19216-220-57, which contains the fix. For managed services using COS, users should follow the respective upgrade instructions. Fixes for earlier COS milestones 117 and 121 are in progress.
A vulnerability in AMD firmware could allow a malicious hypervisor to alter BIOS settings and Memory Mapped I/O (MMIO) routing configurations, compromising the confidentiality and integrity of Confidential VMs using AMD SEV-SNP guests. Google has implemented mitigations for this issue in Confidential VM instances with AMD SEV-SNP. No customer action is required. The vulnerability is documented as CVE-2025-54510 and is rated medium severity.
A vulnerability in AMD firmware allows a malicious hypervisor to execute arbitrary code on the AMD Secure Processor (ASP), escalating Memory Mapped I/O (MMIO) read and write permissions. This compromises the confidentiality and integrity of SEV-SNP guests. Google has applied mitigations to Confidential VM instances with AMD SEV-SNP, and no customer action is required.
A hardware-level vulnerability has been identified in AMD Zen 2 microarchitecture processors, including EPYC and Ryzen series, involving potential corruption within the micro-operation (OP) cache. Under specific conditions, this flaw (CVE-2025-54518) could allow security boundary bypasses or unauthorized data access. Google has deployed fixes across its infrastructure to mitigate this issue.
CVE-2025-10263 is an architectural vulnerability in certain Arm cores (Neoverse V1, V2, N1) used in Google Cloud Arm-based instances. It allows an attacker at a lower exception level to write to memory owned by a higher exception level, enabling privilege escalation. This flaw does not affect memory reads. Google has secured the infrastructure against VM-to-VM and VM-to-hypervisor attacks, but guest OS-level patches from OS vendors are required to mitigate threats within the guest environment. Mitigations for Container-Optimized OS are in progress.
A vulnerability (CVE-2026-6726, also known as TCGVRT0010) has been identified in the Trusted Computing Group's TPM 2.0 reference implementation code, affecting all published revisions of the code (v1.16, v1.38, v1.59, v1.83, v184). This flaw could allow a privileged local attacker to obtain credentials from a TPM-aware Certificate Authority for a falsified TPM key, enabling the creation of fraudulent TPM 2.0 attestations. Google will proactively update affected systems during standard maintenance windows, and no customer action is required.
An incorrect authorization vulnerability (CVE-2026-85299) was discovered in Google Cloud Compute Engine versions prior to August 31, 2026. This flaw allowed users to exfiltrate disk data without having the required IAM permissions. The vulnerability was patched on August 31, 2026, and no action is required from users.
CVE-2026-116506 is a critical command injection vulnerability in the TOTOLINK A950RG router, specifically affecting version 4.1.2cu.5268_B20221104. The flaw exists in the apmib_set function within the firewall.so component. This vulnerability allows remote attackers to execute arbitrary commands on the device. Public exploit code has been disclosed, but no known active exploitation in the wild has been reported.
A resource exhaustion vulnerability exists in seroval versions from 1.4.1 up to but not including 1.6.8. The fromJSON and fromCrossJSON functions do not properly validate that an array node's element list is an actual array, allowing an attacker to supply a crafted length that controls memory allocation and processing iterations. This can lead to denial of service by exhausting process memory or blocking the thread. The issue is a residual bypass of a previous fix and is resolved in version 1.6.8.
A cross-site scripting (XSS) vulnerability exists in the lxsmnsyc seroval library versions 1.6.0 through 1.6.10. The issue arises because the JavaScript-output APIs derive a typed array's constructor name from a property that can be attacker-controlled if untrusted properties are copied onto a typed array and then serialized. This allows execution of attacker-controlled JavaScript in the context of the receiving page. The vulnerability is fixed in version 1.6.11.
A cross-site scripting (XSS) vulnerability exists in the lxsmnsyc seroval library's JavaScript-output APIs prior to version 1.0.0 and from 1.5.5 up to but not including 1.6.10. The issue arises when regular expressions containing a less-than character are serialized without proper escaping for embedding in HTML script elements. This can allow an attacker to inject malicious scripts if untrusted input is serialized and embedded in a web page. The vulnerability is fixed in version 1.6.10.
CVE-2026-116505 is a critical command injection vulnerability in TOTOLINK A950RG version 4.1.2cu.5268_B20221104. It affects the getDeviceInfo function related to the /var/lib/misc/udhcpd.leases file in the device-information handler. The flaw allows remote attackers to inject commands by manipulating the MAC address argument. Public exploit code is available, increasing the risk of exploitation.
A vulnerability in Apple iOS and iPadOS prior to version 27.2 could allow an app to access sensitive user data due to an authorization issue. This flaw was addressed by improved state management in iOS 27.2 and iPadOS 27.2. The issue also affects macOS Golden Gate and visionOS 27.2, where it has been fixed. No public exploit or in-the-wild exploitation is currently known.
A heap corruption vulnerability exists in Apple iOS and iPadOS due to improper bounds checking when processing maliciously crafted images. This buffer overflow issue has been addressed in iOS 26.7.2, iPadOS 26.7.2, iOS 27, iPadOS 27, and macOS versions including Golden Gate 27, Sequoia 15.8, Tahoe 26.7.2, and visionOS 27. Processing a specially crafted image could lead to heap corruption, potentially impacting system stability or security.
An out-of-bounds read was addressed with improved bounds checking. This issue is fixed in macOS Golden Gate 27.2. An app may be able to cause unexpected system termination.
A logic issue was addressed with improved checks. This issue is fixed in macOS Golden Gate 27.2, macOS Sequoia 15.8.2, macOS Tahoe 26.7.2. A local attacker may be able to elevate their privileges.
An out-of-bounds read was addressed with improved bounds checking. This issue is fixed in macOS Golden Gate 27.2, macOS Sequoia 15.8.2, macOS Tahoe 26.7.2. An app may be able to cause unexpected system termination.
A buffer overflow vulnerability in Apple iOS and iPadOS could allow memory corruption when processing a maliciously crafted audio file. This issue affects versions prior to iOS and iPadOS 26.7.2 and 27.2. Apple has fixed the vulnerability in iOS 26.7.2, iPadOS 26.7.2, iOS 27.2, and iPadOS 27.2. No public exploit or in-the-wild exploitation has been reported.