CVE-2025-8088, CVE-2026-111560, CVE-2026-116606, CVE-2026-119977, CVE-2026-58868, CVE-2026-58876, CVE-2026-58882, CVE-2026-58936, CVE-2026-98432
Get tomorrow's brief in your inbox
Today: CISA confirmed exploitation of WordPress Core CVE-2026-119977, with a November 6 remediation deadline for Federal Civilian Executive Branch agencies. Malicious cPGuard updates installed root backdoors on Linux hosting servers; updating removes known artifacts, but exposed credentials still need rotation. Prioritize Barracuda administrative interfaces and critical Linux kernel fixes alongside these incident-response tasks.
WordPress Core Upgrade Race Exploited in the Wild (CVE-2026-119977)
Summary: CISA added CVE-2026-119977, a high-severity WordPress Core race condition rated CVSS 8.1, to its Known Exploited Vulnerabilities catalog. Unauthenticated attackers can execute server-side code during the brief full-package Core upgrade window by targeting predictable, web-accessible staging paths. Enrichment lists EPSS at 0.006, approximately 0.6%, but confirmed exploitation takes precedence over that low prediction score.
setup-config.php paths around upgrades and investigate associated file changes or unexpected execution. The November 6, 2026 deadline applies to FCEB agencies, not automatically to private businesses.Malicious cPGuard Updates Installed Root Backdoors
Summary: OpsShield says attackers compromised cPGuard's update infrastructure on November 2 and distributed malicious versions 5.88.12 through 5.88.19 to Linux hosting servers. The script executed as root, installed a hidden shell, and collected password hashes, database configuration, SSH-related files, and shell history. OpsShield estimated that attackers retrieved 32 collected files associated with separate customer servers; that is not a complete count of potentially affected systems.
cpgcli update --start, then confirm version 5.88.25 or later with cpgcli --version. Search /opt/cpguard/logs/install_log for affected versions, but treat an installation entry as evidence of possible exposure, not proof of execution or exfiltration; incomplete logs do not clear a server.NCSC Warns AI-Powered Malware Is Likely in Critical-Infrastructure Ransomware by 2028
Summary: The UK National Cyber Security Centre assesses that AI-powered malware is "highly likely" to feature in a ransomware attack against British critical national infrastructure between now and 2028. This is a forecast, not a confirmed deployment timeline; the report also describes current abuse of AI security tools, including ARTEX, and considers more autonomous attacks as a future scenario.
Chainstack Discloses Production Access and Customer Password-Hash Exposure
Summary: Chainstack says an attacker entered its development environment on October 26, escalated into parts of production, and accessed a database containing customer names, organization details, email addresses, and password hashes. The company detected and cut off access on October 31; it is treating node credentials as exposed despite finding no evidence they were read. WorkOS separately disclosed an incident affecting business-relationship systems, not production or customer platform data, amid coverage linking these incidents to ARTEX-assisted activity.
mSecure Warns Customers to Protect Stored Secrets After Suspicious Activity
Summary: mSeven Software notified current and former mSecure customers on November 3 that it had detected suspicious activity and engaged external investigators. Its notice recommends protecting stored confidential information, potentially through password changes, 2FA resets, and passkey removal, but does not establish that vault contents or encryption keys were accessed. The scope of the recommended authentication changes remains unclear in the company's email.
Stolen Cloud Credentials Drive LLMjacking Bills Above $100,000 a Day
Summary: Google Threat Intelligence Group and Mandiant describe attackers stealing enterprise access to model services through exposed API keys, service-account credentials, personal access tokens, and active sessions. They observed individual incidents exceeding $100,000 in fraudulent daily consumption, with some organizations accumulating $1 million before access was revoked. One campaign tricked users into submitting active access tokens, bypassing the protection offered by passwords and MFA.
Haiou Bundle Persistently Alters Coding-Agent Instructions
Summary: Check Point Research found that the Haiou jailbreak bundle installs security-related skills and modifies global coding-agent instruction files to insert advertisements into later responses. Researchers observed the bundle in a threat actor's toolkit alongside Cobalt Strike and other offensive tools, but explicitly did not establish that its author was involved in cybercrime. The security issue is persistent modification of trusted instructions, not proof that every advertised gateway is malicious.
Gorbag Steals Browser, VPN, and Document Data (CVE-2025-8088)
Summary: ESET describes Russia-aligned Gorbag targeting Ukrainian government, justice, local-administration, and defense organizations through spear-phishing and information stealers. Earlier attacks exploited patched WinRAR vulnerability CVE-2025-8088 to deliver GIFTEDCROOK; its BeautyStealer evolution collects browser authentication data, OpenVPN configurations, Office documents, KeePass databases, and RDP configuration files. Enrichment identifies the CVE as ransomware-linked KEV, with EPSS 0.940, approximately 94%, and a historical September 2, 2025 remediation deadline.
Fake Windows Installers Enroll Devices in Proxy and Compute Networks
Summary: Microsoft identified trojanized Inno Setup-style installers promoted as cracked software, games, activators, plugins, and utilities. Confirmed outcomes include third-party traffic proxying, distributed-computing enrollment, and utility-style software installations without informed consent; WorldProxyPool also creates recurring SYSTEM scheduled tasks. Legitimate uses of individual downstream products do not make deceptive installation benign.
%TEMP%\is-*.tmp\ launching eld*.exe or set_*.exe, then correlate with new SYSTEM tasks and unexpected network or compute workloads. These patterns are investigation leads, not standalone proof of compromise.ILC Resolves Cybersecurity Contract Allegations for $13 Million
Summary: ILC Dover and ILC Astrospace agreed to pay $13 million to resolve government False Claims Act allegations involving cybersecurity requirements in federal contracts between 2018 and 2024. The government alleged inadequate safeguards for cloud-stored sensitive data and deficient security plans; parent company Ingersoll Rand says it addressed inherited issues and settled without admitting liability.
Linux IPv4 TCP Race Can Allow Unauthenticated RCE (CVE-2026-98432)
Summary: A use-after-free in Linux IPv4 TCP handling can allow remote, unauthenticated code execution and is rated Critical, CVSS 9.8. Red Hat released fixes for RHEL 8, 9, and 10; the article places Ubuntu generic-kernel fixes in its opt-in testing channel and distinguishes other kernel variants. Enrichment gives EPSS 0.008, approximately 0.8%; the article identifies no reported exploitation.
Barracuda Authorization Bypass and Command Injection Can Chain to Root
Summary: CVE-2026-116606, CVSS 9.3 Critical, affects Email Security Gateway, while CVE-2026-111560, CVSS 8.6 High, affects that product and five other Barracuda appliance families. Barracuda says chaining the flaws can bypass authorization and obtain root access, but requires access to an affected administrative interface. Enrichment lists EPSS at approximately 0.7% and 0.5%, respectively.
Android November Updates Address Zero-Interaction RCE (CVE-2026-58868)
Summary: Android's November bulletin addresses 216 vulnerabilities, including Critical Framework flaw CVE-2026-58868, which allows remote code execution without user interaction on affected Android 14 through 17 devices. Critical System vulnerabilities include CVE-2026-58876 and CVE-2026-58936 for RCE, plus CVE-2026-58882 for remote denial of service. Enrichment lists EPSS around 0.2% for each; Google's bulletin identified no exploitation as of November 4.
Trusted update channels, developer tools, and cloud identities recur as intrusion paths across these stories. Patching closes vulnerable code, but stolen sessions, SSH keys, and API credentials require separate revocation. Prioritize confirmed exploitation and reachable administrative surfaces over severity or EPSS alone.