This week's cybersecurity roundup covers the DirtyClone Linux kernel vulnerability, AI focused malware techniques, Turla's STOCKSTAY backdoor, major malware disruptions, OpenAI updates, and emerging cyber threats.
A newly disclosed Linux kernel vulnerability tracked as CVE 2026 46331 allows local users to gain root privileges by poisoning cached binaries in memory, affecting multiple Linux distributions.
Google owned Mandiant reveals attackers exploited Cisco Catalyst SD WAN vulnerability CVE 2026 20245 as a zero day, using anti forensic techniques to gain root access and conceal malicious activity.
CISA has warned that attackers are actively exploiting the critical Lantronix EDS5000 vulnerability CVE 2025 67038 while researchers also report ongoing attacks targeting OpenWRT LuCI devices and UniFi OS flaws.
A security flaw in Anthropic’s Claude Code GitHub Action allowed attackers to potentially hijack public repositories through a single malicious GitHub issue.
Microsoft has disclosed active exploitation of CVE 2026 42897 affecting on premises Exchange Server deployments, allowing spoofing and JavaScript execution through crafted emails in Outlook Web Access.
Instructure confirms an agreement with ShinyHunters following a Canvas breach involving 3.65TB of stolen data impacting nearly 9000 institutions, with threat actors leveraging a vulnerability to exfiltrate sensitive education records.
Palo Alto Networks reports active exploitation of PAN OS CVE-2026-0300 allowing root level RCE, with espionage activity linked to suspected state sponsored threat cluster CL STA 1132.
Dirty Frag Linux kernel vulnerability enables local privilege escalation to root across major distributions including Ubuntu, RHEL, Fedora, and CentOS, with active exploitation risk.
Flowise AI platform suffers a critical CVSS 10.0 code injection vulnerability, exposing over 12,000 instances to remote code execution and full system compromise.
Microsoft identifies active attacks on on-premises SharePoint Server exploiting CVE-2025-53770, a variant of CVE-2025-49706, and shares mitigation guidance while working on a patch.