// Articles
389 articles — page 6 of 17
Ubiquiti UniFi OS Security Bulletin 064: Three CVSS 10.0 Vulnerabilities Enable Unauthenticated Full Device Compromise
Ubiquiti published Security Bulletin 064 on 22 May disclosing five CVEs in UniFi OS devices, three of which score CVSS 10.0: an improper access control flaw, a path traversal enabling arbitrary file read and write, and a command injection that provides root shell access — all exploitable without authentication from the network. Enterprise environments using UniFi Wi-Fi infrastructure must update immediately.
GlobalProtect CVE-2026-0257 Compromise Indicators: Threat Hunting and Forensic Guide for VPN Gateway Authentication Bypass
Organisations running PAN-OS GlobalProtect gateways on versions vulnerable to CVE-2026-0257 must investigate for compromise during the exposure window, not just apply the patch. This guide covers the specific log sources, indicators of compromise, and post-exploitation patterns to hunt for on PAN-OS GlobalProtect gateways after an authentication bypass zero-day.
PAN-OS GlobalProtect CVE-2026-0257: Rapid7 Confirms Second Exploitation Wave — CISA Adds to KEV
Rapid7 MDR confirmed on 21 May that a second, larger exploitation wave of CVE-2026-0257, an authentication bypass in Palo Alto Networks GlobalProtect VPN, began on 21 May targeting enterprise sectors not covered in the initial wave. CISA added the CVE to the Known Exploited Vulnerabilities catalogue with a 1 June remediation deadline. The vulnerability affects PAN-OS 10.2, 11.1, 11.2, and 12.1 as well as Prisma Access.
VPN Authentication Bypass: Identity and Access Containment Response After GlobalProtect Compromise
When a VPN authentication bypass like CVE-2026-0257 is exploited, the attacker enters the network without leaving identity provider audit trails. Standard identity-based detection misses the initial access. This creates a specific response challenge: containing a network breach where the entry event did not generate authentication telemetry and the scope of subsequent access is unknown.
ChromaDB CVSS 10.0 Pre-Auth RCE CVE-2026-45829: AI Vector Database Compromise via HuggingFace Model Injection
HiddenLayer and the Cloud Security Alliance published disclosures of CVE-2026-45829, a CVSS 10.0 unauthenticated remote code execution vulnerability in ChromaDB's Python FastAPI server, on 18–20 May 2026. Attackers can inject malicious code via a crafted HuggingFace-hosted model before the authentication gate fires. Approximately 73% of ChromaDB deployments are internet-exposed. No patch exists for affected versions.
CISA Adds Seven to KEV Catalogue — Including Two Active Microsoft Defender Zero-Days Patched via Silent Engine Update
CISA's 20 May Known Exploited Vulnerabilities batch included CVE-2026-41091 (Microsoft Defender for Endpoint EoP, CVSS 7.8) and CVE-2026-45498 (Microsoft Defender DoS, CVSS 4.0), both patched via a silent Defender engine update pushed on 19 May. The batch also included five legacy Windows and Adobe vulnerabilities from 2008–2010 indicating re-exploitation of outdated systems in active campaigns.
Drupal SA-CORE-2026-004: Highly Critical SQL Injection CVE-2026-9082 — PostgreSQL Sites Must Patch Immediately
Drupal published SA-CORE-2026-004 on 20 May, disclosing CVE-2026-9082, a highly critical unauthenticated SQL injection vulnerability in Drupal's database abstraction API affecting sites running PostgreSQL. The flaw is zero-click and unauthenticated, and Drupal warned that exploit code turnaround would be measured in hours. CISA added the CVE to the Known Exploited Vulnerabilities catalogue on 22 May after confirmed exploitation.
Securing RAG Pipeline Architecture: Vector Databases Are the New Unmanaged Attack Surface in Enterprise AI
The ChromaDB CVE-2026-45829 disclosure exposes a systemic architectural gap in enterprise AI deployments: vector databases used in retrieval-augmented generation pipelines are being deployed without the security controls applied to comparable databases handling sensitive data. The attack surface analysis and architectural recommendations for secure RAG pipeline design apply regardless of which vector database product is in use.
SonicWall EoL Highlights an Asset Management Gap: Network Equipment Lifecycle Tracking in Enterprise Environments
The SonicWall Generation 6 end-of-life situation reveals a consistent gap in enterprise asset management: network equipment EoL dates are not tracked with the same rigour as software licence renewals or server hardware refresh cycles. Organisations with accurate, proactively managed network equipment lifecycle records have a weeks-to-months advantage in responding to EoL-driven security risks.
End-of-Life VPN Appliances: A Security Assessment Framework for Identifying Unsupportable Network Equipment
The SonicWall Generation 6 end-of-life situation is the latest instance of a recurring enterprise security problem: internet-facing network equipment that reaches vendor end-of-life while still actively exploited. A structured assessment approach helps security teams identify, prioritise, and communicate the risk of EoL perimeter equipment.
Exchange CVE-2026-42897 One Week On: Active Exploitation Continues, No Patch Available — Updated Guidance
Microsoft Exchange Server's OWA session hijacking zero-day CVE-2026-42897 entered its second week without a permanent patch. Microsoft's Emergency Mitigation Service (EEMS) rule remains the only automated protection for Exchange Online-connected on-premises environments. Security teams should now focus on identifying whether exploitation occurred during the disclosure week and verifying their mitigation status.
SonicWall Gen6 SSL-VPN: Patch for CVE-2024-12802 Fails to Close MFA Bypass — Akira Ransomware in 86% of Compromises
ReliaQuest published research on 19 May confirming that SonicWall's official firmware patch for CVE-2024-12802 on Generation 6 SSL-VPN devices requires six manual reconfiguration steps to fully close the MFA bypass vulnerability. Devices that reached end-of-life on 16 April 2026 will receive no further patches. Akira ransomware is present in 86% of SonicWall-involved intrusion claims reviewed by ReliaQuest.
AI Coding Agents in CI/CD Pipelines: Mapping the Attack Surface After Pwn2Own AI Category Results
The Pwn2Own Berlin 2026 AI category results — five products exploited — have a compounding implication for organisations where AI coding agents are integrated with CI/CD pipelines, code repositories, and cloud deployment infrastructure. An exploited AI agent running in a pipeline is not a developer workstation compromise; it is a supply chain entry point.
Pwn2Own Week Exposes the Limits of Identity as a Security Control — What IAM Teams Should Review
The week of 12–18 May 2026 produced two distinct scenarios where identity controls — Conditional Access, MFA, and Zero Trust enforcement — provided no meaningful protection: Exchange Server-side RCE (operating below the authentication layer) and Exchange OWA session hijacking (stealing tokens after authentication). Both are active or imminent threats. Both require defences that go beyond the identity layer.
The Pwn2Own 90-Day Clock: How Defenders Should Use the Patch Window Before Public Disclosure
Pwn2Own's 90-day coordinated disclosure rule gives vendors time to patch before technical details are made public. For enterprise defenders, the same 90 days is a known timeline during which the confirmed existence of specific zero-days — but not their technical details — is public. Understanding how to use that window is an underexplored aspect of enterprise vulnerability management.
After Pwn2Own Berlin 2026: A Risk Manager's Assessment of 47 Zero-Days in Enterprise Infrastructure
Pwn2Own Berlin 2026 produced 47 unique zero-day vulnerabilities across Windows 11, VMware ESXi, Exchange Server, SharePoint, Oracle VirtualBox, Red Hat Enterprise Linux, and five AI products. For enterprise risk managers and CISOs, the results require a structured response that goes beyond individual CVE patches and addresses the systemic implications.
Red Hat Enterprise Linux LPE at Pwn2Own: What the Results Mean for Enterprise Linux Patch Strategy
Red Hat Enterprise Linux was successfully exploited twice at Pwn2Own Berlin 2026 via local privilege escalation vulnerabilities. For enterprise security teams running RHEL, and the broader family of RHEL-derived distributions including CentOS Stream, Rocky Linux, and AlmaLinux, the results inform how Linux patching SLAs should be evaluated against the demonstrated threat model.
Why Exchange SYSTEM RCE Bypasses Conditional Access and MFA: The Authentication Architecture Problem
The Exchange SYSTEM RCE chain demonstrated by DEVCORE at Pwn2Own Berlin 2026 achieves code execution at the operating system level, bypassing all identity controls including Conditional Access policies, MFA requirements, and Azure AD authentication entirely. Understanding why server-side RCE renders identity controls irrelevant is essential for accurate risk assessment.
Pwn2Own Berlin 2026 Closes: DEVCORE Wins Master of Pwn with $505K and 50.5 Points — $1.3M Total Across 47 Zero-Days
Pwn2Own Berlin 2026 concluded with DEVCORE Research Team winning the Master of Pwn title with $505,000 in earnings and 50.5 points, driven by Orange Tsai's Exchange SYSTEM RCE chain and consistent results across multiple targets. The three-day competition produced 47 unique zero-day vulnerabilities across enterprise products, cloud infrastructure, and AI tools, with $1,298,250 in total prize money awarded.
Pwn2Own Demonstrates Second Distinct SharePoint RCE Chain — Five Days After Patch Tuesday Fixed CVE-2026-40365
Researchers at Pwn2Own Berlin 2026 demonstrated a multi-bug SharePoint Server remote code execution chain that is entirely distinct from CVE-2026-40365, the SharePoint RCE patched in the 12 May Patch Tuesday. The new chain, targeting SharePoint's server-side processing pipeline, has no patch and will not receive one for up to 90 days.
Windows 11 Yielded Four Independent LPE Paths at Pwn2Own Berlin — Kernel Attack Surface Analysis
By the close of Pwn2Own Berlin 2026, researchers had demonstrated four separate, independently discovered privilege escalation paths from standard user to SYSTEM on fully patched Windows 11. Each exploited a different component and vulnerability class. The results indicate the Windows kernel and user/kernel boundary remain a consistently productive attack surface for skilled researchers.
AI Coding Environments Join Pwn2Own Target List: LM Studio and OpenAI Codex Exploited via Sandbox Escapes
Pwn2Own Berlin 2026 introduced an AI products category and saw both LM Studio and OpenAI Codex exploited on the same day through sandbox escapes and environment variable injection. The results raise urgent questions about the security of AI development tools running inside enterprise environments with access to code repositories, credentials, and production pipelines.
Cisco SD-WAN CVE-2026-20182 Post-Compromise Forensics: Identifying Rogue Device Injection in Catalyst SD-WAN Deployments
CVE-2026-20182, the CVSS 10.0 Cisco Catalyst SD-WAN Manager zero-day added to CISA KEV on 14 May, was exploited before Cisco released the patch. Organisations that ran vManage on publicly accessible addresses during the exposure window must now forensically audit their SD-WAN device inventory and API authentication logs for signs of rogue device registration and traffic interception.
Exchange CVE-2026-42897 Threat Hunting Guide: Identifying Session Hijacking in OWA Logs
With no patch available for the actively exploited Exchange OWA session hijacking zero-day, security teams must hunt for existing compromise rather than waiting for a fix. This guide covers the specific log sources, KQL queries, and behavioural indicators that reveal CVE-2026-42897 exploitation in on-premises Exchange and Microsoft 365 hybrid environments.