Hello, stay ahead with CISO Brief 🚀

🧭 This blog documents an experiment using Google Cloud to deploy a Gemma 3 inference workload across two regional GKE clusters, leveraging TPU v6e instances, managed DRANET for accelerator networking, and a multi-cluster Inference Gateway for cross‑region routing and failover. It describes building VPCs, reserving internal IPs, configuring Cloud Storage FUSE for model storage, creating TPU node pools with managed DRANET, registering clusters into a GKE Fleet, and deploying the inference server and gateway with health checks and autoscaling metrics. The objective is resilient, low‑latency routing to the nearest region with automatic failover to the other region if one fails.

🔒 This post shows how SaaS providers can use Amazon Bedrock AgentCore resource-based policies to control multi-tenant access to a shared AgentCore Runtime and Runtime endpoint. It walks through two tenant scenarios: cross-account access for Example Corp and VPC-restricted access for AnyCompany, demonstrating how to apply resource-level Allow and explicit Deny conditions. The article covers required IAM permissions, example policy files, and verification steps to ensure network- and identity-based constraints are enforced.

🔒 Paul Chichester of the NCSC warned at Infosecurity Europe that escalating technological change, geopolitical tensions and evolving threats make predicting cyber risk harder than ever. He highlighted hyper-connectivity, rapid tech transformation and state-backed cyber operations as key challenges, and urged stronger public-private collaboration. Chichester praised the Cyber Security and Resilience Bill and called for practical steps like reducing attack surface, addressing legacy systems, enforcing access controls and running incident exercises.

🛡️ Joe Slowik warned at Infosecurity Europe that defenders must adopt AI to keep pace with adversaries. He argued that purely human-driven SOCs cannot match the accelerated timescales enabled by AI, ML and LLMs, leaving organisations exposed. Slowik recommended rethinking security operations to integrate AI agents for rapid intelligence, enrichment and remediation, while keeping humans in the decision loop. He used the React2Shell example to illustrate the speed of modern exploits.

🧭 At Infosecurity Europe 2026, Bayer CISO Kevin Jones outlined a shift from checklist-based guidance to psychology-first security awareness to counter AI-enabled social engineering. The firm mandates behavior-focused training, ties AI access to role-based modules, and gates agent development behind completion. Bayer is moving SOCs toward supervised automation and updating supplier contracts and governance to enforce AI transparency and controls.

🛡️ The U.S. Cybersecurity and Infrastructure Security Agency (CISA) has mandated federal agencies to patch an actively exploited Oracle WebLogic vulnerability, CVE-2024-21182, by June 4 under BOD 22-01. The flaw affects Oracle WebLogic Server versions 12.2.1.4.0 and 14.1.1.0.0 and enables unauthenticated remote compromise via T3/IIOP. Shodan reports over 1,592 exposed and vulnerable WebLogic instances, and CISA urges all organizations to apply vendor mitigations or discontinue use if fixes are unavailable.

🔍 Kaspersky analyzed public Wi‑Fi across Mexico City, Guadalajara, and Monterrey ahead of the 2026 World Cup. The team wardrove to log 84,500 signals and 69,500 unique SSIDs, finding about 82% use WPA2/WPA3 but over 10% are unsecured. WPS was enabled on roughly 45% of access points, often even when WPA2/WPA3 was in use, increasing attack risk. The report also warns of other travel threats like malicious QR codes, public USB chargers, NFC/Bluetooth exploits, and evil‑twin networks. Kaspersky recommends using cellular data or an eSIM and a VPN to stay safe when connecting to public networks.

🔒 Google released the June 2026 Android security updates fixing 124 vulnerabilities, including one actively exploited Android Framework zero-day (CVE-2025-48595) affecting devices running Android 14 and later. The company warned of limited, targeted exploitation and urged users to update to the latest Android versions. Two patch bundles (2026-06-01 and 2026-06-05) were issued; Pixel devices will receive updates immediately while other vendors may delay. Google also addressed 18 critical flaws across System, Framework, and Qualcomm components, and previously patched other zero-days this year.

🔒 Bruce Schneier reflects on his 2010 Dark Reading essay arguing that while cryptography provides strong mathematical advantages, it cannot by itself secure modern, interconnected systems. He traces how crypto has been applied since the 1990s and explains that computer security is an ongoing arms race of fragile defenses. Schneier warns that AI changes the landscape by automating vulnerability discovery and exploit creation, shifting the balance between attackers and defenders.

🔒 An anonymous researcher known as “Nightmare Eclipse” has published several significant exploits targeting Microsoft Windows, including a vulnerability that defeats BitLocker. Microsoft has responded with threats of legal action, prompting public debate and recriminations between the company and security community. The situation has raised concerns about disclosure practices, researcher protections, and the balance between security research and corporate legal responses.

🔍 Sophos X-Ops uncovered a lab where a threat actor used AI coding tools to develop and test malware aimed at evading EDR products. The files and Git repository showed Python scripts—many partially AI-generated—used to build and iterate evasion modules against vendors including Sophos, CrowdStrike and Microsoft. Humans retained control of the workflow, using AI to accelerate building, testing and refinement while operating inside an AI-native environment.

🔍 New research from ManageEngine reveals UK IT and business leaders view AI-powered cyber-attacks as their top risk over the next 12 months, with 43% identifying it as the single biggest threat. The survey of 1,500 decision-makers across five European markets shows 41% of UK respondents plan to prioritise spending on tackling AI and advanced threats. Despite strong detection rates, UK organisations report increasing incidents, skills gaps and recovery challenges, alongside rising investment in resilience and governance.

🛡️ Researchers discovered a malicious npm package named codexui-android that impersonated an OpenAI Codex UI and exfiltrated developer authentication tokens. The package was published to npm with malicious code absent from the project's public GitHub repository, highlighting risks in artifact distribution. Security experts warn this pattern exploits trust in legitimate-looking developer tooling and reveals blind spots in software supply chain controls.

🔍 A new paper from Silobreaker and the SANS Institute warns that business leaders often misunderstand threat intelligence and its value, creating an "intelligence–stakeholder gap." The report, launched at Infosecurity Europe 2026, finds that intelligence outputs can be overlooked or misinterpreted, limiting funding and visibility for intelligence teams. To close the gap, teams must tailor briefings to senior leaders, provide forward-looking exposure analysis, prioritise speed and seek regular stakeholder feedback to ensure intelligence changes decisions and drives risk-informed actions.

🔔 Amazon RDS for SQL Server now supports Bring Your Own Media (BYOM), enabling customers to migrate SQL Server workloads to a managed AWS service while reusing existing Microsoft SQL Server licenses and Software Assurance via Microsoft's License Mobility program. The capability integrates with AWS License Manager to help track license usage and maintain compliance. BYOM reduces the need to purchase additional SQL Server licenses or wait for existing agreements to expire when moving to RDS. It aims to simplify migrations from on‑premises, other clouds, or self‑managed EC2 deployments.

🛡️ Discussion-based, low-stress simulations let IT, legal, and business leaders walk through hypothetical incidents to test preparedness, but poorly run tabletops can mislead and harm response capabilities. The article outlines seven common mistakes — from lacking clear objectives and testing only familiar scenarios to favoring conceptual scripts over practical ambiguity — and offers expert recommendations to design realistic, business-relevant exercises. Emphasis is placed on including the right stakeholders, introducing technical detail and uncertainty, and aligning scenarios to actual risks and interdependencies to avoid false confidence and reveal true process gaps.

🔐 Dashlane disclosed a brute-force attack on May 31, 2026, targeting certain personal accounts to bypass two-factor authentication and register new devices. Its security controls triggered temporary suspensions and authentication issues, and although access has been restored, attackers succeeded in downloading encrypted vaults for fewer than 20 personal-plan users. Dashlane stressed that vault contents remain protected by each user's Master Password and that its internal systems were unaffected.

🔬 AWS HealthOmics now lets customers specify the Nextflow engine version at run time via the StartRun API, enabling explicit version pinning for controlled migration. Supported versions include 22.04, 23.10, 24.10, 25.10, and 26.04 via a new engine-settings parameter. This run-time override takes precedence over manifest.nextflowVersion, allowing testing across engine versions without changing workflow source. The feature is available in all AWS HealthOmics regions and supports HIPAA-eligible, production-regulated workflows.