Weekly INK

AI Agents Undermine Progress in Browser Security Source: Dark Reading Early browser security gains are being eroded as AI agents and complex extensions increase attack surface. The piece explains how automated agents can misuse permissions, pivot via tabs, and skirt traditional controls, with practical advice to rein in extensions, harden policies, and monitor browser telemetry in business environments. Link to article Okta SSO accounts targeted in vishing-based data theft at

'Most Severe AI Vulnerability to Date' Hits ServiceNow Source: Dark Reading Authentication issues in ServiceNow potentially opened the door for arbitrary attackers to gain full control over the entire platform and access to the various systems connected to it. ServiceNow is a Fortune 500 company that, according to its promotional materials, acts as an IT services management platform for 85% of the companies that comprise the rest of the Fortune 500. Link to article Microsoft

CTO New Year’s Resolutions for a More Secure 2026 Source: Dark Reading Experts lay out practical security goals for the new year, from operationalizing AI governance to hardening CI/CD and improving CISO–CTO alignment. For smaller teams, the list doubles as a roadmap: standardize secure-by-default build paths, tighten vendor access, and measure security like any other business KPI. Link to article Texas court blocks Samsung from collecting smart TV viewing data Source: Bleepi

Browser Extension Harvests 8M Users’ AI Chatbot Data Source: Dark Reading A popular Chrome extension marketed for privacy was quietly collecting and selling content from users’ AI conversations. For SMBs, that is a reminder to restrict browser extensions, enforce allow lists, and review permissions that can capture on-screen data and network traffic. Link to article Coupang data breach traced to ex-employee who retained system access Source: BleepingComputer Coupang says a fo

CISA Warns of 'Ongoing' Brickstorm Backdoor Attacks Source: Dark Reading US agencies say Chinese state actors are deploying the Brickstorm backdoor in VMware vSphere environments, enabling persistence, VM snapshot theft, and lateral movement. For SMBs supporting critical sectors: harden vSphere, restrict remote access, enforce MFA, and monitor for anomalous VM and DNS-over-HTTPS activity. Link to article Pharma firm Inotiv discloses data breach after ransomware attack Source:

Arizona AG Sues Temu Over “Stealing” User Data Source: Dark Reading Arizona’s attorney general sued Temu, alleging the shopping app secretly harvests sensitive device data and evades reviews. U.S. firms should expect renewed scrutiny of mobile SDKs, background data collection, and consent. Review privacy notices, telemetry settings, and third-party code used in consumer apps. Link to article Google fixes two Android zero-days exploited in attacks (107 flaws total) Source: Ble

Critical Fortinet FortiWeb WAF Bug Exploited in the Wild Source: Dark Reading A newly disclosed FortiWeb flaw lets attackers run admin-level commands on unpatched web application firewalls. For SMBs that rely on WAFs to protect websites and portals, this is a patch-now event: exposed devices can be taken over pre-login, leading to data theft, website defacement, or downtime. Link to article Kraken Uses Benchmarking to Enhance Ransomware Attacks Source: Infosecurity Magazine A

OWASP Highlights Supply Chain Risks in New Top 10 List Source: Dark Reading A major OWASP refresh spotlights software supply chain failures and misconfiguration as top risks. For SMBs, this means looking beyond code bugs to vendor components, CI/CD pipelines, and cloud settings. The takeaway: add supply-chain checks to patching, and tighten configuration governance to reduce real-world breach paths. Link to article Microsoft Uncovers ‘Whisper Leak’ Attack That Identifies AI C

Multiple ChatGPT Security Bugs Allow Rampant Data Theft Source: Dark Reading Researchers found seven weaknesses that let attackers steal chat history and “memories,” bypass safety checks, and plant malicious instructions—no deep technical skill required. For SMBs exploring AI, this signals immediate risk: tighten browsing features, restrict plug-ins, and treat AI tools like any other internet-facing app. Link to article Microsoft: SesameOp malware abuses OpenAI Assistants API