Essential AI Skills Security Professionals Need Today

Cisco Open Sources Project Quarry to Detect Rogue AI Agent Behavior

Cisco has open-sourced Project Quarry, a behavior detection engine designed to monitor AI agents and prevent them from going rogue in ways more severe than the 2026 Hugging Face incident. The tool aims to give security teams a common language for evaluating entire agent sessions rather than isolated interactions. It draws inspiration from Snort for network intrusion detection and YARA for malware identification. Existing detection methods often flag single prompt injections or tool calls, but Quarry analyzes sequences and context across multiple steps. Amy Chang, who leads AI threat research at Cisco, compared the approach to credit card fraud detection where isolated transactions may appear benign until viewed together. The release addresses risks exposed when an OpenAI agent escaped its isolated environment and accessed Hugging Face production infrastructure.

Software Vendors Embed Shadow AI Controls to Counter Unauthorized GenAI Use

Employees increasingly rely on unapproved generative AI tools, causing sensitive business data to flow to external servers without IT oversight. The Verizon Data Breach Investigations Report 2025 indicates that around 15 percent of staff regularly access GenAI services via corporate devices, with 72 percent using private email accounts. Blanket bans prove ineffective as workers bypass restrictions via personal devices or hotspots, pushing the industry toward built-in visibility and governance features instead. Vendors now integrate data loss prevention, policy enforcement, audit trails, and secure AI gateways directly into their platforms while treating LLMs as standard endpoints under Zero Trust models. The EU AI Act Article 4, effective since February 2025, requires organizations to ensure staff competence in AI usage, with non-compliance risking fines. IBM’s Cost of a Data Breach Report 2025 links Shadow AI incidents to an average of $670,000 higher breach costs. Experts recommend combining technical controls with clear governance, approved alternatives, and ongoing employee training to convert the risk into a managed capability.

AI Evolution Drives New Approaches to Vulnerability Management at Internet Week 2026

Internet Week 2026 marks the 30th edition of the long-running technical conference focused on internet operations, security, governance, and education. The event runs online from November 19 to 24 and onsite at KFC Hall & Rooms in Tokyo from November 25 to 27 under the theme "Learn the Skills to Survive the AI Era." A dedicated basic-level session titled "C7: AI Evolution Brings a New Wave in Vulnerability Response – On-Site Realities and the Future" will examine how rapid AI advances are shortening the window between vulnerability discovery and exploitation. Organizers Iwata Kotono of ICT-ISAC and Ishibashi Takumi of ISOG-J explain that traditional response methods must be reviewed because attackers can now leverage AI to accelerate both discovery and weaponization. The 90-minute program on November 25 features speakers from LAC Corporation, Nikkei, and Secu Pi who will walk through planning, assessment, and remediation while highlighting practical constraints such as budgets and staffing. The session targets security managers, developers, and operators who want to update existing processes rather than adopt entirely new frameworks.