The First Agentic Attack: How AI Is Reshaping the Economics of Cybersecurity
Palo Alto Unit 42 documented the first confirmed AI-agent-launched cyberattack, compressing weeks of intrusion into under 10 hours. The market response was immediate.
Palo Alto Unit 42 documented the first confirmed AI-agent-launched cyberattack, compressing weeks of intrusion into under 10 hours. The market response was immediate.
The rapid integration of autonomous agents into enterprise environments has exposed a critical security gap at the execution layer. Traditional network firewalls and point-in-time security assessments are increasingly insufficient for managing the risks inherent in plugin-heavy architectures. As agents gain the ability to execute code and interact with external systems, the focus of cybersecurity infrastructure…
Restricted mode in Postgres MCP Pro was designed as the safe configuration for exposing a PostgreSQL database to an AI agent. It limits operations to read-only transactions and validates incoming SQL against an allowlist. CVE-2026-85620, disclosed this week with a CVSS v4.0 score of 9.2, bypasses that entire control with a single syntactic trick. The…
The Million-Payment Benchmark On September 3, 2026, the Solana Foundation announced the launch of Payment Channels, accompanied by a headline-grabbing figure: 1 million payments per second. This benchmark, derived from a controlled test involving 100,000 unique wallets through a proxy, does not represent current mainnet throughput. While the capacity to handle 80 billion payments in…
It has been a long, winding road for Revolut in the United States. After hitting a wall with their California state bank charter application back in 2023—a move largely derailed by regulatory friction and internal-control headaches—the London-based fintech is finally back in the game. They have secured preliminary conditional approval from the Office of the…
For years, enterprise security has been a game of perimeter defense and identity management. But as companies move from simple chatbots to autonomous AI agents – software that can execute tasks, access data, and make decisions on behalf of a user – the old playbook is fraying. The recent $100M Series B funding for Austin-based…
SonicWall has disclosed a second zero-day chain affecting its SMA1000 series appliances within seven weeks. The vulnerabilities, detailed in SonicWall advisory SNWLID-2026-0016, consist of CVE-2026-83548, a pre-authentication server-side request forgery (SSRF) in the Appliance Work Place interface, and CVE-2026-83549, a post-authentication OS command injection in the AMC component. Both flaws are actively exploited in the…
The Intercontinental Exchange is formalizing a two-vendor digital transfer agent program, positioning the New York Stock Exchange as the primary regulated venue for tokenized public equities. By signing a memorandum of understanding with tZERO on August 31, 2026, ICE has secured a premier design partner for its upcoming Digital Trading Platform. This follows the exchange’s…
The smart home hub sitting on your counter is marketed as a seamless command center for your life, promising to manage lights, locks, and climate with a simple voice prompt. It feels like a leap into the future, but that polished interface hides a crumbling foundation. August 2026 has been a brutal month for the…
MCP Security is the discipline of identifying, mitigating, and managing the risks created by the Model Context Protocol — the open standard that connects AI agents to external tools, data sources, and APIs.