News
AI Agent Bypasses Sandbox Controls in a16z De Fi Study
1+ day, 10+ hour ago (399+ words) An artificial intelligence (AI) agent broke out of the sandbox that a16z crypto engineers built during a test. The engineers wanted to evaluate whether AI agents can move beyond identifying vulnerabilities to building working exploits. Security engineers Daejun Park and Matt…...
Anthropic's Claude Wiped a Company's Database, Then Told on Itself
1+ day, 6+ hour ago (340+ words) Anthropic, one of the most highly valued and controversial AI companies in the world, whose pre-IPO valuation officially hit a record $1 trillion on Monday, had one of its AI agents go rogue. Tom's Hardware reported that Pocket OS founder Jer…...
Ghost tapping: The scam charging your card without you knowing
2+ day, 6+ hour ago (510+ words) Your credit card never leaves your pocket" but somehow, your money does. It's called Ghost Tapping. A proximity-based scam that uses the same tap-to-pay technology you see at legitimate businesses every day to steal from you. Using cheap, easy-to-buy card…...
AI Agent Deletes Startup's Database in 9 Seconds, Founder Says
1+ day, 20+ hour ago (582+ words) A software company founder claims an AI coding agent destroyed his firm's production database, then copped to the mistake and explained how it happened, demonstrating the potential danger of entrusting sensitive access and materials to automated bots. Jeremy Crane, founder…...
North Korea-linked hackers are using fake Zoom meetings to target crypto execs
2+ day, 40+ min ago (413+ words) A clipboard hijack trick replaced benign commands with malware'deploying code The operation enabled rapid credential theft, persistence, and full system compromise Security researchers Arctic Wolf have revealed details of a highly sophisticated campaign targeting North American Web3 and cryptocurrency companies. It…...
North Korea-linked actor targets Web3 execs in social-engineering campaign
2+ day, 20+ hour ago (305+ words) Researchers warn that a financially motivated unit of North Korea's Lazarus Group has been running a social-engineering campaign that uses fake Zoom or Teams calls against senior-level executives in cryptocurrency and blockchain." The threat actor, tracked as Blue Noroff, targeted…...
Red Hat's Open Claw maintainer just made enterprise Claw deployments a lot safer
2+ day, 11+ hour ago (542+ words) On Tuesday, Red Hat principal software engineer Sally O'Malley released a new open source tool called Tank OS to make it easier to deploy and manage Open Claw agents more safely. "This was a fun project that I put together…...
Signal promises new security measures after wave of phishing attacks
2+ day, 13+ hour ago (407+ words) The encrypted messaging app Signal has said it will introduce additional security measures following phishing attacks targeting politicians, military personnel and journalists in several countries in recent months. Signal wrote in a post on social media on Monday "in the…...
This Claude-powered AI agent deleted a company's whole database " and then gloated about it
2+ day, 30+ min ago (220+ words) To make matters worse, all of Pocket OS's recent backups were also deleted because Railway stores them on the same volume. Crane says the most recent recoverable volume the company had access to was three months old. As if the…...
The blueprint architecture for securing the AI data center
2+ day, 13+ hour ago (798+ words) As enterprises turn traditional data centers into AI factories powered by LLMs, they're focused on unlocking new revenue streams, competitive differentiation, and operational efficiencies. But they're also exposing themselves to unprecedented risk. Enterprises are no longer just leasing AI. They…...