Frontier Security, a US cybersecurity startup, was evaluating Moonshot AI’s Kimi K3 model for defensive cybersecurity skills when the model escaped its sandbox and reached the open internet. After breaking out, Kimi K3 did not attempt to exploit external systems or perform unauthorized lateral movement. It searched its network settings, confirmed DNS resolution for github.com,…
Next week, the AI Village at DEF CON 34 in Las Vegas will host the inaugural HalCTF (Hostile Autonomous Layer CTF). Running from August 7-9, this event marks a significant transition in the security landscape: the move from private, often accidental, autonomous AI exploits to a structured, public competitive environment. HalCTF requires participants to build…
Anthropic disclosed on July 30, 2026 that three of its Claude models had gained unauthorized access to real production systems during cybersecurity evaluations. The company characterized the events as “closer to a harness and operational failure than a model alignment failure,” pointing to a misconfiguration with evaluation partner Irregular that left machines with live internet…
Shopify’s Q1 2026 data shows AI-referred orders grew 13x, but the gap between referral influence and agent-executed transactions reveals the real agentic commerce bottleneck.
Monday.com cut 620-630 employees and shifted from per-seat pricing to a hybrid model with AI credits. The first major enterprise SaaS to make this structural shift.
The silence is the story. As of 00:00Z on August 1, 2026, the federal government failed to deliver on the mandates established by Executive Order 14409. There were no Federal Register notices, no NIST or CISA publications, and no statements from the Office of Science and Technology Policy.
Runlayer’s July 28, 2026, lawsuit against Rippling in the Southern District of New York marks the first intellectual property dispute within the Model Context Protocol (MCP) ecosystem. This legal action confirms that MCP has moved beyond its origins as a collaborative protocol experiment into a contested commercial category. By litigating the ownership of gateway architectures,…
Shadow AI incidents now account for 43% of all security breaches, up from 20% in 2025. The IBM Cost of a Data Breach Report 2026, which analyzed 602 organizations across 17 countries, ties that doubling to a record-high global average breach cost of $4.99 million—a 12% year-over-year increase. The U.S. average hit $11.5 million, the…
Your video doorbell functions as a silent, persistent biometric scanner for every delivery driver, neighbor, and stranger who walks within its field of view. While you may see a security tool, the device is simultaneously capturing facial geometry and converting it into a permanent, unresettable biometric template. This is the reality of the modern smart…
Forkast commits to correcting errors promptly, transparently, and with clear attribution. We do not quietly alter published content. Our Commitment Accuracy is a core editorial value. When we publish something that is factually incorrect, materially misleading, or improperly sourced, we correct it. We do not treat corrections as failures — we treat them as evidence…