The Orchestration Framework as an Independent Attack Surface Holding a single model constant—Claude Sonnet—and varying the orchestration framework produces an Agent Compromise Rate (ACR) spread from 11.9% to 31.1%. This 2.6x variance demonstrates that the orchestration framework functions as an independent attack surface, introducing vulnerabilities that model-only evaluation cannot detect. As noted in the Synthetic…
Cisco is not running a pilot program. Starting at the end of July 2026, the company is deploying a personalized AI agent to every single one of its 90,000 employees. This is a structural overhaul of how a Fortune 500 company operates, moving beyond the experimental phase into a full-scale integration of autonomous agents across…
In early 2026, unemployment for new graduates hit 5.6%, a 1.6 percentage point increase from three years prior. This figure is not merely a byproduct of broader economic cycles; it is a direct reflection of how firms are re-engineering their cost structures around artificial intelligence. While a July 2026 policy brief from the Stanford Institute…
The Industry Builds Its Own Regulator Washington is currently occupied with the GENIUS Act, a framework that focuses on stablecoin issuers while leaving the mechanics of machine-initiated transactions largely untouched. As Congress stalls on the specifics of AI-driven finance, the private sector has begun to construct its own regulatory architecture. On August 4, 2026, the…
A colleague of researcher Sammy Azdoufal recently bought a baby monitor on Amazon and, like any sensible person, asked if it was actually safe. It’s the kind of question we all ask ourselves before plugging in a new device, usually followed by a shrug and a hope that the brand name on the box knows…
On August 6, 2026, the agent ecosystem got its own npm moment. A coalition of major platform operators—Amazon, Microsoft, OpenAI, Vercel, and Cursor—simultaneously adopted Agent Plugins 1.0.0, an open, vendor-neutral standard for packaging Agent Skills and Model Context Protocol (MCP) servers into portable plugins. This is not a proposal or a white paper. It is…
Alibaba is pioneering a revenue-share model on open-weight Qwen3.8, the first Chinese lab to formalize monetization beyond API pricing. Tests whether Max-class performance can command deployment royalties.
The security architecture of AI agents is facing a fundamental challenge: the trust-through-defaults model. As developers increasingly rely on autonomous coding agents like Claude Code, Gemini, and ChatGPT to streamline workflows, these tools are being weaponized in a campaign codenamed FakeGit. This technique, termed AgentBaiting, exploits the agents’ propensity to autonomously discover and recommend external…
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,…
Zenity Labs disclosed a new vulnerability class at Black Hat on August 5 that it calls Intent Collision, a zero-click attack vector that hijacks agentic browsers by injecting hidden instructions into any web page the agent visits. The demonstration compromised Claude in Chrome, Gemini, Perplexity Comet, ChatGPT Atlas, and Copilot Edge without requiring the user…