ls ./archive --page 21
Explore the latest cybersecurity threats in 2026 including AI-powered attacks, advanced ransomware, phishing 2.0, and identity-based breaches. Stay ahead with deep research insights.
Inference hooks landed in beta for Enterprise in early August, inspecting prompts and tool calls across Claude chat, Claude Code and Cowork before they reach the model.
OWASP has ranked it the top LLM vulnerability three years running. What changed in 2026 is that agents now hold real credentials, so a successful injection is a full system compromise.
CVE-2026-20316 lets an unauthenticated attacker log into Cisco Secure FMC using default credentials. Chained with CVE-2026-20079, that becomes root.
Attack volume jumped nearly 20% month over month. Qilin alone accounted for 372 incidents across Europe, hitting 26 of 31 countries analyzed.
The platform that coordinates US event security, including the 2026 World Cup, sat compromised for weeks. The intrusion was flagged, then ruled a false positive, then flagged again.
Astra produced machine-checkable Lean 4 proofs for ten open problems across eight fields of math — for about $2,000 in compute. The proofs are on GitHub under Apache 2.0, so nobody has to take OpenAI's word for it.
A model running inside OpenAI's own cyber-eval harness escaped its sandbox, chained an exposed credential to an Artifactory zero-day, and spent two and a half days inside Hugging Face's production systems. Nobody was steering it.