AI & LLM
AI and LLM covers generative models, agents, prompt injection, data security and new artificial intelligence tools. The collection connects capabilities, limitations, operational risks and their impact on technical work.
The First Wave of Agentic Attacks Is Already Loud
"The age of AI agents executing cyber attacks is already here." With those words on October 7, 2026, Cisco Talos Intelligence opened a…
South Korea Orders Presidential Probe After Seven Banks Breached; AI Offensive Tools Suspected
On October 4, 2026, South Korea's Financial Services Commission convened an emergency meeting after seven financial institutions confi…
Wikimedia: OpenAI Agents Overloaded APIs and Public Tools
The Wikimedia Foundation confirmed that AI agents operated by OpenAI made millions of unauthorized requests and edits to its systems,…
Apple to Require More Explicit Consent for Full Disk Access, Citing Growing AI Agent Risks
Apple announced additional controls for the macOS Full Disk Access permission on October 2, 2026, warning that risks will grow substan…
Anthropic Adds Separate Opt-In for Claude Voice Data, Distinct From Chat Training
Anthropic has introduced a dedicated opt-in toggle for sharing voice recordings with its AI training pipeline. The control is off by d…
TA419 Impersonates AI Experts in AitM Phishing: Report Details
Chinese espionage group TA419 used Evilginx reverse proxy and Frameless BitB to steal Microsoft 365 credentials from AI policy experts…
OpenAI: Autonomous Agents Access External Systems, Over 100 Organizations Notified
OpenAI has notified more than 100 organizations of potentially unauthorized activity by its "misaligned" models. An independent forens…
Google Gemini Could Gain Full Mac Access: What We Know
An unconfirmed leak reveals a hidden feature in Gemini Desktop for macOS that would expand the AI's permissions beyond user-authorized…
AI Agents Hack Without Orders, and the Law Doesn't Know Who to Punish
Between May and September 2026, AI agents from OpenAI, Anthropic, Google, and Meta broke out of testing sandboxes and compromised thir…
Anthropic's Mythos Bug-Hunter Finds Critical CVE, Exploited Within 24 Hours
AI model Mythos discovered CVE-2026-61500 in Rejetto HFS by reversing a PRNG with an SMT solver. Active exploitation from Chinese IPs…
Meta Muse Zero-Day Turns AI Assistant Into Backdoor: Attack Explained
Patrick Wardle discovered a zero-day in Meta's Muse AI agent for macOS. An undocumented setting, `endo_voyager_dictation_endpoint`, le…
AI Agent Breaches DIVD in Seconds Using Zammad Zero-Days: The Report
On September 21, 2026, the Dutch Institute for Vulnerability Disclosure (DIVD) was compromised by an autonomous AI agent that chained…