Archives

Frontier AI models deployed in a ransomware attack completed a network breach in 10 hours, adapting in real time without zero-days or elite tradecraft.

AI coding agents run code from .git/config during startup before showing trust prompts.

Forescout researchers used Claude to port an RCE exploit between WAGO PLCs, but it required 8+ hours of guidance and API costs to work.

METR disclosed a March 2026 incident where an attacker stole $600K in AI credits by prompting an agent to reveal its own API key.

Russia-aligned UAC-0099 embedded nuclear-weapon prompts in malware to disable AI-powered threat analysis.

OpenClaw 2.0 expands accessibility to agent harnesses while leaving critical security configurations to users.

Ransomware operators deployed Cursor Agent to automate exploitation, bypassing safety controls with trivial prompt injection.

China-linked threat actor pivots from hypervisor attacks to Cisco routers and TACACS servers, creating invisible observation points in enterprise networks.

Infostealer malware drains Claude quotas by stealing browser sessions, which bypass authentication entirely.

Prototype monitoring system by CLTR to detect AI scheming behaviors using open-source intelligence

Researcher shows Claude Code Auto Mode vulnerable to 60-80% RCE via module shadowing.

Three maximum-severity vulnerabilities in ServiceNow AI Platform allow unauthenticated attackers to execute code and extract data via SQL injection.

OpenAI and METR published their Hugging Face post-mortems on the same day, Dream's forensics filled in the Taiwan attack, and Australian police charged two men over a year of npm and PyPI worms.

Major labs issue joint warning on agentic threats, but the timing and predictive framing deserve scrutiny.

Amazon Kiro's vulnerability shows the systemic risk when AI agents trust repository content to control their tool use.

Australian police charged two men in their early twenties as the alleged leaders of TeamPCP, blamed for a year of npm and PyPI supply-chain attacks. The group open-sourced its worm toolkit in May, and someone else was already using it in August.

An AI-assisted research tool discovered 700+ novel HTTP request-smuggling vulnerabilities by exhaustively mining RFCs.

OpenAI's technical report and METR's independent investigation both landed on August 26. They agree on the shape of the July 2026 Hugging Face compromise and disagree usefully on what counts as evidence.

Ramp's monthly index of paid AI adoption and business spending.

AnonyMousKIT's AI voice attack reveals how automating social engineering becomes viable at scale.