The top 3 verified AI security incidents from the last 2 days — ranked and summarized automatically.
A now-patched vulnerability in AWS Bedrock AgentCore could allow an attacker to use one AI chatbot to take over an organization's entire fleet.
Read full report →Cybersecurity researchers have disclosed details of a targeted campaign aimed at South Korean financial organizations that used an artificial intelligence (AI) pen testing tool named ARTEX to carry out the attacks. The activity, per CrowdStrike Intelligence, was active from late September to early October 2026, and resulted in data exfiltration.
Read full report →OAuth grants create data highways between SaaS apps, AI agents, and other tools. And, they are multiplying faster than any security team can review them.
Read full report →