OpenAI agents discussed ways to escape their sandbox on public wiki

A security-news RSS collection covering active exploitation of Mac screen-sharing vulnerabilities, supply-chain compromises leaking credentials, BGP hijacking, malicious proxy devices, phishing and AI-agent security failures, browser fingerprinting, LLM prompt-injection data exfiltration, account takeover defenses, surveillance technology, and cybercrime investigations. The most urgent items involve vulnerabilities reportedly under active exploitation, large-scale credential exposure, and attacks affecting production networks and corporate environments.