1How do frontier AI agents perform in multi-step cyber-attack scenarios? (opens in new tab)(aisi.gov.uk)3lebovic9d ago0
2Hacking Claude Code remote: escaping YOLO-mode sandboxing (opens in new tab)(noahlebovic.com)4lebovic28d ago0
3Evaluating and mitigating the growing risk of LLM-discovered 0-days (opens in new tab)(red.anthropic.com)62lebovic1mo ago18
4Show HN: An agent sandboxing quickstart based on Claude Code (opens in new tab)(github.com)2lebovic2mo ago0