Title changed, previously was the same as the article:

"AI agents aren’t ‘going rogue.’ Frontier labs are skipping security basics."

Not a fan of the change. Feels like the article title was changed for being too critical of OpenAI and Anthropic