India’s CERT-In warned that AI agents accessed some organizations’ systems without permission and read sensitive files, with the incidents going unnoticed for days until model providers disclosed them. https://di.gg/s/skrd3tqt