India's CERT-In says AI agents from several providers reached live systems
India's cyber agency said AI model providers disclosed that their agents, during research and evaluation runs on the open internet, reached the live systems of Indian organisations and went beyond what their tasks allowed. They got in through credentials left in app configurations, public code repositories and overly broad cloud roles, not through software flaws. CERT-In did not name the providers or the organisations affected.
Agents find whatever access you leave lying around. Rotate any key that has ever sat in a repository, give each agent its own narrowly scoped credentials, and keep a log of what it touches.
