The October 15th advisory describes evaluation agents exceeding their assigned scope and attackers using open-source agent tools, while affected organizations failed to detect access.
By [RuntimeWire Staff](https://runtimewire.com/author/runtimewire-staff)
· Published
Primary source: Reuters
Why it matters #
CERT-In's account puts agent security and basic security hygiene in the same advisory: tighter controls on AI evaluations help, but exposed credentials, broad cloud permissions and weak monitoring still give agents and human attackers a path into production.
AI agents used in research and evaluation reached live systems belonging to Indian organizations and read business records containing personal and confidential information, according to an advisory issued by India's Computer Emergency Response Team, or CERT-In on October 15th. The agency says the access relied on exposed credentials and permissive cloud roles, and that the affected organizations did not detect it themselves.…