
CERT-In Warns Autonomous AI Agents Accessed Indian Systems Without Authorization
CERT-In warns autonomous AI agents accessed Indian systems without authorization, exposing business data, as attackers misused open-source tools.

CERT-In warns autonomous AI agents accessed Indian systems without authorization, exposing business data, as attackers misused open-source tools.

Huntress reports unknown actors used a compromised ConnectWise Cloud Backup credential for unauthorized sign-ins to 173 Microsoft 365 tenants.

Sansec reports BlockStrap template flaw could let unauthenticated attackers run code on Adobe Commerce and Magento servers; no exploitation observed.

Researchers say a PoC chain using shared-memory overflow and use-after-free gives local root on OpenBSD 7.9, with fixes via syspatch.

Attackers briefly altered AOMEI Backupper updates to deploy CROSSROAD and SHINYROAD backdoors via unsigned files, researchers report.

UAT-11985 uses fake event invites and spoofed Google login pages to relay credentials and MFA prompts targeting Taiwan researchers.

Burger King Japan disclosed app breach potentially exposing 11.18M membership records; credit cards not stored, no evidence passwords accessed.

Authenticated Langflow flaw enables OS command execution via MCP Stdio config; Previdian reports exploitation attempts and urges upgrade to 1.9.0.

BeyondTrust Remote Support zero-day exploited in hosted and self-hosted appliances, enabling file operations and root access, linked to Armory Castle.