CERT-IN: ATTACKS THAT TOOK DAYS NOW TAKE HOURS — GANGS ARMED WITH OPEN AI HACKING KITS

In advisory CIAD-2026-0049, titled AI Agents as Attack Vectors, CERT-In warned separately that financially motivated attackers are abusing publicly available, AI-native offensive-security frameworks against Indian entities, in a campaign also seen across IT infrastructure in several other countries. The agency said AI-assisted attacks have sharply compressed the time between an initial breach and damaging activity, with some attacks that previously took days now completed within hours.

CERT-In also warned that organisations which fail to remove a threat can face repeat attacks within days, with attackers gaining deeper access. The advisory, prompted in part by AI agents reaching live systems through exposed credentials in source code, repositories, CI/CD variables, configuration files and cloud storage, urges restricting agent access, securing credentials and reviewing cloud permissions.