Attackers use AI agents to sweep GitHub for leaked secrets and abuse service principals to break into Azure tenants

Microsoft Threat Intelligence said on November 5, 2026 that since August it has seen threat actors use an agentic AI offensive framework to search public GitHub repositories for credentials, application credentials and AI inference API keys, then use them for further malicious activity against victim organizations ^[1]. Microsoft says it observed successful service principal sign-ins from a specific IP since October 28 involving multiple applications at some customers, including a vendor-owned multitenant app credential stolen at the source and used against multiple downstream tenants, and iThome reports one case went from searching GitHub to exfiltrating data from Azure Key Vault in five minutes ^[1][2]. Microsoft advises organizations to scan their GitHub code for secrets and immediately rotate any exposed keys or credentials, since exposed credentials stay usable until revoked or rotated and deleting the original disclosure does not fix it ^[1].

Industries mainly affected: Cross-industry

Microsoft advises organizations to scan their GitHub code for secrets and immediately rotate any exposed keys or credentials, since exposed credentials stay usable until revoked or rotated and deleting the original disclosure does not fix it