Microsoft Says AI Agent Went From GitHub Secrets to Azure Key Vault Data in Five Minutes

Microsoft says in a security blog post that threat actors are using an agentic AI offensive framework to search public GitHub repositories for credentials, application secrets and API keys. In one observed case, an AI agent moved from searching GitHub to extracting data from Azure Key Vault within five minutes.

Microsoft began seeing the framework submit large volumes of queries to GitHub in August 2026, and points to agentic tooling like ARTEX as the kind of tool that lets attackers scan and follow up with little hands-on work. In activity beginning as early as October 28, attackers used exposed application credentials to access Microsoft Entra ID tenants and connected Azure and Microsoft 365 resources. The agents searched for credentials, mapped cloud resources and accessed services including Azure Key Vault, Storage Accounts, SharePoint and OneDrive.

Microsoft advises organizations to scan public repositories, revoke or rotate exposed credentials and enable secret scanning and push protection. Removing a secret from a repository does not prevent misuse if the credential remains active. The company also recommends replacing stored application secrets with managed identities where possible.