Examples

1. 1.

   CyberArk's non-human identity security platform manages service accounts for 2,000+ enterprise clients - including AI agent identities - enforcing credential rotation, just-in-time access, and usage anomaly detection.

2. 2.

   GitHub Advanced Security flags over-permissioned service accounts in CI/CD pipelines - a common finding is AI agent service accounts with org-admin scope that should be scoped to a single repository.

3. 3.

   OWASP's Non-Human Identities Top 10 (2024) highlights that AI agents are the fastest-growing category of NHIs - with risks including long-lived tokens, hardcoded credentials, and missing revocation processes.