Credential and Secret Exposure in Agent Output

Detects when an AI agent exposes API keys, secret tokens, private keys, database connection strings, JWT tokens, or other sensitive credentials in its output. Covers all major cloud provider key formats, CI/CD tokens, payment processor keys, SSH keys, .env file content patterns, and generic secret assignment patterns. Credential leakage in agent output poses a critical security risk leading to unauthorized access, lateral movement, financial loss, and full account compromise.

建議回應

參考資料