Permission-centric adaptive security
ML-CRPA keeps a session from becoming blanket trust. After login, an employee completes a manager-code check and a logged-in network and device acknowledgment. Each sensitive operation is scored and, when policy requires it, verified again.
1 Login
2 Session re-auth
3 Acknowledgment
4 Role dashboard
5 ML decision
6 Step-up grant
Sign in
Demo usernames: admin, accountant, pm, data. Passwords and manager codes are in the project README.