On the Automated Correction of Security Protocols Susceptible to a Replay Attack