Well, you could have your script tail /var/log/nsm/securityonion/sguild.log for lines containing "Alert Received:". Those lines include the basic metadata associated with Snort events as they occur. You could even configure OSSEC to monitor /var/log/nsm/securityonion/sguild.log for specific Snort alerts and to call a custom active response script to take action about them. Or you could interact with the securityonion_db mysql database. Start with the event table.
Perceiving which Snort events are truly of a high priority is not a simple task though. I have never found the Snort rule class types and associated priority numbers to be of much help when trying to automate the generation of administrative alerts about "important" Snort events. There are many factors to consider in auto-classification of events, some of which can be very site specific. I ended up building a rule set that a script uses to evaluate Snort events to determine what if any administrative action should be taken (like cell phone page, admin email, or auto-quarantine of the local host). It's another layer of rules, but it had enabled me to stay aware of high value Snort events in near real time.
Regards,
Kevin