I think you misinterpreted what a SNMP-trap is. A SNMP-trap is a SNMP message sent to your monitoring system/service from a network device such as a router, switch, blade, cluster, ..
I guess the thing you want to do is search the MIB-file for the particular network device you want to monitor and search the OID that matches the information you want the gather from that specific device.
The device you want to monitor through SNMP has to have SNMP enabled in it's configuration (webbased or something..).
You can do a SNMPwalk to this device to see all available OIDs :
snmpwalk -v 2c -c public <ip address network device>
-c stands for 'community' and by default this is 'public', you can edit this in the configuration of your network device.
-v stands for the version of SNMP you want to use.
When you find the OID which provides you the device's information you wore looking for you can do the following command (or put this in a perl or bash script) :
snmpwalk -v 2c -c public <ip address network device> <OID>
When you made this script you can define a command for this script in commands.cfg :
#'check_lefthand' command definition
define command{
command_name check_lefthand
command_line $USER1$/lefthands.pl $ARG1$ $ARG2$
}
You can now use this check_ in your service definitions of Nagios.