In one of my groups we have an animal companion which is a wolf. Since it only has one natural attack (bite), once it reaches an appropriate level it gets an iterative attack in place of Multiattack. Thing is, the wolf gets trip.
The way I see it, there are two possibilities. The CMB for trip is linked to the functional BAB used in each attack, or the CMB is static.
Example... let's assume the wolf has +10/+5 for its attack routine and a CMB of +8.
Option #1. Wolf attacks at +10 and trips at +8 then attacks at +5 and trips at +3.
Option #2. Wolf attacks at +10 and trips at +8 then attacks at +5 and trips at +8.
The numbers are completely artificial in this example, but demonstrate my point... does the CMB for a secondary attack's use of trip suffer the same -5 that the attack itself does?
When you attempt to perform a combat maneuver, make an attack roll and add your CMB in place of your normal attack bonus. Add any bonuses you currently have on attack rolls due to spells, feats, and other effects. Combat maneuvers are attack rolls, so you must roll for concealment and take any other penalties that would normally apply to an attack roll.
Your first scenario is the correct one.
Edited to add the actual sentence that I meant to add in the first place.
Just to back up Brotato with the rules from the book, here is this little snippet:
Combat maneuvers are attack rolls, so you must roll for concealment and take any other penalties that would normally apply to an attack roll.
Any penalties to attack apply to CMB, including the penalty for iterative attacks.