1st layer functions don't know anything about 2nd layer stuff. So you can't escrow 2nd layer tokens.
But you can create a 2nd layer escrow function.
You can achieve all functions on the 2nd layer just like on the first layer: it's just a different group of people running different software.
Essentially with 2nd layer, you can run one or more separate "blockchain protocols" with their own rules that just piggy back of the p2p network and transaction ordering functionality of the 1st layer. So the "trustworthiness" of a specific 2nd layer protocol just depends on how that protocol is designed. There's no reason it can't be just as trustworthy as the 1st layer protocol.
Creating this token at 1st layer would slow the chain?
Too much data to be tracked?
It is possible that a malicious update could be successful?
At least for the amount of time it takes to revert the changes?
Keychain would be my example.
I couldn't defend against a malicious update absent the social aspect of hive tipping me off to delete the app.
So with open source distributed nodes keeping everybody honest a 2nd layer is just as secure as hive?
How are you envisioning the functioning of 'bonds'?
A '2nd layer' gui that manages 'bond tokens' according to the contract at the time of escrow?
ie, I stake hbd for a set period of time and receive interest, to liquidate the position I send the token to a 2nd layer contract that executes according to its parameters?