OIP-1 Pause Guardian

There is often a window of 3-9 minutes between an attack contract deployment and the execution of the attack. If the contracts can be paused in this time, the funds can be saved.

Attacks are often first noticed by people external to the organization.

To pause currently requires:

1. Someone to notice the exploit
2. Find out how to contact us
3. Actually contact someone
4. Convince someone
5. Initial origin responder has to contact another signer
6. We have to create the pause transaction
7. Second signer has to sign

This takes too long.

Instead we could create a contract that has one time use pause actions. After 12 hours OUSD can be unpaused unless the pause is confirmed by a multisig. These pause authorizations can then be distributed to external security researchers that are monitoring the blockchain. Then if they see an attack come through they can pause immediately in one step. This could cut the reaction time down to 2-3 minutes.

Even if this was only used internally, this would still be quiet the speed difference.

Extra thoughts

• Pause authorizations would have an expire time. mapping(address =>uint256)
• Pause authorization is zeroed on use
• Pause would be in three states: unpaused, temp-pause, confirmed pause
• Anyone can unpause when a temp-pause expires.

No files, tests, or entry points are named in the issue. Start by clarifying the contract design and reviewing the existing pause flow; done would mean implementing and testing temporary, confirmed, expiring, and single-use pause authorizations.