It's almost as if having a readable contract that the signers can all understand and a judge who can rule on reasonable interpretations of the contract is a feature, not a bug.
Whatever component of the "smart contract" that allowed them to do this would have been negotiated out, but since it's a decentralized "the contract enforces itself" cryptosystem, people just trust that the contract does what it says on the tin.
What should be the procedure before using these? Pay a security researcher for an independent review, like you would have a lawyer go over a contract? If you have to do that, why not just use a regular contract and lawyers where you have recourse when something goes wrong?