trusted bridge

A trusted bridge is one where you must trust a specific set of people or machines beyond the two blockchains themselves. The bridge appoints an external group — call them guardians, validators, oracles, or a multisig — who watch the source chain, and when they see a valid deposit they collectively sign a message authorizing the matching release or mint on the destination. The destination contract does not check the source chain's consensus; it only checks 'did enough of our designated signers approve this?'

This is by far the easiest kind of bridge to build, which is why it dominated early cross-chain infrastructure. It works between any two chains, however different their consensus or virtual machines, because the signers do all the cross-chain reasoning off-chain and the on-chain part is just signature verification. The cost is the trust assumption: if a threshold of those signers is compromised or colludes, they can mint unbacked tokens or release locked collateral, and the chains themselves will see nothing wrong. The honest-majority assumption is the whole security argument.

Most of the largest bridge hacks in history were trusted bridges whose signer keys were stolen. The Ronin bridge lost about 625 million dollars when an attacker obtained five of nine validator keys. The mitigations are familiar: raise the threshold, diversify and harden key custody (HSMs, MPC, geographic and organizational separation), add per-transfer rate limits and time delays so a sudden drain can be paused, and publish the signer set transparently. But none of these change the fundamental shape — you are trusting a club, not mathematics.