I used to think that "cross-chain" just meant transferring tokens from A to B, and that the chains would recognize each other; at most, trusting the bridge's smart contract would suffice. Now I understand that, in a single cross-chain operation, you're trusting more than just one component: who generates the message, who relays/packages it, how verification is done (light clients/multisig/external validators), how the destination chain executes it, and even things like oracles/pricing—any anomaly in these details can easily trigger everyone to collectively "wait for confirmation"... Recently, bridges have been hacked again, and often it's not because the code is terrible, but because the level of trust is too high; if one link is bypassed, it's game over. Anyway, when I look at bridge/IBC-like systems now, my first step is to draw a trust linkage diagram, or I won't have a clear picture in my mind.