My finger was over “Confirm” when I noticed the destination network was wrong. The amount was small, but that was exactly the point: bridges do not forgive a rushed click.
That is the catch with mantabridge and any cross-chain bridge. You are not simply sending a token from one wallet to another. You are selecting a source chain, a destination chain, an asset route, and a contract that may require a wallet approval. A mistake in any one of those can leave funds delayed, unsupported, or unrecoverable.
The decision I actually make
When I need to move an asset, the options are usually simple: use the first bridge in a search result, use a bridge I have used before, or stop and verify the route before signing. I used to choose the second option because familiarity felt like safety. After one transfer took longer than expected and the status was unclear, that stopped being good enough.
Now I check four things before connecting a wallet: the domain, the source and destination networks, the exact token, and the wallet transaction. I also look at whether the displayed route matches what I intended—not merely whether the token symbol looks familiar. Then I send a small test amount. If the test arrives and the wallet shows the expected asset on the expected chain, I continue. If the interface asks for an approval that does not fit the action, I reject it and investigate.
For the Manta route, the reference I use is the mantabridge bridge, opened by typing the address myself or using a bookmark I created after checking it. The useful part is not treating the page as proof that everything is safe; it is using it to inspect the intended route before the wallet presents a transaction.
Hindsight has made the process slower by perhaps two minutes and safer by a great deal. A bridge is useful when it removes a real network boundary. It is not useful when it persuades you to sign before you understand where the funds are going.