I followed one credential out of a Saudi flow into a UAE system.
It didn’t resolve.
Same schemaId.
Same attestation.
Still nothing.
I thought I pulled the wrong one.
Ran it again.
Same result.
No rejection.
No mismatch.
Just... empty.
That didn’t hold.
So I flipped it.
Tried a UAE credential back into a Saudi flow.
Same outcome.
At that point I thought I was breaking something.
Because nothing was failing.
It just wasn’t showing up.
So I stayed on it longer than I planned.
Tracing where it actually stops.
The identity exists.
The issuer is trusted.
Inside the system, everything resolves cleanly.
Outside it...
nothing points to it.
The credential doesn’t fail.
It just disappears from the surface.
I checked another one.
Different schema.
Same behavior.
That’s when it clicked.
Sovereign silo.
Trust holds inside the boundary.
And disappears the moment it crosses it.
That part stuck.
Because nothing here is broken.
Saudi works.
UAE works.
Jordan works.
All of it is real infrastructure.
But it doesn’t move.
I saw it again with a profile moving across systems.
Already verified on one side.
Nothing usable on the other.
Had to start over.
Later it showed up in a different flow.
Same identity.
Investor this time.
Still no resolution.
Then again during a distribution check.
Eligible on one side.
Invisible on the other.
The identity exists.
The proof exists.
It just doesn’t travel.
That’s where it shifted for me.
This isn’t a verification problem.
It’s a resolution problem.
And resolution doesn’t fail loudly.
It just stops existing at the boundary.
No error.
No signal.
No indication that something valid just became unusable.
That’s what makes it hard to catch.
Because everything still looks correct.
Until you try to use it somewhere else.
Movement didn’t stop.
People.
Capital.
Access.
All of it is already crossing systems.
The verification layer just isn’t moving with it.
So every time something crosses that boundary...
the system behaves like it’s seeing it for the first time.
That’s the gap.
Not infrastructure.
Resolution.
$SIGN only matters here if an attestation issued under one schema can actually be resolved by another system without restarting verification from scratch.
Because right now...
every boundary resets trust.
And that cost compounds.
Quietly.
Across everything trying to move.
So the real question becomes this.
At what point does this stop being acceptable?