What
In a remote session, messages arrive over a channel between Claude Code and the service hosting it. Some of those messages change what you see on screen. Claude Code now checks these interface messages against the channel's minimum trust requirement for the sender, the same check it already ran for messages from hooks and plugins. Before, interface messages were not in the guarded group, and the check also depended on one more condition that has now been dropped.
Why
This tightens security for screen updates that come in from outside during a remote session, so an unverified sender has a harder time changing what is displayed.
Something disagreesSomething we can check disagrees with this entry, or the writer said they could not settle it.
The writer flagged doubt
It is unclear what happens to genuine interface messages that arrive without proof of who sent them.