Under the hoodTier: how much it should matter to you
1Useful: my rating, 1 to 5
1Signal: worth watching, 1 to 5
PermissionsArea: what it touches
Internal ChangesKind: in v2.1.304,
Internal ChangesSection of the release
What
Before Claude runs some actions, it asks for permission, and the answer can come from somewhere other than the terminal, such as a program built on Claude Code. An answer can also change the input the action runs with.
When such an answer arrives late, Claude Code now checks that changed input. If it is too large or too deeply nested, the action is not run and Claude sees this message:
"The call was not run. The permission answer gave it an input past what this session takes from an answer"
Why
This is a new safety limit on permission answers that come from remote or programmatic sources, so an outsized answer cannot slip an oversized input into an action.
