Stream-json output now includes the stop reason field tools expect on assistant messages.
What's wrong with this entry?
When Claude Code assembles an assistant message from accumulated text on the stream-json path, it now sets stop_reason: "end_turn" and stop_sequence: null. Tools consuming stream-json output see a complete, correctly shaped message instead of one missing those fields.
stop_reason: "end_turn"
Strings lifted out of the shipped bundle, so the claim above can be checked against them.
Related
Other releases about the same thing. Found by shared names or similar wording; neither means one caused the other.
-
v2.1.235
Eval results written inside a plugin warn that they may load as components
Both mention output
-
v2.1.238
Files changed during a turn are uploaded to the cloud session worker
Both mention output
-
v2.1.248
Detection of "I'll check back later" phrasing in model output
Both mention output