Unclear The finding does not say what the prefixed wrapped blocks contain or where they come from.
What
Claude Code can mark a message it sends with humanTurn: true, which flags the message as a real turn typed by a person. This applies to prompt messages from certain origins, such as remote sessions. The marker is now withheld when the message text consists entirely of wrapped blocks. The text is checked after joining it and trimming leading whitespace.
The existing on/off check still has to pass as well. That check sits behind the tengu_parsed_willow flag and can be overridden with the CLAUDE_CODE_PARSED_WILLOW environment variable. Its built-in fallback applies the check. Nothing has been read about this flag yet.
Why
Messages that carry only wrapped, automatically added content stop being counted as turns a person actually took. This can affect how a session tells your own messages apart from generated ones.
tengu_parsed_willow Not enough to sayNothing here resolved what this flag was doing on this version, so nothing here should be read as on or off.
This account: no value returned · anonymous baseline: no value returned · compiled default in v2.1.281: on
These values were read against a different version of Claude Code, so treat them as the nearest reading available instead of one taken on this release.
Read once, for one account on one subscription tier, against v2.1.281. It isn't a statement about your account. What a flag value here can and cannot tell you
The finding does not say what the prefixed wrapped blocks contain or where they come from.