The instructions Claude Code gives Claude for the Agent tool (which hands work to a subagent, a separate helper run of Claude) and for the Skill tool can now be built with extra inputs: the model in use and an option named leanRewrite. The Agent and Skill tool descriptions did not change in this release, so these inputs have no visible effect yet.
Separately, the artifact tool now records whether edits were sent with the publish call itself, and Claude Code now stores a frozenPublishEdits value.
These are prepared variations of Claude's instructions and the artifact tool. Nothing you see changes until they are put to use.
tengu_velvety_sprout 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 v2.1.301: off
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 v2.1.301. It isn't a statement about your account. What a flag value here can and cannot tell you