Group of 2 Nothing to try yet
Behind an experimental flag, the monitor tool's watch timeout can now expire and require a manual re-arm instead of running indefinitely.
What
- The Monitor (background watch) tool's timeout behavior changed: behind an experimental flag, it now uses a bounded timeout schema with a capped, defaulted
timeout_ms, instead of the previous unbounded schema. - When the bounded timeout expires, Claude Code now sends a specific notice that includes how much time elapsed and how many events were delivered (or none), and instructs the model to "re-arm" the watch rather than keep polling.
- The system prompt text describing background monitors changes to match: instead of saying a monitor sends "one notification per occurrence... indefinitely", it now says "until the monitor expires (re-arm to continue)" when this flag is on.
Why An unbounded watch could run forever consuming resources; a bounded, self-expiring watch forces a periodic check-in ("re-arm") so long-running monitors don't run unattended indefinitely.
Feature flag
tengu_breezy_crescent 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: off · compiled default in v2.1.268: off
Read once, for one account on one subscription tier, against v2.1.268. It isn't a statement about your account. What a flag value here can and cannot tell you
Flag reading moved
The flag server now returns on for tengu_breezy_crescent, read as this account.
A reading is one sample. Claude Code evaluates its flags remotely, so no client sees the targeting rule behind a value and this says nothing about your account.
One source agreesOne thing we can check says the same as this entry.
The name it cites is new in this build
New in this build: tengu_breezy_crescent