What
When a background session's turn is interrupted or paused waiting for a permission answer, Claude Code "parks" it and can resume it later. The settings for this behaviour have changed:
- The settings (such as
resumeInterruptedTurn,adoptUnderivableandparkedStopRetires) used to be built by reading environment variables likeCLAUDE_CODE_RESUME_INTERRUPTED_TURNandCLAUDE_CODE_PARKED_STOP_RETIRESdirectly, plusanswerWaitMs. They are now taken from one of two prepared sets, a saved one or a served one. The environment variable names still appear in the code, so they are probably read elsewhere. - A new flag,
tengu_polished_salamander, is read as the value 'on' to decide whether a parked permission message is held open (holdsParkedMessageOpen). The flag server returned off for this site's account and on for an anonymous check, but no reading has been taken under this release yet. - The resume settings now include
holdOpenFor, taken from the message that opened the interrupted turn, which keeps that parked message open.
Why
The defaults for parking and resuming may now come from these prepared sets rather than from environment variables you set, so check how they behave if you rely on those variables.
tengu_polished_salamander Gate removed from the codeThis release deleted the gate from the code while it was still reading on for the account this site reads, so the code path no longer asks a flag before running.
This account: on · anonymous baseline: on · compiled default in v2.1.288: not a boolean we can read
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 v2.1.288. It isn't a statement about your account. What a flag value here can and cannot tell you
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.
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.
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.
The code path no longer asks a flag before it runs.
What a reader actually sees differently, and what applies when the server sends nothing, are not settled.