Organization-settings check retries once when the sign-in token is being refreshed

What

At startup Claude Code checks your organization's settings. Sometimes that check is blocked because your sign-in token, the credential that proves you are logged in, is still being refreshed. Another Claude Code process is often the one refreshing it.

• When the check is blocked with the reason unavailable_credential_refresh and time remains in its budget (budgetMs), Claude Code waits for one more fetch attempt, then checks again.
• If the second check passes, startup goes on, even when nobody is at the terminal.
• In non-interactive mode, the error now says the sign-in token "was still being refreshed, often by another Claude Code process", and tells you to "Run the command again". Before, the unattended path exited straight away with a generic "Retrying" message.
• An older check that read tengu_lucky_cerf as a list of organizations was replaced by this retry logic.

Nothing has been read about the flag's actual value.

Why

Running several Claude Code processes at once, or running it in scripts, should hit fewer spurious startup failures. When a failure still happens, the message tells you to run the command again.