I'm not sure if this is a bug ...
From the README:
"Commands that operate on an instance (claude-send-command, claude-code-switch-to-buffer, claude-code-kill, etc.) will prompt you for the Claude instance if there is more than one instance associated with the current buffer's project."
When already in a eat/vterm buffer with a claude instance, some commands, eg. toggling read-only mode, should use that instance without prompting. If there are strong reasons to retain the current behavior, maybe consider a variable to override it ?
I use this workaround right now:
(define-advice claude-code--get-or-prompt-for-buffer (:around (orig-fn) use-current)
"If already in a claude buffer, use it directly."
(if (claude-code--buffer-p (current-buffer))
(current-buffer)
(funcall orig-fn)))
I'm not sure if this is a bug ...
From the README:
"Commands that operate on an instance (claude-send-command, claude-code-switch-to-buffer, claude-code-kill, etc.) will prompt you for the Claude instance if there is more than one instance associated with the current buffer's project."
When already in a eat/vterm buffer with a claude instance, some commands, eg. toggling read-only mode, should use that instance without prompting. If there are strong reasons to retain the current behavior, maybe consider a variable to override it ?
I use this workaround right now: