ltm

You can control prompt cost by toggling low-token mode, which modifies prompt behavior for a single command or an entire session using sentinel files to persistently track and enforce your token reduction settings.

Low-token mode is a toggleable state that alters prompt behavior to reduce token consumption, affecting either the current command or all subsequent prompts in a session by relying on sentinel files to maintain explicit and predictable state changes.

Check the low-token mode status by running the command with the status argument, which reads the current sentinel files to determine whether the mode is active, disabled, or overridden for your session.

Yes, you can apply low-token mode to a single command by using the once argument, which temporarily overrides the session state for that specific prompt without permanently changing the behavior for the rest of the session.

The best way to enable low-token mode for all subsequent prompts is to use the session or on argument, which updates the sentinel files to ensure persistent token reduction behavior across the rest of your active session.

To re-enable normal prompt behavior after overriding low-token mode, use the re-enable argument, which adjusts the sentinel files to restore the default or session-level token management settings predictably.