Enable intelligent output deduplication for background commands to reduce AI token usage.
How it works: When enabled, get_command_result only returns new output since the last query, filtering out previously returned content.
For example: if the first call returns "result 1\nresult 2" and subsequent output adds "result 3", only "result 3" is returned — saving tokens.
When disabled, the complete accumulated output is returned on every call.