Add emulate command for mobile/device viewport emulation

Add --timeout flag (default 30s) to prevent hung processes

Add --annotate flag to screenshot command

Add --agent-id for parallel agent isolation and atomic file writes

Add upload command for file inputs

Add --full flag docs for screenshot and fix --full being used as save path

README: add comparison with agent-browser CDP mode

Add tab/window create and close commands

Initial release: agent-chrome CLI for AI agents to interact with running Chrome via CDP