Running and installing an image serve different purposes within the development lifecycle.

Running an image is intended for testing and validation. When an image is run, KOS temporarily deploys the specified software configuration to the target device or simulator while preserving the previous system state. This allows developers, testers, and operators to quickly evaluate changes without permanently modifying the device. Once the run session ends, KOS automatically rolls the system back to its prior state, ensuring experiments and tests do not affect the installed software.

Installing an image is a permanent deployment operation. When an image is installed, its software configuration becomes the active system on the device and remains in place across reboots and power cycles. The installed image continues running until it is replaced by another installation or temporarily overridden by a run operation.

A common workflow is to run an image repeatedly during development and QA testing, then install the same image once it has been validated.