Why not execute the shell command described by the RemoteCmd but the Communicator?

[qia01ng]

There is one ExecuteCommand field in Communicator struct and trigger the ExecuteCommand execution in the Start() method; But what confused me is that why the command to be executed is not the command described by the RemoteCmd?

// https://github.com/hashicorp/packer-plugin-sdk/blob/main/shell-local/communicator.go
type Communicator struct {
	ExecuteCommand []string
}
func (c *Communicator) Start(ctx context.Context, cmd *packersdk.RemoteCmd) error {
	if len(c.ExecuteCommand) == 0 {
		return fmt.Errorf("Error launching command via shell-local communicator: No ExecuteCommand provided")
	}

	// Build the local command to execute
	log.Printf("[INFO] (shell-local communicator): Executing local shell command %s", c.ExecuteCommand)
	localCmd := exec.CommandContext(ctx, c.ExecuteCommand[0], c.ExecuteCommand[1:]...)
	localCmd.Stdin = cmd.Stdin
	localCmd.Stdout = cmd.Stdout
	localCmd.Stderr = cmd.Stderr

	// Start it. If it doesn't work, then error right away.
	if err := localCmd.Start(); err != nil {
		return err
	}

	// We've started successfully. Start a goroutine to wait for
	// it to complete and track exit status.
	go func() {
		var exitStatus int
		err := localCmd.Wait()
		if err != nil {
			if exitErr, ok := err.(*exec.ExitError); ok {
				exitStatus = 1

				// There is no process-independent way to get the REAL
				// exit status so we just try to go deeper.
				if status, ok := exitErr.Sys().(syscall.WaitStatus); ok {
					exitStatus = status.ExitStatus()
				}
			}
		}

		cmd.SetExited(exitStatus)
	}()

	return nil
}