AGENT-EYE

Linux

agent-eye works the same on Linux, with these differences:

  • The background process is a systemd user service. agent-eye install writes ~/.config/systemd/user/agent-eye.service and starts it with systemctl --user; its log is journalctl --user -u agent-eye. It runs while you are logged in; loginctl enable-linger $USER keeps it running after you log out. Where there is no systemd, run agent-eye daemon yourself (install --no-service skips the service).
  • Resume and Start open a terminal window when there is a display (DISPLAY or WAYLAND_DISPLAY), with the first of xdg-terminal-exec, x-terminal-emulator and $TERMINAL that is installed ([board] resume = "auto", the default). The board says which: TERMINAL (xdg-terminal-exec). With no display or no such program they copy the command instead, and the board says COPY. resume = "copy" always copies; resume = "terminal" skips the display check. In the picker, Enter and Ctrl-G run the command in your current terminal as usual.
  • The background service may have no display. A systemd user service only gets DISPLAY or WAYLAND_DISPLAY when your session passes them on (systemctl --user import-environment DISPLAY WAYLAND_DISPLAY, which most desktops do at login). Without them the board copies. A terminal the service opens stays open when the service restarts (the unit uses KillMode=process).
  • Copying uses wl-copy, xclip or xsel, the first that works. With none, the command is printed instead. The background process copies for the board, and it may have no display; then the board shows the command.
  • eye board on --open opens the board with xdg-open.

Edit this page on GitHub