Skip to content

Troubleshooting

No Console Output in UI

If the "Console Output" box in the UI is empty but the agent seems to be running:

  1. Check Backend Terminal: Look at the terminal where you started run_ui.sh. Do you see logs there?
    • AgentCommander uses Dual Logging: Logs are printed to the terminal and sent to the UI.
  2. Browser Refresh: If the terminal has logs but the UI doesn't, the WebSocket connection likely dropped. Refresh the page to reconnect.
  3. Check "Console" Visibility: Ensure you are on the Control tab. The console is located at the bottom of the Control Panel.

"Run Task" Error: Unexpected keyword argument 'no_exec_list'

This error indicates a mismatch between the running ui_server.py and the installed pylib.

  • Fix: Restart the UI Server (Ctrl+C then bash run_ui.sh). Python code changes in the library require a restart to be loaded.

Metric Not Found / Infinite Score

If the agent keeps failing with "Metric is INF":

  1. Check Evaluator Output: Use the "White-box Debugging" method to inspect eval_out.txt in the experiment directory.
  2. Check Print Format: Ensure your evaluator.py prints exactly Best metric: 0.123 (Case sensitive: lowercase 'm' in metric).
  3. Check Leakage: If the Anti-Cheating check fails, the score is forced to Infinity. Check if your model is modifying y_test in place.

Windows WSL2 Issues

  • Accessing UI: If localhost:8080 doesn't work, try the WSL IP address (run wsl hostname -I).
  • Permissions: Ensure your project folder is inside the WSL filesystem (e.g., ~/projects/), not on the mounted Windows drive (/mnt/c/), for better performance and permission handling.