governed is early, and its whole point is that you should not have to take the author's word for it. If you have a spare hour, I would value an honest outside look. Specific things I want to know:
Does the clean install and the 60-second example work on your machine? Please say which OS and Python version.
Do the contracts actually catch a misbehaving agent, or can you get it to act without a plan?
Is the trace readable and useful when you go back to understand a run?
Is the threat model in docs/RESPONSIBLE_AI.md sound, or am I overclaiming anywhere?
Drop your notes in a comment here, or open a separate issue if you find a bug. No finding is too small.
governed is early, and its whole point is that you should not have to take the author's word for it. If you have a spare hour, I would value an honest outside look. Specific things I want to know:
Does the clean install and the 60-second example work on your machine? Please say which OS and Python version.
Do the contracts actually catch a misbehaving agent, or can you get it to act without a plan?
Is the trace readable and useful when you go back to understand a run?
Is the threat model in docs/RESPONSIBLE_AI.md sound, or am I overclaiming anywhere?
Drop your notes in a comment here, or open a separate issue if you find a bug. No finding is too small.