#What is our stance on fully automated
1 messages · Page 1 of 1 (latest)
I think we should require a human in the loop for all PRs.
We want to talk to a human when reviewing PRs.
I think we should adopt something similar as this:
https://github.com/astral-sh/.github/blob/main/AI_POLICY.md
I like that very much
I'd agree with that, frontend too. I'd also like the bot/copilot to check if the template is followed and has been completed by a human - have a list of known bots, not foolproof but still something
Seems like this claw-explorer bot is now hanging around in integrations starting with A
do we want to keep it around or block it for now
block it?
block the account
like if we don't want AI generated contributions without a human in the loop, should we keep it open
I say no.
I meant "block it."
there are still a couple claw-explorer PRs open, should I close them?
I reckon we could programatically do the PR template checks with a bot. similar to the cla/stale bot. Check if the template has been followed, add a label (so we can filter these) and draft the PR. Give a shorter time to update, then close if not compliant then.
Probably something core, frontend. home-assistant.io etc could all benefit from
I opened a PR for an AI policy (parts taken from astral, because I like that one) https://github.com/home-assistant/core/pull/167047