Flag users with large numbers of contributions in small time periods

[melissawm]

Hi folks, thank you so much for this! ❤️

One pattern I've been observing in a few projects for a few weeks is users who are not necessarily easily flagged as bots, but who may be using automation on their side to create a huge amount of PRs in a very small time period - something like >300 PRs in >200 repos, over 24h.

For example:

I'd love to be able to have a column/tag for that in the triage board. This would only run for first-time contributors, and would not necessarily be exposed in the PR itself, only as part of the board metadata, showing that this user has made a significant number of PRs in a short period of time across many different repos, which may be a sign of automation and a violation of many projects' AI/automation policy.

I see from your principles:

2. The calculation of the values for Project Fields should be all deterministic, so they are stable between runs as long as the PR itself doesn't change.

I am not sure if this suggestion would imply a violation of this principle, since this number is external (user may keep creating more PRs after the number is first computed, which is admittedly a limitation of this approach).

Is that an acceptable contribution to this bot, or is that something I should implement on my own board? If it is acceptable, I'd be happy to try my hand at a PR.

Thanks again!

[yuvipanda]

Thank you for your thoughtful issue here, @melissawm! I would love a PR for this :)

I opened #55 to clarify the 'deterministic' principle with a few more examples. We already move things around because of changes outside the PR, and I've more explicitly counted that there. The intent behind the principle was to avoid doing probabilistic labeling work (by asking an LLM, for example), as that can change wildly without real explanation.

I'm sorry for the late reply here!