Tom does the math on /dev/null
Tag the bot in a public post and it now answers with contributor rankings, production merges, stars, forks and member tallies. Every number comes from public GitHub data or public profiles, stamped with when it was read.
You tag @tom in a public post and ask who has landed the most commits, or how many people joined this week. Back comes a short reply with a ranked list, a star and fork count, or a member total, each saying when it was read.
Credit goes to paulinevos, who gave Tom two read-only tools. On the GitHub side Tom lists up to ten contributors, bots excluded, ranked by GitHub's all-time commit count, plus ten recent merges into production from the last three UTC days, ten open issues labelled good first issue or help wanted, stars, forks, separate open issue and pull request counts, and code sizes in bytes. On the community side he counts onboarded members, new members and members who posted publicly over 7 or 30 days, and matches location or one of 17 languages in public profiles. GitHub answers sit in a 15-minute cache. Member answers last 5 minutes. Tom gets two calls per tool per reply.
There is one catch. Tom counts profile text, not people. Ask about Portugal and a member whose location reads only "Lisbon" is not in the tally. Ask about Go and he wants "Golang developer" or "Go engineer". A bare "Go" stays out.