can you do me an aggregation over the last month across all PRs about how reviewed how many PRs?

Here's the breakdown for entireio/cli, reviews submitted 2026-05-30 → 2026-06-30 (last month):

Reviewer	PRs reviewed	Review events
copilot-pull-request-reviewer 🤖	210	232
Soph	113	172
cursor 🤖	87	131
pfleidi	47	94
toothbrush	36	110
pjbgf	35	53
computermode	29	51
dipree	16	20
nodo	8	10
peyton-alt	7	14
alishakawaguchi	7	13
claude 🤖	5	5
khaong	3	9
evisdren / squishykid	3	3–4
paxos / evjan / stiak	2	each
blackgirlbytes, suhaanthayyil, ashtom, stale2000, dvydra	1	each

Totals: 255 distinct PRs received at least one review; 951 review events overall.

Notes:

• PRs reviewed = distinct PRs that person submitted at least one review on. Review events = individual review submissions (a single PR often gets multiple from one reviewer across re-review rounds), which is why events ≫ PRs.
• Three reviewers are bots/automation (🤖): Copilot, Cursor, and claude. Excluding those, the top human reviewers are Soph (113), pfleidi (47), toothbrush (36), pjbgf (35), and computermode (29).
• Window is keyed on each review's submittedAt, so this counts review activity in the last month regardless of when the PR was opened.

Want me to split this differently — e.g. humans-only, approvals vs. change-requests, or count only PRs merged in the window?

A review event is one submitted review on a PR — the thing that happens when someone clicks "Approve," "Request changes," or "Comment" in GitHub's review tab and submits it. Each click-and-submit is one event.

The key distinction from "PRs reviewed":

• PRs reviewed counts distinct PRs a person touched — PR #500 counts once for you no matter how many times you reviewed it.
• Review events counts each submission — if you review PR #500, the author pushes changes, and you review again to approve, that's 2 review events on 1 PR.

That's why events are always ≥ PRs reviewed, and the gap tells you how much back-and-forth re-review happened. For example, toothbrush has 36 PRs but 110 events — averaging ~3 review rounds per PR, lots of iteration. dipree has 16 PRs and 20 events — mostly one-and-done reviews.

One caveat: a review event is not the same as a single inline comment. If you leave 8 line comments and then hit "Submit review," that bundles into one event. A standalone comment posted outside the review flow isn't counted here at all — this only counts formal review submissions (which carry a state: APPROVED, CHANGES_REQUESTED, or COMMENTED).

Want me to break the events down by that state, so you can see approvals vs. change-requests vs. plain comments per reviewer?

Here's each reviewer with the top authors whose PRs they reviewed most (distinct PRs). Self-reviews are now excluded, which is why some totals dropped from the first table and a few reviewers (e.g. khaong, ashtom) disappeared — their only reviews were on their own PRs.

Reviewer	PRs	Events	Top authors reviewed (PRs)
copilot 🤖	210	232	Soph (52), toothbrush (37), dipree (20)
Soph	98	122	dependabot (20), toothbrush (20), dipree (18)
cursor 🤖	87	131	toothbrush (17), pfleidi (14), Soph (12)
pfleidi	34	48	Soph (13), computermode (12), toothbrush (3)
pjbgf	33	50	Soph (22), toothbrush (6), pfleidi (2)
computermode	18	23	pfleidi (12), alishakawaguchi (2), peyton-alt (1)
toothbrush	12	17	entire (4), Soph (4), stiak (3)
dipree	10	11	Soph (8), toothbrush (2)
nodo	8	10	toothbrush (4), Soph (3), pjbgf (1)
claude 🤖	5	5	suhaanthayyil (3), MuskanPaliwal (1), stale2000 (1)
alishakawaguchi	5	7	suhaanthayyil (1), computermode (1), evisdren (1)
peyton-alt	4	6	Soph (2), dipree (1), khaong (1)
evisdren	3	3	pfleidi (1), dipree (1), computermode (1)
squishykid	3	4	Soph (2), pjbgf (1)
blackgirlbytes	1	6	suhaanthayyil (1)
paxos / stiak / dvydra / evjan	1	1–2	khaong / toothbrush / toothbrush / dipree (1)

A few patterns worth noting:

• pjbgf → Soph is the strongest single pairing among humans (22 of 33 PRs).
• computermode ↔ pfleidi is a tight mutual pair: pfleidi reviewed 12 of computermode's PRs, and computermode reviewed 12 of pfleidi's.
• Soph is the broadest reviewer — top targets are dependabot (20, mostly dep bumps) and toothbrush (20), then dipree (18).

Want me to drop the bots and dependabot PRs to get a cleaner human-to-human review map, or show this as a directed matrix/heatmap?

Confirmed. The split:

• On their own PRs: 24 PRs, 93 events — all COMMENTED (zero approvals/change-requests, which makes sense: GitHub doesn't let you approve your own PR).
• On others' PRs: 12 PRs, 17 events.
• Total: 36 PRs, 110 events — exactly the first table.

So the first table (36) counted all of toothbrush's review submissions, including the 24 PRs where toothbrush was the author replying to reviewers via "Submit review" comments. The second table excludes those self-reviews, leaving the 12 PRs that were actually someone else's work — which is the more meaningful "reviewing for others" number.

This is also why toothbrush's events-per-PR looked so high in the first table (110/36 ≈ 3): those 93 self-comments are the back-and-forth of an author answering review threads, not review effort on other people's code.

If you want, I can regenerate the first table with self-reviews excluded too, so both tables are consistent on the "reviews for others" definition.

Yes — done. Scanned 56 active repos across both orgs (41 had PRs in the window), reviews submitted since 2026-05-30, self-reviews excluded:

Reviewer	PRs	Events	Top authors reviewed (PRs)
cursor 🤖	638	1138	evisdren (115), toothbrush (66), jagregory (66)
copilot 🤖	238	261	Soph (59), toothbrush (38), dipree (20)
Soph	118	142	dipree (23), toothbrush (21), dependabot (20)
pjbgf	53	76	Soph (34), toothbrush (6), nodo (5)
dipree	43	44	evjan (16), Soph (14), matthiaswenz (5)
pfleidi	38	52	Soph (14), computermode (13), toothbrush (3)
nodo	31	37	Soph (22), toothbrush (6), pjbgf (2)
toothbrush	26	35	stiak (8), Soph (6), entire (4)
computermode	21	26	pfleidi (12), alishakawaguchi (3), paxos (2)
evjan	21	25	dipree (16), alishakawaguchi (2), matthiaswenz (1)
matthiaswenz	20	22	dipree (6), Soph (4), evjan (3)
evisdren	19	20	alishakawaguchi (15), pfleidi (1), dipree (1)
paxos	11	12	khaong (3), dipree (3), matthiaswenz (2)
jagregory	11	11	Soph (4), toothbrush (2), dipree (1)
alishakawaguchi	9	12	evisdren (3), suhaanthayyil (1), computermode (1)
georg	7	7	stiak (3), jdx (2), toothbrush (1)
khaong	6	11	toothbrush (2), nodo (2), Soph (1)
claude 🤖	5	5	suhaanthayyil (3), MuskanPaliwal (1), stale2000 (1)
stiak	5	6	toothbrush (2), khaong (1), Soph (1)
peyton-alt / squishykid / blackgirlbytes / dvydra	2–4	—	(tail)
suhaanthayyil / jakub-nespor / kramuenke / ashtom	1	—	(tail)

Notes & caveats:

• Top human reviewers org-wide: Soph (118), pjbgf (53), dipree (43), pfleidi (38), nodo (31).
• Cursor dominates (638 PRs) — it's auto-reviewing across many repos, heavily on evisdren's work. Copilot is second.
• Some strong human pairings cross repos: evjan↔dipree (evjan reviewed 16 of dipree's, dipree reviewed 16 of evjan's), computermode↔pfleidi (12 each way), evisdren→alishakawaguchi (15).
• Filter: I only scanned repos pushed within the window (56 of 116). A repo with PR reviews but no branch pushes since 2026-05-30 would be missed — unlikely, but I can scan all 116 repos if you want it exhaustive.
• Anything I lack access to was skipped silently (none errored loudly here).

Want a humans-only version, a per-org split, or a directed reviewer→author matrix/heatmap?