You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: .agents/skills/babysit/SKILL.md
+6-9Lines changed: 6 additions & 9 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -1,6 +1,6 @@
1
1
---
2
2
name: babysit
3
-
description: Drive a PR to a clean review (Greptile 5/5, zero open threads) — ships if needed, keeps it mergeable against staging, triggers Greptile/Cursor Bugbot, fixes real findings, replies to and resolves every thread, and loops until clean
3
+
description: Drive a PR to a clean review (Greptile 5/5, zero open threads) — ships if needed, keeps it mergeable against staging, triggers Greptile, fixes real findings, replies to and resolves every thread, and loops until clean
4
4
---
5
5
6
6
# Babysit PRs
@@ -58,9 +58,9 @@ round. Always check both conditions freshly after every push.
58
58
2.**If the PR has a merge conflict**, merge `origin/staging`, resolve the conflicts, run the
59
59
usual pre-push checks, push, and go to step 8 to re-trigger review.
60
60
61
-
3.**If no review has run yet** (fresh PR, no Greptile/Cursor comments): they usually run
62
-
automatically on PR open — confirm via `gh pr checks <n>` (look for `Cursor Bugbot` /
63
-
`Greptile Review`) and wait for that first round before doing anything else.
61
+
3.**If no review has run yet** (fresh PR, no Greptile comments): Greptile usually runs
62
+
automatically on PR open — confirm via `gh pr checks <n>` (look for `Greptile Review`) and
63
+
wait for that first round before doing anything else.
64
64
65
65
4.**If a review round has landed and it isn't clean**: for every thread where
66
66
`isResolved: false`, triage the finding on its own merits — this is the part that requires
@@ -113,15 +113,13 @@ round. Always check both conditions freshly after every push.
113
113
rounds; checking sync only before the push (step 6) and never after is how a bad push or a
114
114
PR whose commit history quietly went stale between rounds goes unnoticed.
115
115
116
-
8.**Re-trigger review** by posting `@greptile` and `@cursor review` as **two separate PR
117
-
comments** — never combine them into one comment, each bot only responds to its own mention:
116
+
8.**Re-trigger review** by posting `@greptile` as its own PR comment:
118
117
```bash
119
118
gh pr comment <n> --body "@greptile"
120
-
gh pr comment <n> --body "@cursor review"
121
119
```
122
120
123
121
9.**Wait for the new round**, then go back to step 1. Pace the wait with `ScheduleWakeup` using
124
-
a fallback delay of ~250–300s (Greptile/Cursor typically take 1–3 minutes) — never busy-poll
122
+
a fallback delay of ~250–300s (Greptile typically takes 1–3 minutes) — never busy-poll
125
123
in a sleep loop. Pass the same `/loop babysit PR <n>` prompt on each wakeup so the loop
126
124
resumes correctly.
127
125
@@ -147,7 +145,6 @@ notification email.
147
145
148
146
## Hard rules
149
147
150
-
- Never post the two re-review mentions as a single combined comment.
151
148
- Never paste prod evidence into a reply without scrubbing it first (see above).
152
149
- Never resolve a thread without replying to it first.
153
150
- Never fix a finding with a hacky workaround — if the clean fix isn't obvious, find the sibling
0 commit comments