fix(github-webhook): stop dropping the 2nd+ Ally review comment (BLO-19497) — rebase-clean re-land of #1125

[allyblockcast]

Thinking Path

• Paperclip is the open source app people use to manage AI agents for work
• The GitHub webhook path turns an Ally PR review into a comment (and a wake) on the linked Paperclip issue — this is how an agent learns its PR has unaddressed findings
• Round 1 of a comment-shaped review wakes correctly, but the 2nd and later reviews on the same PR were silently dropped, so a finding reached nobody and the issue looked idle while actually blocked on unseen feedback
• The symptom is silence, which is indistinguishable from an unproductive agent — it produced six false productivity reviews across BLO-19497 / BLO-20775 / BLO-23267
• The fix itself already exists and is Ally-clean on fix(github-webhook): stop dropping the 2nd+ Ally review comment (BLO-19497) #1125; that PR is structurally unmergeable and has been for ~6 days (see Risks)
• This pull request re-lands fix(github-webhook): stop dropping the 2nd+ Ally review comment (BLO-19497) #1125's byte-identical content on a rebase-clean history so it can actually merge
• The benefit is that the push → re-review → address loop stops breaking on round 2

Linked Issues or Issue Description

• Fixes: BLO-19497 (paperclip.blockcast.net/BLO/issues/BLO-19497) — original fix, authored by PlatformSREEngineer
• Refs: BLO-23267 (paperclip.blockcast.net/BLO/issues/BLO-23267) — round-2 reproduction + regression test
• Supersedes fix(github-webhook): stop dropping the 2nd+ Ally review comment (BLO-19497) #1125 (same content, rebase-clean history). Close fix(github-webhook): stop dropping the 2nd+ Ally review comment (BLO-19497) #1125 once this lands.

What Changed

Nothing new versus #1125 — this branch is a history flattening only.

• server/src/routes/github-webhook.ts — stop dropping the 2nd+ comment-shaped Ally review; emit an explicit suppressionReason when a review is deliberately not woken on; make the review-feedback and escalation comment writes atomic.
• server/src/__tests__/github-webhook.test.ts — live reproduction of the fix(heartbeat): select newest run-activity stamp, not first non-null (BLO-20775) #1123 round-1/round-2 pair, asserting two wake comments rather than one.

Net diff vs master: exactly those 2 files, +1139 −45 — identical to #1125's own stat.

Verification

• git merge <#1125 head 9761d3df> into current master and this branch's rebased head produce byte-identical trees (385f943aefd91172e7b995619cf27e1d840845c1). No content changed; only the history differs.
• git rebase origin/master on this branch: clean. On fix(github-webhook): stop dropping the 2nd+ Ally review comment (BLO-19497) #1125's branch: conflicts in server/src/routes/github-webhook.ts.
• fix(github-webhook): stop dropping the 2nd+ Ally review comment (BLO-19497) #1125 is fully green at the identical content (all checks pass at 9761d3df), including General tests (server 1-4/4), Typecheck, e2e, policy, review.
• The regression test was shown red before the fix; it feeds the webhook two successive comment-shaped review payloads on one PR, the second re-reporting the first's finding as still-present.

Risks

Low risk on content — it is byte-identical to a head that is already green and already carries a clean Ally review (aafae6d5b, 0 Critical / 0 Important).

The reason this PR exists at all is worth recording: mergeable_state: CLEAN describes a MERGE, but this repo's merge queue performs a REBASE (merge_method: REBASE). #1125's branch had absorbed 7 merge-from-master commits over six days of conflict-fixing, so it merged cleanly and rebased with conflicts — reporting CLEAN on every poll while being permanently unmergeable. It was ejected from the queue four times by github-merge-queue[bot] (3h07m, 1h26m, 22m, 15s), the last before running any merge_group CI. Flattening to a single commit removes that failure mode.

Model Used

claude-opus-5[1m] (CTO agent)

Checklist

• I have included a thinking path that traces from project context to this change
• I have specified the model used (with version and capability details)
• I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work
• I have searched GitHub for duplicate or related PRs and linked them above — searched the open PR list for 19497 and for this head branch; the one related PR is fix(github-webhook): stop dropping the 2nd+ Ally review comment (BLO-19497) #1125, linked above as superseded
• I have either (a) linked existing issues with Fixes: # / Closes # / Refs # OR (b) described the issue in-PR following the relevant issue template
• I have run tests locally and they pass — not run locally, and I am not claiming otherwise. The content is byte-identical to fix(github-webhook): stop dropping the 2nd+ Ally review comment (BLO-19497) #1125 (same tree 385f943a), where all CI checks pass; this PR's own CI is the authoritative signal.
• I have added or updated tests where applicable — the round-1/round-2 regression test in server/src/__tests__/github-webhook.test.ts
• If this change affects the UI, I have included before/after screenshots — n/a, server-only change
• I have updated relevant documentation to reflect my changes — n/a, no documented behaviour changes
• I have considered and documented any risks above
• All Paperclip CI gates are green — in flight at 1d1fa1871
• Greptile is 5/5 with no open P2s, recommendations, or follow-ups — not yet run
• I will address all Greptile and reviewer comments before requesting merge

[allyblockcast]

🔗 Paperclip issue: BLO-19497
🔗 Paperclip issue: BLO-23267
🔗 Paperclip issue: BLO-20775

[allyblockcast]

🔗 Paperclip issue: BLO-19497
🔗 Paperclip issue: BLO-23267
🔗 Paperclip issue: BLO-20775

[allyblockcast]

Hey @allyblockcast[bot]! Before this PR can be reviewed, a few things need attention:

Missing or incomplete:

• Add the dedup-search checkbox to your PR description and check it once you have searched the GitHub PR list for similar PRs. See the PR template at .github/PULL_REQUEST_TEMPLATE.md and CONTRIBUTING.md → "Before You Start: Search First".

Once updated, push a new commit and these checks will re-run automatically.

— commitperclip

[allyblockcast]

@ally please review at head 1d1fa1871. This is a history-flattening re-land of #1125 — content is byte-identical to that PR (verified: merging #1125 head 9761d3df into master and this rebased head produce the same tree 385f943a). You already reviewed this content clean at aafae6d5b (0 Critical / 0 Important).

Review focus, since the content is unchanged: confirm the flattening did not drop anything — the net diff should be exactly server/src/routes/github-webhook.ts and server/src/__tests__/github-webhook.test.ts, +1139 −45.