Repository navigation
Description
Summary
Every scanner kick since 2026-11-05T00:04Z has listed the same five contributor PRs as MERGE-ELIGIBLE (CI green, hive-reviewed approve), and every hive-merge request for them has been refused with the same policy error:
authorization denied: contributor-lane policy refuses the merge relay for @thatguywob's PR:
contributor-lane:no-claim (author tier "contributor", require_tiers merger, trusted)
Affected PRs (15+ consecutive refusals each, 2026-11-05 00:04Z → 04:12Z):
- 🐛 fix(intake): restrict extraction worker environment dibs#602 — 🐛 fix(intake): restrict extraction worker environment
- 🐛 fix(settle): replace stale trailing markers despite matching draft text dibs#603 — 🐛 fix(settle): replace stale trailing markers despite matching draft text
- 🐛 fix: keep superseder decision questions parked #16687 — 🐛 fix: keep superseder decision questions parked
- 🐛 fix(notify): back off ntfy sends after rate limiting #16691 — 🐛 fix(notify): back off ntfy sends after rate limiting (closes [selfreport] unclassified_error_repeated: ntfy returned error #16689, and [selfreport] unclassified_error_repeated: ntfy errors suppressed #16708 depends on it)
- ✨ feat(settings): expose merge-train controls in Auto-merge #16680 — ✨ feat(settings): expose merge-train controls in Auto-merge
Why this is one incident, not five
The governor's merge-eligibility computation and the merge relay's contributor-lane policy disagree about the same class of PR: approved, CI-green PRs from a contributor-tier author whose relay record has no claim. The governor keeps dispatching them to the merge sweep; the relay keeps refusing; each kick burns five merge requests and a bead and nothing changes. Fixes already in flight (e.g. #16691 for the ntfy 429 selfreports #16689/#16708) sit unmerged behind it.
Asks (pick whichever is the intended behaviour)
- Eligibility side: exclude PRs the relay will refuse (
contributortier without a claim / not inrequire_tiers) from the MERGE-ELIGIBLE list, and surface them in a distinct "needs maintainer merge" section so lanes stop retrying. - Relay side: if an approved, CI-green contributor PR that
hive-reviewapproved should be mergeable, treat the review approval (or a linkedCloses #Non an issue the author held) as satisfyingno-claim, or document the/claimstep the author must take and have the relay say so in its refusal. - Either way, the relay refusal should post one comment on the PR naming what unblocks it; today the refusal is visible only in the merge-request result file.
Decision taken by hive: filing as a tracker for the shared root cause rather than re-recording the refusal per kick; the lane will stop retrying these five until this closes or the policy/eligibility changes.
Evidence
/var/run/hive-metrics/merge-requests/scanner-1793851945662442632.result.json(dibs#602, 2026-11-05T04:12:28Z) and sibling results for the other four this kick, allok:falsewith the identical error above.- Scanner beads 2026-11-05 00:04Z, 00:18Z, 01:21Z, 01:35Z, 01:47Z, 02:05Z, 02:21Z, 02:29Z, 02:59Z, 03:13Z, 03:28Z, 03:38Z, 03:48Z, 03:57Z, 04:06Z — same five PRs, same error.
— hive: agent=scanner backend=copilot model=claude-fable-5.1 copilot=1.0.88
Activity
hivecommons-hive commentedon Nov 5, 2026
🔓 Claim by @thatguywob released (lease revoked).
Claim history
thatguywob(contributor on hive hosted-available-oke-11-placeholder-r05x)thatguywob(contributor; lease revoked)🐛 fix(automerge): align contributor merge eligibility with relay poli…