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
{{ message }}
Repository navigation
🐛 contributor lane: relay fix-pr task fails instantly, hive logs the nudge as delivered and never retries; stale unauthorized pill hides the real blocker #14668
A PR returned to a relay contributor gets a fix-pr task that the relay fails instantly; the hive still logs the nudge as delivered (mode=spawn), posts the "back in your court" comment, and never retries or falls back. The PR then sits at stage/072-returned-to-contributor (24h dwell) with a stale hive:unauthorized pill that names a blocker that is no longer the one stopping it.
Nothing further for #397 in the next 40 min: no re-enqueue, no second nudge, no adoption (adoption due but the author maintains the repository; not adopting fires only for #14524), no comment telling the contributor the relay task failed. The relay failure reason is not logged either.
Current state (18:25Z): mergeable=CONFLICTING/DIRTY, labels stage/072-returned-to-contributor + hive:unauthorized; head e3e9f5c still has 3 action_required runs (CI, Copilot DCO, Mutation) — so the unauthorized marker is factually true, but the row's lead pill tells the owner "approve runs" when the thing that actually moved the PR to 072 is the merge conflict, which only the contributor can fix. Approving the held runs now would be wasted: a rebase push will create new held runs.
Expected
fix-pr task failed must be treated as a dispatch failure: re-queue with backoff (relay may be mid-turn), surface the relay's error, and after N failures fall back to the hive's own path (adopt / rehome when registry-trusted) or post one comment telling the contributor the relay could not take the task.
stage labels: contributor nudged … mode=spawn should only be logged/commented when the task was accepted by the relay, not when the dispatch already failed.
Pill precedence on 072: when the 07x reason is conflicting, the row's lead pill should be conflicting/needs rebase, and unauthorized should fold into the "+1" tooltip (or be cleared and re-marked on the next head) — it is not the current blocker.
Related: #14589 / PR #14595 introduced the unauthorized marker (auto-clears only on successful approval). #14607 (conflicting masks decision_requested) is about 070, not the relay path.
🔒 Claimed by agent scanner (agent on hive hosted-available-oke-11-placeholder-r05x). The claim lasts 2h0m0s without activity and renews while the work continues. Reply /unclaim to release, or /claim to take it over if you outrank the holder.
@clubanderson this issue was closed by @hivecommons-hive[bot] as completed.
If this was closed by a merged PR, the fix only takes effect once it ships in a tagged release and your deployment updates to it — check your running version against the fix before concluding the fix failed.
If you are still seeing the problem, comment /reopen and it will be reopened for you — you do not need write access. Please add what you are still observing (running version, logs, steps) so it is not closed the same way again.
Activity
hivecommons-hive commentedon Oct 27, 2026
🔒 Claimed by agent
scanner(agent on hive hosted-available-oke-11-placeholder-r05x). The claim lasts 2h0m0s without activity and renews while the work continues. Reply/unclaimto release, or/claimto take it over if you outrank the holder.github-actions commentedon Oct 27, 2026
@clubanderson this issue was closed by @hivecommons-hive[bot] as completed.
If this was closed by a merged PR, the fix only takes effect once it ships in a tagged release and your deployment updates to it — check your running version against the fix before concluding the fix failed.
If you are still seeing the problem, comment
/reopenand it will be reopened for you — you do not need write access. Please add what you are still observing (running version, logs, steps) so it is not closed the same way again.