Repository navigation
Description
Decision required: yes
Risk: human-intent — two conflicting CI-gate approaches are both in flight and closing contributors' issues/PRs is a maintainer call.
Summary
Triage pass over the 23 open issues and 8 open PRs (#17025, #17151, #17152, #17220, #17222, #17226, #17229, #17230) in hivecommons/hive as of 2026-11-08. Found 2 near-duplicate pairs, 1 conflicting-PR pair, and 1 branch-twin merge-order note. No actions have been taken; this is a checklist for maintainers. (Open-PR search was returning empty from the index this pass; the PR set above comes from the hive snapshot plus issue timelines, so a very recent PR outside that set may be missing.)
Near-duplicates: close one
| Keep | Merge into it | Reason |
|---|---|---|
| #17221 crux-file-size-ratchet.yml pins a stale crux baseline | #17149 (same title, same finding) | Both report the same drift: workflow pins crux 13372ca7 whose baseline (1658/1389) is looser than crux main; both cite the same 2026-11-07 crux main ratchet break (crux#124) and #15331 as the origin. #17221 carries the full 7-file drift table and the git diff 13372ca7 7f931914 evidence, and quotes #15331's "crux is the source of truth" language; keep it. They differ only in proposed remedy (re-pin vs ref: main) — that choice is the PR decision below. |
| #17228 test (rest 3/4) failing across hive | #17227 [Tests] v5 test failure on 58f593d | Same failure: #17227's run 42480676503 @58f593d is the exact run #17228 cites as its v5 evidence (TestCloseOnMergeLedgerSweepSettledClaimsSkipAPI, retention time-bomb pruning the fixture's 2026-10-09 claims on reload). #17228 has the root cause and is covered by PRs #17229/#17230; #17227 is the auto-filed generic ci-failure issue with no cross-references. |
- [scanner] crux-file-size-ratchet.yml pins a stale crux baseline (1658/1389) looser than crux main (1657/1348) #17149 — close as duplicate of [ci-maintainer] crux-file-size-ratchet.yml pins a stale crux baseline — hive allowed agent/manager.go to 1658 while crux enforces 1657, breaking crux main for 4h on 11-07 #17221 (same stale-pin finding; remedy choice tracked by the PR pair below)
- [Tests] v5 test failure on 58f593d #17227 — close as duplicate of [shared-ci] test (rest 3/4) failing across hivecommons/hive #17228 (same run 42480676503, same head 58f593d, same test)
Conflicting PRs: merge order matters
- PR 🌱 ci: re-pin crux file-size ratchet to current crux main #17152 (re-pin to crux
fefb7a3, Fixes [scanner] crux-file-size-ratchet.yml pins a stale crux baseline (1658/1389) looser than crux main (1657/1348) #17149, heldsentinel-alert) and PR 🌱 ci: read the crux file-size ratchet baseline from crux main instead of a stale pin #17222 (read baseline from cruxmain, Closes [ci-maintainer] crux-file-size-ratchet.yml pins a stale crux baseline — hive allowed agent/manager.go to 1658 while crux enforces 1657, breaking crux main for 4h on 11-07 #17221) edit the same workflowcrux-file-size-ratchet.ymlwith opposite designs — both already carryconflicts-with-open-pr. Pick one: 🌱 ci: read the crux file-size ratchet baseline from crux main instead of a stale pin #17222 removes the drift class permanently (the baseline is supposed to move with crux); 🌱 ci: re-pin crux file-size ratchet to current crux main #17152 keeps the pinned-SHA design and will drift again on the next crux tightening. Whichever lands, close the other PR and its issue as superseded.
Related but distinct: cross-link only
- 🌱 test: load the reloaded sweep ledger with the test clock #17229 (base v5) and 🐛 [v6] test: load the reloaded sweep ledger with the test clock #17230 (base v6) — identical fix for [shared-ci] test (rest 3/4) failing across hivecommons/hive #17228 on both branches, not duplicates. Note the automated v5→v6 top-up (cf. 🌱 Forward-merge v5 into v6 (1 commits) #17025) carries v5 commits into v6: if 🌱 test: load the reloaded sweep ledger with the test clock #17229 merges first and the top-up picks it up before 🐛 [v6] test: load the reloaded sweep ledger with the test clock #17230, 🐛 [v6] test: load the reloaded sweep ledger with the test clock #17230 becomes redundant and can be closed; if v6 needs green now, merge both and let the top-up no-op.
- ✨ sidecar: /v1/knowledge/advise callback so sidecar hives get the mid-task adviser (advisor phase 2, slice 5) #17157 (sidecar
/v1/knowledge/advisecallback) and 📝 RFC: LLM-ranked knowledge advisor (OMP-style) #17054 (RFC: LLM-ranked knowledge advisor) — the sidecar callback is one delivery surface for the advisor epic, not a duplicate of it.
Checked and not duplicates
- [selfreport] unclassified_error_repeated: contributor-lane auto-promotion repo search failed; counting the PR's repositor… #17148 / PR 🐛 fix(automerge): isolate repos GitHub refuses to search in contributor auto-promotion #17151 — paired issue/PR; the PR's first commit says
Fixes #17148(branchfix/17148-autopromo-search-422), so the issue is referenced and will close on merge to v5; no closing keyword in the PR body, but no action needed. - 🐛 stage engine: proposal/kind-design issues read as Actionable after their RFC lands #17150 / PR 🐛 fix: unclaimed proposal / kind/design issues derive stage/018-open-for-discussion #17220 — PR deliberately uses
Refs #17150(AC5 — theproposallabel on 📝 RFC: LLM-ranked knowledge advisor (OMP-style) #17054 — stays open for a maintainer); correct as written. 🐛 fix: unclaimed proposal / kind/design issues derive stage/018-open-for-discussion #17220's "Supersedes 🐛 fix: unclaimed proposal / kind/design issues derive stage/018-open-for-discussion #17171" is settled: 🐛 fix: unclaimed proposal / kind/design issues derive stage/018-open-for-discussion #17171 is already closed. - [scanner] v6 readiness live exercise: chat surfaces (Slack, Discord, Matrix, Telegram) #9699 / [scanner] v6 readiness live exercise: escalation (email, push/on-call) and admin MCP endpoint #9701 / [scanner] v6 readiness live exercise: dashboard chat, inception via chat, standby S5-S6 #9735 — three v6-readiness live-exercise issues, deliberately split by surface (chat / escalation / dashboard chat), all held.
- [shared-ci] github.com release-download 504s (golangci-lint, docker, FOSSA, runtime-tests, typos) failing across hivecommons/hive #17224 (release-download 504s) vs [Tests] v5 test failure on 58f593d #17227/[shared-ci] test (rest 3/4) failing across hivecommons/hive #17228 — different failure class: tool-fetch 504s at install steps vs a deterministic test failure; not the same incident.
Filed by dedupe agent (issues-only mode). Recommendations only — nothing has been closed, labelled, or merged.
🐝 Hive Agent: dedupe | Instance: hosted-available-oke-11-placeholder-r05x | SHA: unknown
— hive: agent=dedupe backend=copilot model=claude-fable-5 copilot=1.0.88
Decision taken by hive: Hive did not park this: engineering decision taken by hive — [dedupe] triage: duplicate and overlapping open issues/PRs to consolidate. Owner may override with /hive hold or needs-human.
Activity
hivecommons-hive commentedon Nov 8, 2026
Hive did not park this: engineering decision taken by hive — [dedupe] triage: duplicate and overlapping open issues/PRs to consolidate. Owner may override with
/hive holdorneeds-human.hivecommons-hive commentedon Nov 8, 2026
Hive status — updated in place; edits do not notify. Full history is in the Activity Monitor.
Parked on
stage/045-needs-decision— Risk: human-intent. The checklist's remaining actions are maintainer calls, not hive work:main) be the design — close 🌱 ci: re-pin crux file-size ratchet to current crux main #17152 as superseded.Already resolved without a decision: #17227 and #17232 are the same
TestCloseOnMergeLedgerSweepSettledClaimsSkipAPIincident as #17228; #17229 carriesCloses #17227/Closes #17228and #17232 is closed as a duplicate. Note #17222 is itself already parked onstage/085-needs-decision, so one ruling there settles the PR pair here too.— hive: agent=scanner backend=copilot model=claude-fable-5.1 copilot=1.0.88
hivecommons-hive commentedon Nov 8, 2026
Scanner status on the checklist (2026-11-08 14:xx UTC):
test (rest 3/4)pair is settled.crux-file-size-ratchet.yml— 🌱 ci: re-pin crux file-size ratchet to current crux main #17152 (@Karibhou, pinned-SHA re-pin, heldsentinel-alert) and 🌱 ci: read the crux file-size ratchet baseline from crux main instead of a stale pin #17222 (hive, read baseline from cruxmain, parkedstage/085-needs-decision). Closing a contributor's PR is the maintainer call already parked on 🌱 ci: read the crux file-size ratchet baseline from crux main instead of a stale pin #17222, so the hive is not closing [scanner] crux-file-size-ratchet.yml pins a stale crux baseline (1658/1389) looser than crux main (1657/1348) #17149 ahead of that ruling; whichever PR lands should close both the other PR and its issue as superseded.Nothing else on the list is hive-actionable.
🐝 Hive Agent:
scanner| Instance:hosted-available-oke-11-placeholder-r05x| SHA:unknown— hive: agent=scanner backend=copilot model=claude-fable-5.1 copilot=1.0.88
hivecommons-hive commentedon Nov 8, 2026
Closed as read (advisory acknowledged) by @clubanderson from the Hive dashboard.