@feigi/fleet-ctl 3.47.3 → 3.47.5

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -1,5 +1,5 @@
1
1
  ---
2
- description: review a PR, apply recommended actions, push, watch checks until green
2
+ description: review a PR, apply recommended actions, push — standalone, then watch checks until green; under the fleet, report and stop (the controller owns the CI wait)
3
3
  argument-hint: [pr-number]
4
4
  ---
5
5
 
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@feigi/fleet-ctl",
3
- "version": "3.47.3",
3
+ "version": "3.47.5",
4
4
  "description": "Agent fleet: run-team controller, merge bot, PR reviewer, and the ticket pipeline they share",
5
5
  "license": "Apache-2.0",
6
6
  "repository": {
@@ -1761,7 +1761,9 @@ owes beyond its row:
1761
1761
  runner — **Reviewers** says which, per harness) and record `reviewed=` off its
1762
1762
  digest with `ledger.mjs review <M>`, then run the tick: a PR with survived or unverified findings prints as
1763
1763
  `DISPATCH fix-pr PR#<M>`. No survived or unverified finding and nothing to file → skip the
1764
- fix-applier and dispatch the **finisher** directly: no member will touch that
1764
+ fix-applier and dispatch the **finisher** directly — or, when that head's
1765
+ bound conclusion is red and not solely staleness, the red-check fixer
1766
+ instead (**Then dispatch a red-check fixer**, below): no member will touch that
1765
1767
  PR, so no push is coming and nothing will wake you. A review that failed is
1766
1768
  not a result — **Reviewers** says who retries it and what takes over.
1767
1769
  - **Monitor: `ready-to-merge` appears** → nothing to record; run the tick, and
@@ -1812,7 +1814,9 @@ owes beyond its row:
1812
1814
  place. The diff-validating `check` job green with no heavy job (the
1813
1815
  diff-validating suites, not the `rebase-check` currency gate) in `failure`
1814
1816
  → a finisher CANDIDATE (gated below), a `check`
1815
- **failure** → a fixer. A `check`-green board whose heavy jobs are merely
1817
+ **failure** that is not solely staleness → the **red-check fixer** (**Then
1818
+ dispatch a red-check fixer**, below), behind the same review-returned and
1819
+ fix-applier-reported gate as a finisher. A `check`-green board whose heavy jobs are merely
1816
1820
  `skipped` (behind-count staleness, the normal case after a merge) still labels — do NOT
1817
1821
  gate on `ci-state --quiet` exit 0, which a behind PR never reaches. **This
1818
1822
  edge fires for every CI run the PR produces, and the implementer's own
@@ -1835,10 +1839,15 @@ owes beyond its row:
1835
1839
  first — including any ruling you have withdrawn or reversed — then dispatch
1836
1840
  (below); a final report is not proof it stopped. A PR that needs nothing fixed
1837
1841
  reaches a finisher through the **fix-applier reports `no-op`** bullet below,
1838
- never by waiting for a second CI event this head will never produce.
1842
+ never by waiting for a second CI event this head will never produce. The
1843
+ same goes for a red one: a red that fired before the review never fires
1844
+ again, so the **fix-applier reports `no-op`** and **review result lands**
1845
+ bullets are also the only routes to the red-check fixer there.
1839
1846
  Then run the tick.
1840
1847
  - **A fix-applier reports `no-op`, or a SHA you have already bound** → dispatch
1841
- the finisher **now**, against the existing head. No push means no new run, and
1848
+ the finisher **now**, against the existing head — when that head's bound
1849
+ conclusion is red, the red-check fixer instead (**Then dispatch a red-check
1850
+ fixer**, below), never a finisher and never a wait. No push means no new run, and
1842
1851
  the Monitor above is edge-keyed on `<run-id>:<attempt>:<conclusion>` — that
1843
1852
  head's terminal state already fired once and will never fire again, so waiting
1844
1853
  for the CI event waits forever. This stays a common outcome on a clean PR:
@@ -2559,6 +2568,46 @@ is a review fix-applier, dispatched as above. On a PR carrying
2559
2568
  bot refuses it `head-moved-after-label-#<M>`: a fresh finisher, per **Failure
2560
2569
  handling**.
2561
2570
 
2571
+ **Then dispatch a red-check fixer** — on a `check`-job `failure` bound to the
2572
+ PR's current head that is not **solely** staleness: a `rebase-check` red, or
2573
+ heavy jobs `skipped`, that comes only off the behind-count gets no fixer (the
2574
+ five-condition note in the fix-applier block below states how to establish
2575
+ that). The tick prints no `DISPATCH` for it, so you reconcile it yourself, and
2576
+ it always goes after the review, never before. Take the common case, the
2577
+ implementer's own first CI firing turning red before any review exists: the
2578
+ review is still dispatched against that head as normal, because a red suite is
2579
+ review evidence and not a blocker. While a review or a review fix-applier on
2580
+ the PR is live or still owes its report — a dispositions mismatch awaiting its
2581
+ retry and a conflict hold awaiting its fix-applier included — the red-check
2582
+ fixer does not fire: their push starts a new run, and a fixer racing them
2583
+ would work on a SHA that is still moving. These edges dispatch it, each the
2584
+ finisher's own: **Monitor: CI run completes** binding a red on a head
2585
+ whose review has returned and whose fix-applier, if any, has reported;
2586
+ **A fix-applier reports `no-op`**, or a SHA you have already bound, when that
2587
+ head's bound conclusion is red; and **A review result lands** with no finding
2588
+ and nothing to file, when that head's bound conclusion is red. The Monitor is
2589
+ edge-keyed on `<run-id>:<attempt>:<conclusion>`, so a red that fired before the
2590
+ review never fires again, and with no push there is no new run — the other
2591
+ edges are the only route there, so do not wait for an event. Edit one of these
2592
+ bullets without the others and what strands is exactly the red-only PR.
2593
+
2594
+ Name it `fix-pr-<M>`, or the next free suffix (`-b`, `-c` …) — `dispatch`
2595
+ refuses a reused name — record it with `ledger.mjs dispatch <pr#> <name>`
2596
+ before the call, and dispatch it with the `agent` that command printed (`null`
2597
+ while no conflict hold stands, as for a review fix-applier), in the PR's
2598
+ existing worktree. Its prompt takes the prompt below's worktree, branch and
2599
+ path discipline, but in place of its review paragraphs: read the failing job's
2600
+ log, fix the cause, run `<testCmd>`, push, report the new head, and **stop**.
2601
+ It is told outright not to watch, poll or wait for CI: your persistent Monitor
2602
+ owns the next run on the pushed head, and the finisher gate — review returned
2603
+ AND fix-applier report sent — applies to this member the same way. Settle it
2604
+ `applied:<head>` once a fix is pushed. Settle it `no-op` only when the red
2605
+ turns out not to be this member's to fix — a flake that passes on rerun, or
2606
+ staleness — with nothing pushed; `no-op` counts as landed, so it never stands
2607
+ for "could not fix". Settle it `failed` when it cannot reach green, which the
2608
+ **Failure handling** row "Reviewer, fix-applier or red-check fixer cannot
2609
+ reach green" covers.
2610
+
2562
2611
  **A `DISPATCH fix-pr PR#<M>` on a dispositions mismatch** — the PR's row carries
2563
2612
  `dispositions-mismatch=<member>:<run>` for its latest review's run, and no later
2564
2613
  fix-applier has landed — is one automatic retry, a review fix-applier named the
@@ -4147,7 +4196,7 @@ Plan it as a PR the run's own merge queue lands, never as an edit-and-continue.
4147
4196
  | Implementer bails before implementing | → `needs-triage` if under-specified, `ready-for-human` if it needs human hands or leaves a fork open; drop `in-progress`, comment the cause, release the claim, refill (phase 3) |
4148
4197
  | Implementer blocked or ambiguous *mid-implementation* | Free the slot, leave `in-progress`, report — the row above is the pre-code bail, not this one |
4149
4198
  | `review-pr` workflow throws or returns no tree | Retry once under the next letter (`review-pr-<pr#>-b`); still failing, report and leave the PR at review-due. Never enqueue the PR as reviewed — the workflow is not a member, so no other row here covers it |
4150
- | Reviewer or fix-applier cannot reach green | Report, leave the PR unlabeled, free the slot |
4199
+ | Reviewer, fix-applier or red-check fixer cannot reach green | Report, leave the PR unlabeled, free the slot — a fix-applier or red-check fixer settles `failed` |
4151
4200
  | Merge bot hits the hold rule | Report `held-behind-#<lower>`, PR stays queued |
4152
4201
  | Merge bot finds the worktree ahead of the PR head **on the local-rebase fallback** | `worktree-diverged-#<pr>`, PR stays queued. Read the stray commit; push-or-discard is yours, and the maintainer's if the evidence cannot settle it |
4153
4202
  | Merge bot finds the head moved after `ready-to-merge` was applied | The label binds to the head's net change, not to one SHA: a proven rebase-carry keeps the label — the bot proceeds and reports `rebase-carry-#<pr>` with the labelled and accepted SHAs, and nothing is dispatched. Any other move is `head-moved-after-label-#<pr>`, PR stays queued, label untouched. Dispatch a **fresh** finisher against the new head — the first audit verified a different tree |