Repository navigation
execute: fleet expansion per census #2017 — canonical list, umbrella + autocommit changes, per-repo management decision #2230
Description
Execution follow-on from the fleet expansion census on #2017 (comment: #2017 (comment)), under fleet epic #2055.
The census found there is no single canonical "fleet list (the 39)" — fleet membership is five separate mechanisms (sweeps/queues, the dezren39/fleet umbrella, the discovery-based laptop repo-keeper, the laptop autocommit map, the sandbox hatch-autocommit). Execute the expansion as the concrete change set below.
Cheap steps (no decision needed)
- Canonical list: create the fleet list as an explicit file in dezren39/fleet (e.g.
FLEET.mdor a machine-readable manifest) naming every managed repo and which mechanisms cover it, seeded from the investigate: fleet expansion census — CAIRN, hatch, hatch-personal, openclaw-personal, kaneo, and the rest of the touched-but-unmanaged set #2017 census table. From then on "the fleet list" is that file. - Umbrella PR (dezren39/fleet): add submodules for the census candidates not already present: openclaw-personal, cairn, cairn-private, kaneo, nix-cache, hatch-drawer, hatch-personal-drawer, opencode-muse, opencode-auto-continue, hardware-doc, dezren39 (profile), muse-esp32-screen. (Already present: code, opencode, openchamber, nix, zettel, hatch, hatch-personal.)
- Autocommit PR (developing-today/code): extend the repos map in
nixos/home/laptop-autocommit.nixfor any of cairn / cairn-private that live as laptop primaries (verify local paths first; skip any that don't). - Keeper: no change — it discovers every laptop primary clone automatically (code#1723). Verify in the next keeper report (
~/backups/repo-keeper/last-run.mdon the laptop) that the added primaries appear.
Decision for Drew (checklist — reply on this issue)
Which candidates get full management (label taxonomy + backlog + producer/queue/sweep coverage, Phase 3 pattern) vs umbrella + autocommit only (the census recommendation)?
- developing-today/kaneo — census recommends FULL (pm-stack already works its boards; issues are fleet work)
- dezren39/cairn (+ cairn-private) — census recommends FULL-lite (read-heavy research corpus; sweeps for hygiene, not implementation)
- developing-today/zettel — census recommends FULL-lite (actively pushed, fleet submodule already)
- developing-today/nix-cache — census recommends FULL-lite (fleet infra; CI/robustness issues belong in the queue)
- hatch, hatch-personal, openclaw-personal, hatch-drawer, hatch-personal-drawer, dezren39/fleet, dezren39/dezren39, opencode-muse, opencode-auto-continue, hardware-doc, muse-esp32-screen — census recommends umbrella (+ autocommit where applicable) ONLY
Also
- Add every repo Drew marks FULL to the format auto-fix replication set tracked by design+PoC: format auto-fix workflow in one repo, ready to replicate fleet-wide #2015 once its PoC lands.
- Note the census gap: org GitHub Projects membership could not be enumerated (token lacks
read:projectscope) — if Projects are part of "management", re-check after a scope refresh.
Program: Drew's 2026-10-28 directive, LIST 2 (fleet expansion). Census: #2017. Epic: #2055.
Activity
developing-today commentedon Oct 29, 2026
CODE-1102
developing-today commentedon Nov 3, 2026
agent-note:v1
state: reviewed (coverage loop-back)
Coverage pass 2026-11-03, loop-back review: this issue was re-reviewed (drift / overlap / already-handled / stale / blockers / ordering) so every open issue in the four repos is reviewed and in the system. Verdict: still live; in the system as task 7312 (triage, queued, priority 50) in repo-health-code-pilot. Labels audited and synced. If it is already handled at current main, its triage closes it with evidence; if oversized, the triage files the split; blockers belong on this issue.
Found via: coverage loop-back 2026-11-03
dezren39 commentedon Nov 3, 2026
agent-note:v1
state: triaged (escalated-to-user for the decision checklist)
task: 7312 (repo-health-code-pilot, lane code-o)
Triage: what was executed vs what remains - verified fresh 2026-11-03
Umbrella (dezren39/fleet) - PARTIALLY done. Current
.gitmoduleson fleet main: code, opencode, openchamber, nix, zettel, hatch, hatch-personal, cairn, dezren39-zettel, org-readme, zulip-archive, hardware-doc. Of this issue's 12 candidates, cairn and hardware-doc are in. Still missing (10): openclaw-personal, cairn-private, kaneo, nix-cache, hatch-drawer, hatch-personal-drawer, opencode-muse, opencode-auto-continue, dezren39 (profile), muse-esp32-screen.Canonical fleet list - NOT done. No FLEET.md or manifest in dezren39/fleet root (README.md only). "The fleet list" still has no single home.
Autocommit (code) - NOT done, now unblocked and split out.
nixos/home/laptop-autocommit.nixat current main maps only openclaw, hatch, hatch-personal. Verified on the laptop this shift: cairn IS a primary at~/git/cairnand the repo-keeper report listscairn: ok (main, worktrees: 1)- so the keeper step is verified done for cairn, and the autocommit addition is actionable. cairn-private has no laptop primary (not under~/git) - skipped per this issue's own rule. The cairn autocommit slice is enqueued as task 8918 in repo-health-code-pilot.Decision checklist - still awaiting Drew (escalated with this note, label
user/escalated). No reply on this issue since filing (2026-10-29). The question: which candidates get FULL management (label taxonomy + backlog + queue/sweep coverage) vs umbrella + autocommit only. Census recommendations already on the issue: kaneo FULL; cairn, zettel, nix-cache FULL-lite; the rest umbrella-only. This gates the remaining scope: the umbrella submodule adds are cheap either way, but FULL/FULL-lite coverage for kaneo/cairn/zettel/nix-cache should not start until Drew rules.Remainder routing: FLEET.md + the 10 missing umbrella submodules live in dezren39/fleet (outside this queue's repo) and stay tracked here on #2230 as the named remainder; task 8918 covers the code-repo slice; the #2015 format auto-fix replication item stays gated on Drew's picks plus the #2015 PoC, as written.
Verdict: issue stays OPEN; triage task 7312 completes with the split above.
dezren39 commentedon Nov 3, 2026
agent-note:v1
state: done
task: 7312 completed as triaged (repo-health-code-pilot, lane code-o). Split and escalation are in the triage note above: task 8918 carries the autocommit slice; the decision checklist awaits Drew under
user/escalated.dezren39 commentedon Nov 4, 2026
🤖 Agent note · handed off
dezren39 commentedon Nov 4, 2026
agent-note:v1
event: picked_up
task: 8918 (small_fix; repo-health-code-pilot; worker repo-worker-code; OpenChamber scheduled lane 9aca43b4-5439-4939-9f71-323f5d32521b)
at: 2026-11-04T06:44Z
state: Atomic queue claim acquired and first checkpoint saved with origin=openchamber-scheduled-lane, branch queue/8918-2230-cairn-autocommit-map, and planned worktree /home/user/git/fleet/8918.
next: Freshly revalidate issue #2230, related PR ownership, and hard-stop labels before editing.
dezren39 commentedon Nov 4, 2026
agent-note:v1
event: done
task: 8918 (small_fix; repo-health-code-pilot; worker repo-worker-code; OpenChamber scheduled lane 9aca43b4-5439-4939-9f71-323f5d32521b)
at: 2026-11-04T06:51Z
outcome: PR opened as #5326: #5326. Commit 8d83473 on queue/8918-2230-cairn-autocommit-map.
evidence: Added only cairn = "${home}/git/cairn" to nixos/home/laptop-autocommit.nix. Nix parse, targeted amd Nix evaluations (timer and generated service path), and git diff --check pass. PR checks are not yet available (only CodeRabbit neutral at verification); queue task is completed as pr_opened without waiting for CI.
labels: removed queue/claimed, status/in-progress, work/assigned. Kept user/escalated for the separate Drew decision excluded from this code slice.
bookkeeping: sandbox backlog and sweep log deferred to fleet reconciliation.