decisions, build-log, sessions: 110, Q109 follow-ups filed in Vikunja P19 (#153-#161)

Co-Authored-By: Claude Opus 5.5 <[email protected]>

docs/SESSIONS.md

@@ -752,3 +752,4 @@ are never rewritten or removed; corrections are new entries.
 2026-10-12T14:59:20Z | sage | decision 107, I1-3 deferred | Jason deferred Q104 via Mos sage-0077: no WebAuthn in the I1 brief; reopens only as a new question once containers run; a compromised Console holding human authority is recorded as open
 2026-10-12T15:00:59Z | sage | decision 108, I1-4 answered | Jason chose Q105 A via Mos sage-0078: subscription token as a read-only file secret per session container, no API key or gateway; copy-out risk accepted, bounded by the egress allowlist; all four I1 items answered, brief is Sage's on 2026-10-13
 2026-10-12T17:27:49Z | sage | cutover live, decision 109 | verified the mosaic.woltje.com Console read-only after Mos's 17:21Z start (status active, send on, healthz 200); decision 109: landings touching webui, bus or cli in the canonical checkout are live changes until the Console runs from a pinned copy; BUILD-LOG entry
+2026-10-12T17:38:19Z | sage | decision 110, follow-ups to Vikunja | Jason's Q109 via Mos sage-0081: Vikunja project 19 holds Mosaic work items; filed #153-#161 with the pm slot, read back with sync; #146 and #147 cited for the 10-13 rows; queue and Gitea stay the review and landing record; BUILD-LOG entry

docs/plans/2026-09-26_lead-decisions.md

@@ -2437,3 +2437,28 @@ which stay with him. Each item names who decided it and what happened.
       being deploys. Pointing the unit at it is a reinstall, which is
       Mos's step under decisions 102 and 103. It goes in the same window
       as decision 101's uid reinstall.
+110. **Vikunja holds Mosaic work items (2026-10-12, Q109).** Jason, relayed
+    by Mos in sage-0081: "Always use Vikunja for task tracking." Mosaic
+    work items live as tasks in project 19 (Mosaic Stack) on
+    tasks.woltje.com. Sage files each new one there when raising it and
+    stops keeping follow-up lists in BUILD-LOG, standups or decisions.
+    Those records cite the Vikunja id.
+    - Filed at 17:37:55Z with the pm token slot, as Mos filed #146 to
+      #152, since Sage has no key of its own: #153 I1 brief, #154 M1
+      brief and row, #155 row 77 brief (all three due 2026-10-13), #156
+      compose.yaml:40 mount, #157 stale LAYERS and CONDUCTOR docs, #158
+      board Reply without tmux, #159 Console cosmetic notes from rows 76
+      and 83, #160 Console flag tests, #161 row 67 test 4 (Dewey N1).
+      Each was read back with the sync key. No labels, assignees or
+      bucket moves. Evidence:
+      `agents/sage/work/vikunja-setup/2026-10-12_followups.json`.
+    - Decision 109's pinned-copy row is #146 and decision 101's uid
+      reinstall is #147. Both are due 2026-10-13, and the 10-13 rows
+      cite them.
+    - The queue and its Gitea issues stay the review and landing
+      record, because `queue review` and `verify-commit` gate landings
+      on them. When a row opens for a Vikunja task, the task names the
+      row and the row's brief names the task. Moving the queue itself
+      into Vikunja would be a separate change and goes to Jason first.
+    - If the pm slot fails for Sage, Mos files the task (sage-0081).
+      Agent key scopes are #151, owned by ops-01.