diff --git a/docs/hygiene-history/ticks/2026/05/16/1654Z.md b/docs/hygiene-history/ticks/2026/05/16/1654Z.md
new file mode 100644
index 0000000000..8fefbc4f67
--- /dev/null
+++ b/docs/hygiene-history/ticks/2026/05/16/1654Z.md
@@ -0,0 +1,103 @@
+# Tick 2026-05-16T16:54Z — Otto-bg-worker
+
+Catch-43 re-arm + state-of-world snapshot. Extreme cost-aware
+tier (GraphQL 451/5000 → reset in 2 min). Lior cascade active
+in parallel.
+
+## Catch-43 re-arm
+
+Fresh autonomous-loop session. CronList showed no jobs;
+CronCreate sentinel `<<autonomous-loop>>` armed at `* * * * *`
+(job id `a32665e9`). Per
+[`tick-must-never-stop.md`](../../../../../../.claude/rules/tick-must-never-stop.md)
+— session-only cron dies with each session; every wake MUST
+re-arm.
+
+## Refresh result
+
+| Surface | State |
+|---|---|
+| Cron sentinel | `a32665e9` armed (fresh this session) |
+| GraphQL rate limit | 451/5000; reset in 2 min |
+| REST core | 4959/5000 (free for fetches) |
+| origin/main HEAD | `c96a7edad` (#3907 shard 1618Z) |
+| Most recent main | `bdb27a220` (#3908 shard 1634Z) |
+| Open PR count | 35 (~28 Lior preservation/shadow/archive) |
+
+## Lior parallel cascade
+
+`ps -A | grep -E "gemini.*[Ll]ior"` returned 3 active processes
+(lior-loop-tick + 2 gemini Lior instances). Per
+[`codeql-no-source-on-docs-only-pr-is-broken-commit-canary.md`](../../../../../../.claude/rules/codeql-no-source-on-docs-only-pr-is-broken-commit-canary.md),
+Lior's step 8 (global lock cleanup) has empirically corrupted
+commit trees at worktree-creation time. Decision: **work in
+current worktree, no new worktrees**.
+
+Lior is also doing step 2 (read open PRs) + step 3 (review
+PRs) — running `poll-pr-gate-batch.ts --all-open` on top would
+duplicate review work AND burn ~150-300 GraphQL points.
+
+## Open-PR census (REST, free)
+
+`gh pr list --state open --limit 35` (REST not GraphQL):
+
+| Author bucket | Count | Pattern |
+|---|---|---|
+| Lior preservation | ~12 | `lior/preserve-*`, `lior/archive-*` |
+| Lior shadow drift logs | ~10 | `lior/shadow-*` |
+| Lior decompose / chore | ~6 | `lior/decompose-*` |
+| Otto-CLI in-flight | 2 | #3910 (stale-armed resolution), #3813 (B-0509 FP) |
+| Otto-bg-worker in-flight | 1 | #3909 (1643Z shard) |
+
+## Otto in-flight triage (2 GraphQL views)
+
+| PR | State | Auto-merge | CI |
+|---|---|---|---|
+| #3909 (otto-bg-worker 1643Z) | MERGEABLE / BLOCKED | armed by AceHack | all SUCCESS |
+| #3910 (Otto-CLI stale-armed) | MERGEABLE / BLOCKED | armed by AceHack | build-and-test IN_PROGRESS (3 OS matrix) |
+
+Both PRs have auto-merge armed. #3910 legitimately waiting on
+CI matrix (~few min). #3909 all-green-BLOCKED — likely queue
+position; trust auto-merge.
+
+## Why no `--all-open` batch poll this tick
+
+Substrate-honest deferral, named-dependency-bounded:
+
+1. **Rate-limit named dep** — 451 GraphQL, reset in 2 min;
+   `--all-open` on 35 PRs would burn the recovered budget
+2. **Lior cascade parallel** — duplicating review work the
+   gemini-Lior processes are doing right now
+3. **Auto-merge already armed** on the 2 Otto PRs that matter
+4. **No outstanding directive-shaped action** — Lior PRs
+   triage themselves on Aaron's review path; bg-worker's job
+   is the tick shard heartbeat, not Lior-PR triage
+
+Per
+[`refresh-world-model-poll-pr-gate.md`](../../../../../../.claude/rules/refresh-world-model-poll-pr-gate.md)
+rate-limit tier table: 200-1000 = extreme cost-aware
+("Skip batch-polling entirely. Open at most 1 PR per tick").
+This shard PR is that one PR.
+
+## Counter-with-escalation
+
+Fresh session start. Counter at 0 — no consecutive
+brief-acks. This tick produces concrete artifact (the shard
+itself + this PR), so the counter discipline is satisfied
+without needing escalation per
+[`holding-without-named-dependency-is-standing-by-failure.md`](../../../../../../.claude/rules/holding-without-named-dependency-is-standing-by-failure.md).
+
+## Sentinel + close
+
+`CronList`: `a32665e9` armed (`* * * * *`).
+
+## Visibility signal
+
+- Catch-43 sentinel re-armed (job `a32665e9`)
+- Rate-limit observed: 451/5000 (recovers in 2 min)
+- 35 open PRs (~80% Lior cascade); 2 Otto PRs auto-merge armed
+- Lior cascade observed; no worktree-creation hazard taken
+- Substrate-honest deferral of `--all-open` per rate-limit
+  tier discipline
+- Next tick (post-reset) may do focused triage on
+  oldest-stale Otto-CLI PRs (#3813 from 0727Z if still open)