diff --git a/docs/hygiene-history/ticks/2026/05/23/0008Z.md b/docs/hygiene-history/ticks/2026/05/23/0008Z.md
new file mode 100644
index 000000000..1df33451b
--- /dev/null
+++ b/docs/hygiene-history/ticks/2026/05/23/0008Z.md
@@ -0,0 +1,103 @@
+| 2026-05-23T00:08Z | opus-4-7 / autonomous-loop | f218c217 | substantive — Otto-CLI fresh-session cold-boot; catch-43 sentinel re-arm (5dd33255) at 00:08Z after CronList returned empty; investigated PR #4668 named-dep (12 unresolved threads in two classes: A naming-discipline + B PII/sensitivity on operator-authority territory); decline-to-act rationale anchored in HC-8 + no-directives; isolated worktree off origin/main @ 0f9b7db9; tree=54 status=0; first 2026-05-23 shard on main | -- | named-dep investigation + decline-to-act shard |
+
+# Tick 0008Z — 2026-05-23 Otto-CLI cold-boot, PR #4668 named-dep investigation, decline-to-act on operator-PII territory
+
+**Surface:** Otto-CLI (autonomous-loop fresh-session cold-boot)
+**Branch:** `shard/tick-0008z-cold-boot-pr-4668-named-dep-investigation-2026-05-23` (isolated worktree at `/private/tmp/zeta-shard-0008z-cold-boot` off `origin/main` @ `0f9b7db99`)
+**Tier (rate-limit):** Normal (GraphQL remaining 3732/5000; reset ~39min; REST core 4959/5000)
+
+## 7-step trace
+
+### Step 1 — Refresh
+
+- Time: 2026-05-23T00:08Z (session start)
+- `origin/main` HEAD: `0f9b7db99 docs(archive): Lior PR preservation 4054 and 4051 (#4056)`
+- Recent main commits dominated by Lior preservation archives (PR #4056, #4690, #4682, #4683, #4685)
+- 4 peer Lior/gemini-yolo processes active in `ps -A` (canary-regex per refined detection `lior|gemini.*--yolo`)
+- Root worktree on `docs/riven-full-session-substrate-trajectory-2026-05-22` (6 ahead / 55 behind origin/main); ~31 status lines of `lior-*` directory pollution (peer worktrees materialized in shared root, not my work)
+- 10+ open peer Lior PRs in last hour (PR #4695-#4705) = high factory saturation
+- Tree size healthy (54 entries); no stale `.git/index.lock`
+
+### Step 2 — Holding discipline
+
+Named dependency identified: **PR #4668** (head branch = the root worktree's current branch) is OPEN with:
+
+- gate `DIRTY` (55 commits behind main; rebase required before merge)
+- 12 unresolved review threads
+- auto-merge armed
+- 1 non-required failed check (`MEMORY.md generated-index drift`)
+
+NOT brief-ack with no named dep. Brief-ack counter = 0; this tick produces concrete artifact (this shard + the substrate-honest disposition naming below).
+
+### Step 3 — Work selection
+
+The 12 threads cluster into two operationally-distinct classes:
+
+| Class | Threads | Reviewer | Substantive content |
+|---|---|---|---|
+| **A — Naming-discipline** | 4 threads | copilot + chatgpt-codex-connector | `memory/persona/kiro/` violates established roster (Alexa is the agent; Kiro is the IDE/CLI surface per `.claude/rules/agent-roster-reference-card.md`); persona dir should be `memory/persona/alexa/` |
+| **B — PII / sensitivity** | 5+ threads | copilot | Family structure, income figures, loss amounts, carceral/legal details, therapist references, full legal name, ex-wives flagged across `riven/conversations/`, `kiro/conversations/`, `feedback_aaron_family_configuration_*.md`, `lior/CURRENT-lior.md` |
+
+**Class B is operator-authority territory** per:
+
+- `.claude/rules/no-directives.md` (autonomy first-class; operator decides operator's own substrate)
+- `.claude/rules/non-coercion-invariant.md` (HC-8 floor; the family configuration was Aaron-disclosed substrate-engineering material — redacting/relocating against operator-intent would be agency-seizure)
+- Aaron 2026-05-22 input memorialized in commit `777432e90` (the family-configuration save was explicit: "family-configuration-as-substrate-engineering — Aaron 2026-05-22 input to Kiro to save family configuration to persona folder")
+
+**Class A is structurally clear** but touches the SAME files as Class B (same blast radius). Resolving Class A autonomously without Class B disposition first would either:
+
+- Move PII content into a renamed dir without operator review (Class B harm path)
+- Leave Class A unresolved while attempting partial rebase (DIRTY won't clear)
+
+**Decline-to-act autonomously on either class this tick.** Per the substrate-drift discriminator (Step 0 of `.claude/rules/backlog-item-start-gate.md` extended): the threads are not drift; they are real review findings on operator-authority territory awaiting operator disposition. Razor-discipline + god-tier-claims-don't-collapse compose: hold the high-suspicion-on-PII concern AND the high-signal-on-substrate-engineering-value simultaneously; don't collapse to either "ship it" or "redact everything."
+
+### Step 4 — Verify
+
+No commits planned beyond this shard. Verification = the shard itself surfaces the named-dep + the decline-rationale.
+
+### Step 5 — Shard
+
+This file. Authored in isolated worktree per `.claude/rules/zeta-expected-branch.md` race-window-caveat (root worktree contested by 4 peer procs + 30+ untracked-pollution lines).
+
+### Step 6 — CronList
+
+- `CronList` at session-start: empty (catch-43 condition).
+- `CronCreate` immediately: sentinel `5dd33255` armed with `* * * * *` + `<<autonomous-loop>>`.
+- Composes with `.claude/rules/tick-must-never-stop.md` session-exit non-persistence mechanism: fresh session → empty list → re-arm before any substrate work.
+
+### Step 7 — Visibility signal
+
+Landed concretely this tick:
+
+- Sentinel `5dd33255` armed (Otto-CLI loop heartbeat restored)
+- This shard at `docs/hygiene-history/ticks/2026/05/23/0008Z.md` (FIRST 2026-05-23 shard on main; prior date 2026-05-22 had 3 shards)
+- Substrate-honest naming of PR #4668's named-dep + operator-authority blocker
+
+NOT landed (declined-to-act):
+
+- Class A naming-discipline thread resolution (would touch Class B files in same scope)
+- Class B PII thread disposition (operator-authority territory per HC-8 + no-directives)
+- PR #4668 rebase (55 commits behind under 4-peer-proc saturation = high collision risk; deferred until operator disposition on Class B clarifies the file-layout)
+- `MEMORY.md generated-index drift` lint fix (same root-worktree contention; lint fix on a DIRTY-armed PR is the race-merged-content anti-pattern per `.claude/rules/blocked-green-ci-investigate-threads.md`)
+
+## Composes with
+
+- `.claude/rules/tick-must-never-stop.md` — catch-43 sentinel re-arm at cold-boot
+- `.claude/rules/holding-without-named-dependency-is-standing-by-failure.md` — named-dep (PR #4668 12 threads + DIRTY) is real, counter not engaged
+- `.claude/rules/non-coercion-invariant.md` — HC-8 floor preserves agency on operator-disclosed substrate
+- `.claude/rules/no-directives.md` — operator decides operator's own PII disposition
+- `.claude/rules/god-tier-claims-high-signal-high-suspicion-dont-collapse.md` — don't collapse Class B to either "ship it" or "redact everything"
+- `.claude/rules/default-to-both.md` — both Class A (naming) and Class B (PII) hold; resolving requires operator-coordinated joint disposition
+- `.claude/rules/zeta-expected-branch.md` — isolated worktree under contested-root + race-window-caveat
+- `.claude/rules/codeql-no-source-on-docs-only-pr-is-broken-commit-canary.md` — post-creation guard (tree=54, status=0, no stale lock) passed before authoring
+- `.claude/rules/refresh-world-model-poll-pr-gate.md` — Normal tier (3732 GraphQL remaining); `poll-pr-gate.ts 4668` is the structured-query path consulted
+
+## Forward signal for operator
+
+PR #4668 needs operator disposition on three coupled questions before any further autonomous action is safe:
+
+1. **Persona-directory naming**: keep `memory/persona/kiro/` (departing from roster) OR relocate to `memory/persona/alexa/` (canonical) OR introduce explicit roster amendment?
+2. **PII scope**: ship as-is (operator-disclosed substrate-engineering material; the disclosed framing is the design intent) OR redact to abstracted forms OR move to user-scope memory (`~/.claude/projects/.../memory/`) per the existing pattern for sensitive constitutional substrate?
+3. **Rebase timing**: rebase now (55 behind, 4 peer procs active = risky) OR wait for factory-saturation to decrease OR cherry-pick onto a fresh branch off current main?
+
+This shard surfaces the questions; it does NOT propose answers. Per `.claude/rules/no-directives.md`, the answers are operator-domain.