shard(tick): 0813Z — PR #3384 Codex P2 fix (cross-PR file-ref); restraint discipline

docs/hygiene-history/ticks/2026/05/15/0813Z.md

@@ -0,0 +1,52 @@
+| 2026-05-15T08:13:00Z | claude-opus-4-7 | 596e842c | shard: PR #3384 Codex P2 fix (cross-PR file-ref chicken-and-egg); 6 PRs in CI wait | (PR #3387) | compact maintenance tick: address review thread, restrain from new substantive threads, document the cross-PR-ref discipline |
+
+# Tick 0813Z — PR #3384 Codex P2 fix; 6 PRs in CI wait; restraint discipline
+
+## Headline
+
+- **PR [#3384](https://github.com/Lucent-Financial-Group/Zeta/pull/3384)** (ID-allocation rule update) picked up one Codex P2 thread: my empirical-anchor cited the 0742Z shard file path, but that file doesn't exist on #3384's commit (chicken-and-egg with #3381 which lands the shard). Fix: cite PR #3381 + PR #3323 (durable on GitHub) instead of file path. Commit `024f947`; thread resolved.
+- **6 PRs still in CI wait** (#3379, #3381, #3382, #3383, #3384, #3385) — all auto-merge armed. No new failures.
+- **Restraint discipline applied**: I had two next-tick candidates (claude-loop substrate investigation, `(PR TBD)` mechanization). Did NOT start either this tick. Per [`.claude/rules/holding-without-named-dependency-is-standing-by-failure.md`](../../../../../../.claude/rules/holding-without-named-dependency-is-standing-by-failure.md), the "Holding"-failure-mode test asks if there's a named dependency — there are 6 named dependencies (the in-flight PRs awaiting CI). That IS the named-dependency-wait. Address review threads when they surface; don't pile on new substantive threads.
+- Cron sentinel `596e842c` armed.
+
+## Δ since 0804Z
+
+| What | At 0804Z | At 0813Z |
+|---|---|---|
+| PR #3384 | OPEN, auto-merge armed | OPEN, 1 Codex P2 → fixed + resolved (`024f947`) |
+| Other 5 PRs | OPEN, wait-ci | OPEN, wait-ci (unchanged) |
+| New substantive threads started | n/a | 0 (restraint discipline) |
+
+## Substrate-honest observations
+
+### Cross-PR file-ref chicken-and-egg
+
+When a rule update (PR #3384) cites a shard file (`0742Z.md`) that's introduced by a different PR (#3381), the cross-PR reference is broken on the rule-update PR's branch. Codex correctly catches this — the file genuinely doesn't exist there.
+
+The fix pattern: cite PR numbers (durable on GitHub) instead of file paths (branch-state-dependent). PR URLs are stable; navigating to them shows the file content in their PR-diff view.
+
+This is a structural discipline emerging from parallel-PR work: **cross-PR references should be by PR# not by file-path** when the file is being introduced by another in-flight PR.
+
+### Restraint applied at tick-scope
+
+The carry-forward from 0804Z named two next-tick candidates:
+
+1. Investigate `com.zeta.claude-loop` substrate (last unknown in worktree-prune-race investigation)
+2. Mechanize `(PR TBD)` fix (small TS helper)
+
+Both are valid. Neither is urgent. The named-dependency-wait is real (6 PRs in CI). Per restraint discipline, the right move is address-review-threads-as-they-surface, not pile-on-new-threads. This tick ships 1 fix + brief shard + stops.
+
+This composes with the earlier B-0527 restraint pattern (after 2 unactioned bus advisories, no third republish). The same shape: when there's no new signal demanding action, the discipline is to NOT generate one for its own sake.
+
+## Cron sentinel
+
+`596e842c` armed.
+
+## Next
+
+Cron-driven. Next tick:
+
+1. Verify the 6+ in-flight PRs auto-merge fires
+2. Address any new review threads
+3. Consider claude-loop investigation OR `(PR TBD)` mechanization if no fix-work surfaces
+4. Watch PR #3323 for Lior response (still no movement)