docs(PENDING-181): ADDENDUM 2 — instance 6, a mechanism that renders correctly

The repair of 4 and 5 introduced 6: REVIEWED-138's header placed with two
leading spaces. CommonMark allows it, so it renders as a heading and reads as
correct — but every reader here anchors to ^##, and grep -c '^## REVIEWED-13[89]'
returns 1, not 2. The ruling is placed, correct, and invisible to the wake
digest, the drift check, governance_item and every scan this sitting ran.

Three mechanisms now, not variants: wrap at ~150 cols, dropped clause/row, and
leading indentation. The first two are catchable by reading. The third is not.

The executor's handover format is a cause: fenced text to copy is what adds
indentation. A whitespace-only repair should be handed over as a command, which
does not traverse the transit path at all.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01T1i5VRfHjD79hfaXjWsBXA

PENDING.md

@@ -4510,3 +4510,49 @@ exact strings and a negative control. **The transit path is not self-reporting,
 nothing in the current arrangement checks a paste except a party deciding to.**
 
 **Awaiting:** unchanged — steward authorization on the parent proposal.
+
+### ADDENDUM 2 — 2026-09-10 — Instance 6, and it is a THIRD mechanism: corruption that renders correctly and is invisible to every id-keyed reader
+
+**Filed by the executor. The repair of instances 4 and 5 introduced instance 6**, which is
+the reason this is a separate addendum rather than another example in ADDENDUM 1's list.
+
+**The defect.** REVIEWED-138's repaired header was placed with **two leading spaces**:
+`  ## REVIEWED-138 — PENDING-168 — …`. The 139 list's continuation lines carry four
+where two were handed over, which is harmless.
+
+⚠ **THE HEADER ONE IS NOT HARMLESS, AND IT IS THE MOST DANGEROUS FORM SEEN SO FAR.**
+CommonMark permits up to three spaces before a `#`, so **it renders as a correct heading
+and a human reader sees nothing wrong.** But every reader in this system anchors to line
+start — `^## `. Measured immediately after placement:
+
+> `grep -c "^## REVIEWED-13[89]"` returns **1, not 2.**
+
+**REVIEWED-138 is currently invisible to the wake digest's LAST RULINGS, to
+`governance-drift-check.py`'s block parser, to `governance_item`, and to every scan this
+sitting used.** A ruling that was placed, is correct, renders correctly, and **does not
+exist to any machine that reads the register.**
+
+**Three mechanisms now, and they are not variants of one.**
+
+- **Wrap at ~150 columns** — destroys headings and table rows; prose survives. Instances 3, 4, 5.
+- **Dropped clause or row** — the content simply is not there. Instances 2, 5.
+- **Leading indentation** — *nothing is missing and nothing looks wrong.* Instance 6.
+
+⚠ **The first two are detectable by reading. The third is not**, and it is the one a
+proposal aimed at reducing clerical load is least likely to catch, because it survives
+every human check including a careful one.
+
+**⚠ THE EXECUTOR'S HANDOVER FORMAT IS A CAUSE, NOT A BYSTANDER.** The replacement was
+handed over as fenced text to be copied, and indentation is exactly what a fenced block
+in transit adds. **For a whitespace-only repair, text-to-paste is the wrong instrument —
+the fix should be handed over as a command the steward runs, which does not traverse the
+transit path at all.** Recorded as a standing lesson for this executor, not as a proposal.
+
+**A one-line canary exists and is NOT built here.** The invariant is checkable in a
+single comparison: the count of `^## REVIEWED-` against the count of register headings
+found without the anchor. ⚠ **Not proposed for the verification ladder — REVIEWED-123's
+freeze is general.** ⚠ **And not added to `governance-drift-check.py` unilaterally**, which
+carries its own open item (PENDING-139). Named so the next hand does not have to find it
+by having a ruling go missing.
+
+**Awaiting:** unchanged — steward authorization on the parent proposal.