diff --git a/claude/memory/skill-harvest-register.md b/claude/memory/skill-harvest-register.md
index 0003085..ee0fd8a 100644
--- a/claude/memory/skill-harvest-register.md
+++ b/claude/memory/skill-harvest-register.md
@@ -218,3 +218,6 @@ The single place proposed skills live so they don't evaporate between sessions.
 | 175 | ``/wrap-up` §5 (KG append)` | patch | Add a `prevention` predicate — `{subject: <banked lesson>, predicate: "prevention", object: <the | same | PROPOSED | New proposals (2026-07-29, steward-raised —  |
 | 176 | ``/wake-up` §2.b.2` | patch | Surface one `prevention` alongside the drift-patterns. Currently the wake greps only `drift-patt | same | PROPOSED | New proposals (2026-07-29, steward-raised —  |
 | 177 | ``symmetria` §2 / `/wrap-up` §1` | patch | Retire the self-report framing of the standing question. The 2026-07-29 literal question — *"is  | 2026-07-29 | PROPOSED | New proposals (2026-07-29, steward-raised —  |
+| 178 | `/fool` | create | **Build WHEN STABLE, not now.** The differently-formed-checker trial protocol, derived twice this session: withhold the ruling; pre-register grading before the run; prompt gives *form* not target (+ anti-contrarian + anti-echo clauses); `enable_thinking` ON (off produces silence, not brevity); **one variable per trial**; no standing granted to the Fool — only checkable claims get standing. Codifying it now would freeze an unstable protocol; the steward has directed next session at stabilising it. Interim home: `dotfiles/claude/governance/fool-trial-log.md`. | 2026-08-01/02, trials 01–02 | PROPOSED (build-when-stable) |
+| 179 | `wake-digest.py` | patch **[FIX-class]** | **The wake instrument silently hides open items.** `sec_pending()` drops a `PENDING-N` whenever a `REVIEWED-N` exists — **matching the number alone**, never checking the ruling is *about* that item. Numbering has drifted (REVIEWED-84 rules on PENDING-87), so **PENDING-84 was invisible at wake on the very morning the steward's pulling thread pointed at it**; closing it later produced no visible count change. Measured: 9 suppressed, 8 correctly, 1 falsely. Fix: resolve by the ruling's *subject line*, not its number. | 2026-08-01 (found at wake, surfaced not fixed) | PROPOSED |
+| 180 | `normalize_ocr.py` (chamber fleet) | patch **[FIX-class]** | **Silent degradation with no disclosure.** `dict_state` reports only the static word list, so a `--lang fr` run with **wordfreq absent** falls back to its own documented *"old, weaker behaviour"* — validating French de-hyphenations against `/usr/share/dict/words` (English) — and says nothing. §VII: *a measurement carrying a known reliability caveat states it on its own output, every time*. Fix: report wordfreq active/absent and the effective language on every run. (wordfreq installed into `~/.local/chamber-tools-venv` 2026-08-01.) | 2026-08-01 (Derrida conversion) | PROPOSED |