From 3ae21f90470bd72e1859334623d82930db8a4218 Mon Sep 17 00:00:00 2001 From: David F Glidden Date: Sat, 4 Jul 2026 22:54:21 +0200 Subject: [PATCH] =?UTF-8?q?session=202026-07-04:=20skill-harvest=20?= =?UTF-8?q?=E2=80=94=20CI-for-ESCALATE=20ladder=20entry=20+=20model-handof?= =?UTF-8?q?f=20reinforced?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- claude/memory/skill-harvest-register.md | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/claude/memory/skill-harvest-register.md b/claude/memory/skill-harvest-register.md index 3ea3450..2a6a2c3 100644 --- a/claude/memory/skill-harvest-register.md +++ b/claude/memory/skill-harvest-register.md @@ -307,3 +307,12 @@ The single place proposed skills live so they don't evaporate between sessions. | **`/wrap-up` §4.a** | **patch** | The §4.a drawer-filing step instructs passing `tags:` to `mempalace_add_drawer` — **the tool rejects it** (`MCP error -32602: Unknown parameter 'tags'`). Misfired in use this wrap. Fix: drop the `tags:` param from the §4.a instruction; fold retrieval keywords into the drawer **content** (a `[tags: …]` line) instead, since content is the semantic surface. One-line correction; caught live. | `~/.claude/skills/wrap-up/SKILL.md` §4.a | **PROPOSED (patch — caught in use)** | *(Two, both earned in use. `/spec-amendment`'s deferral condition [first real use] is genuinely met — this session was that use, so it's a build not a re-propose. The §4.a patch is a real tool-instruction mismatch that cost one failed drawer call this very wrap. Neither manufactured; source-not-derived-as-drafting-method is NOT re-proposed as a skill — it's already a discipline-named KG fact + reinforces the existing trusted-derived-artifact §3 flag.)* + +### Harvest 2026-07-04 (Seam-1 ESCALATE + engine deep-research day) + +| Element | Kind | One-line | Status | +|---|---|---|---| +| **verification-ladder: CI-upper-bound + drop-one-robustness = STANDARD for every ESCALATE** | verification-ladder entry | The jurist RULED (2026-07-04) that grading on the one-sided 90% Clopper-Pearson **upper bound** (not the point estimate) + the **drop-one-confirmed-case robustness check** are standing practice for *every* ESCALATE, not a one-off. A point estimate at n=40 can't tell 1% from 5% (P(0or1\|.05,40)=0.399); the CI does the real work. → `reference-verification-ladder.md`. | PROPOSED | +| **`/model-handoff`** (premium-model scope-charter) | REINFORCED (existing proposal, 2026-06-12) | Used tonight end-to-end: produced `studium-engine/docs/stage-1-replan-scope-charter-2026-07-05.md` on Opus (§0 discipline / §1 settled / §2 already-read / §3 bounded-reading / §4 deliverable / §5 report-and-stop) as the Fable-5 handoff; clear+wake not switch-in-place; artifacts-are-the-handoff. Second clean use → strengthens the standing proposal. | PROPOSED (reinforced) | + +*(Both genuine, both recurring-shaped. The CI-for-ESCALATE is a jurist-ruled standing practice worth a ladder entry; the model-handoff is its second proven use. No NEW skill emerged — the chamber-specific instruments (exclusion-path, archive-audit, Instrument-B hardening) are corpus work, not skills.)*