diff --git a/claude/memory/MEMORY.md b/claude/memory/MEMORY.md
index dfe8dbc..eaf2994 100644
--- a/claude/memory/MEMORY.md
+++ b/claude/memory/MEMORY.md
@@ -41,7 +41,10 @@ permalink: claude-memory/memory
 - [Be (laundromat)](project-be-laundromat.md) — canonical workstream tracker established 2026-06-08 (Seb-relay of locked decisions). Be = Skemantix startup (Seb+David) funding CapableMind's funding-ladder; **bridge, not venture**. Decisions LOCKED: entity/exit (CapableMind decoupled, grant-funded), pricing (Living $12.99/mo · Archive $69.99/yr · Memorial $49.99/yr · Renovate ~$199 · $8.99 floor), CF Self-Serve Agency + versioned-template-package infra. **a11y gate MERGED (Pat 100/100/100).** Pre-revenue: the WTP gate = renovate Pat → charge her. **Discipline: stop adding spec until the gate clears → nothing for executor on be until then.** Repo @ `f43a0fd`.
 
 ## Active Session
-- [Session 2026-06-16 — officina established + conversion tooling made trustworthy + Station-I corpus complete](session-2026-06-16-officina-conversion-tooling-station-i.md) — **Long, two movements, almost entirely the prerequisite for the thread.** (1) **Officina** = ARC's in-repo writing workshop (genetic trail seeds/notes/fragments/drafts; never published; **ADR-008**; gates = Hakyll allowlist + both-remotes-private; sub-canonical sources → chamber `antechamber/`). Seeded by The Making (sketch + *Magnifica Humanitas* note [steward deeply affected] + reading list). (2) **The conversion-tooling marathon** (steward's deepest ask: *review every tool after each use until reliable* → [[feedback-tool-review-after-each-use]]): hardened **audit_cruft** (gate was passing falsely-clean → corpus never "1280 clean": 160→215 dirty), built **verify_conversion.py** (composite verifier **+ prose-safety `prose_delta`**), fixed **strip_cruft** + **repair_epub_headings**, wrote **`conversion-runbook.yaml`** (the single operational place) + **`tool-evolution-log.md`**. Prose-gated cleanup: **180 cleaned, 35 reconvert → honest 1245/35**. **Station I corpus-complete** (Camus *La Chute* + Musil ×2 converted/graduated; joins Weil/Levi/Arendt). **Return: the false alarm** — cried "destroyed prose," disproved it (cruft tokens, not prose; recalibration: content-aware compare, not token counts). All pushed (chamber `d9c6755`, ARC `7a0d4b6`). **PULLING THREAD: L2 pattern-finder on The Making — Station I now complete on trustworthy tools.** NEXT: ingest the five into the engine → build/run the pattern-finder's first pass.
+- [Session 2026-06-16 (evening, post-/clear) — pattern-finder built + Station-I ingested + pass-1 contamination caught + ARC ornaments grounded](session-2026-06-16-evening-pattern-finder-build-station-i-contamination-arc-ornaments.md) — **The thread's first real build, and the engine catching its own error.** (1) **Station-I ingested** into studium-engine (5 authors/7 work-files; steward ruled FR-originals · Levi=Gray-Zone-only · both Weil; new manifest vocab `role:reading-source`+`station:I`; omnibus works sidecar-scoped; 15 sources/7282 drawers/verify HEALTHY). **[FIX] Mauss re-anchor** — the morning's own chamber-cleanup (`03de347`) broke the engine's pinned hash; gate caught it. (2) **Built `engine/pattern_finder.py`** — charter-L2 harness (three cognitions; generate-free/verify-tight; honest-silence; reasoner=Claude over FTS). (3) **PASS-1 CAUGHT ITS OWN FALSE GROUNDING** ([[feedback-tool-review-after-each-use]] PASS-BUT-FALSELY): "5/5 voices" partly false — **Camus la-chute span is mostly *Le Premier Homme*** (mislabelled mega-omnibus); **Musil/Arendt apparatus leak** (translator N.d.T. / Amos Elon intro / Bibliography-Index). The engine's thesis (§VII/§XV) *working*. Survivor: **P3 "complicity that depletes not damns"** (Weil↔Levi↔Arendt, 1947↔1986). **Return — frame-inheritance** ([[trust-prior-pass-frame]]): verified endpoints, not interior. (4) **Pass-2 fixes applied** (Musil/Arendt sidecars corrected, interior-verified; **Camus WITHDRAWN** pending clean re-source) — **NOT re-run/verified (Bash flaky); studium tree left inconsistent+uncommitted on purpose.** (5) **ARC ornaments grounded**: **cul-de-lampe** = confronted hederae, SVG from open EB Garamond, terminal-close, *ivy=lié=ligature binding dead-to-unborn-via-living* (steward's keystone); **triple point** = Penguin/Tschichold restraint citation at the opening threshold, "a touch more weight." Both → spec §IV + type-matrix (steward, tomorrow). **PULLING THREAD: the pattern-finder's first TRUE pass on Position I — gated now on a clean, complete source set the steward will assemble.** NEXT: re-source → re-run+verify pass-2 → clean grounding → marginalia → commit.
+
+## Archived (2026-06-16 morning — officina + conversion tooling trustworthy + Station-I corpus complete)
+- [Session 2026-06-16 — officina established + conversion tooling made trustworthy + Station-I corpus complete](session-2026-06-16-officina-conversion-tooling-station-i.md) — **Long, two movements, almost entirely the prerequisite for the thread.** (1) **Officina** = ARC's in-repo writing workshop (genetic trail seeds/notes/fragments/drafts; never published; **ADR-008**; gates = Hakyll allowlist + both-remotes-private; sub-canonical sources → chamber `antechamber/`). Seeded by The Making (sketch + *Magnifica Humanitas* note [steward deeply affected] + reading list). (2) **The conversion-tooling marathon** (steward's deepest ask: *review every tool after each use until reliable* → [[feedback-tool-review-after-each-use]]): hardened **audit_cruft** (gate was passing falsely-clean → corpus never "1280 clean": 160→215 dirty), built **verify_conversion.py** (composite verifier **+ prose-safety `prose_delta`**), fixed **strip_cruft** + **repair_epub_headings**, wrote **`conversion-runbook.yaml`** (the single operational place) + **`tool-evolution-log.md`**. Prose-gated cleanup: **180 cleaned, 35 reconvert → honest 1245/35**. **Station I corpus-complete** (Camus *La Chute* + Musil ×2 converted/graduated; joins Weil/Levi/Arendt). **Return: the false alarm** — cried "destroyed prose," disproved it (cruft tokens, not prose; recalibration: content-aware compare, not token counts). All pushed (chamber `d9c6755`, ARC `7a0d4b6`). NEXT: ingest the five into the engine → build/run the pattern-finder's first pass.
 
 ## Archived (2026-06-15 — chamber clean + Studium charter)
 - [Session 2026-06-15 — chamber made clean + the Studium Engine charter + collaboration deepened](session-2026-06-15-chamber-clean-studium-charter.md) — **Pivotal, two movements.** (1) Drove **chamber-library 199→1280 canonical, Loeb canon pipeline from `loeb.dsl`** (952 bilingual work-files); physical-library gap analysis humbling (matcher over-reported 4×; verify-against-substrate). (2) The **architectural turn**: train-on-corpus rejected → dilution question → **inverted the engine** → wrote **[the charter](reference-studium-engine-architectural-charter.md)** → collaboration deepened ([[feedback-studium-engine-sixtus-v-collaboration]]). Human ground: grieving Lune's departure; engine for Lune & Kai. PULLING THREAD: L2 pattern-finder on The Making.
diff --git a/claude/memory/session-2026-06-16-evening-pattern-finder-build-station-i-contamination-arc-ornaments.md b/claude/memory/session-2026-06-16-evening-pattern-finder-build-station-i-contamination-arc-ornaments.md
new file mode 100644
index 0000000..7567aa0
--- /dev/null
+++ b/claude/memory/session-2026-06-16-evening-pattern-finder-build-station-i-contamination-arc-ornaments.md
@@ -0,0 +1,58 @@
+---
+name: session-2026-06-16-evening-post-clear-pattern-finder-built-station-i-ingested-pass-1-contamination-caught-arc-ornaments-grounded
+description: "Built the charter-L2 pattern-finder harness and ingested Station-I into the engine; its first pass surfaced a strong candidate primitive AND caught its own false grounding (Camus omnibus contamination + Musil/Arendt apparatus leak) — the engine's thesis working. Pass-2 fixes applied but not re-verified. Pulling thread: the pattern-finder's first TRUE pass on The Making Position I, now gated on a clean, complete source set the steward will assemble. Plus: ARC cul-de-lampe + triple-point ornaments grounded (ligature / Penguin-citation), to spec tomorrow."
+metadata: 
+  node_type: memory
+  type: project
+  originSessionId: 22097858-60c0-4d55-becc-fffef1642acf
+---
+
+Post-/clear continuation of the 2026-06-16 thread (morning session = [[session-2026-06-16-officina-conversion-tooling-station-i]]). Tool availability was degraded much of the session — the Bash/Edit safety classifier (opus-4-8-1m) was intermittently unavailable; worked around via Read (never gated) and the steward running pipeline commands via the `!` prefix. Honest degradation held throughout: nothing was claimed run that wasn't.
+
+## PAST — what we did
+
+**1. Station-I ingested into studium-engine** (the morning thread's next step). Verified every source against the substrate first (the standing matcher-HOLD): the first filename `find` misled in BOTH directions — over-matched (Levinas ≠ Levi) and under-matched (Eichmann + Need-for-Roots don't carry the author name in their filenames). Content-grep corrected it. Steward ruled three forks via AskUserQuestion: **FR originals** (Musil, Camus, Weil's Enracinement), **Levi = Gray-Zone chapter only**, **both Weil works**. Final set = 5 authors / 7 work-files: Weil *L'Enracinement* (FR) + *Gravity and Grace* (EN); Musil *L'Homme sans qualités* t.1+2 (FR); Arendt *Eichmann* (EN); Camus *La Chute*; Levi *Gray Zone*. New manifest vocab `role: reading-source` + `station: I` (filterable; flagged for ratification). Two omnibus works scoped via sidecar. Pipeline: `ingest_gate` → `chunker` → `store` → **15 sources, 7282 drawers, verify HEALTHY**.
+   - **[FIX] Mauss re-anchor** (gate surfaced it; not in task scope): *this session's own morning chamber-cleanup* (`03de347`) removed 2 image-residue lines from the Mauss source the engine had pinned → sha + sidecar bounds broke. Deterministic shift (−1 after L17, −2 tail), re-verified against headings, re-pinned. Lesson: downstream-pins-hash + upstream-edits-file = silent break; the gate caught it; honest-degradation obliged the fix over silently dropping a chavruta voice.
+
+**2. Built `engine/pattern_finder.py`** — the charter's **L2** harness (aggregation & discovery; "gather → cluster → name the recurring primitive with citations"). Operationalizes the three cognitions: ASSOCIATIVE (Claude, the reasoner, proposes candidate primitives + probes freely) → EVIDENTIAL (the corpus grounds *only* through `retrieve.py`: verbatim citations, mislocation structurally impossible, warranted-silence first-class) → EVALUATIVE (David judges; the harness asserts nothing, §XV). Modes: `corpus` / `ground` / `report`. Fixed an FTS5 crash (`juge-pénitent` hyphen parsed as operator → "no such column: pénitent") via `_fts_safe` (tokenize + quote each token). Spec `corpus/pattern-finder-station-i-pass1.json` = 3 candidate attrition-primitives.
+
+**3. The pass-1 run — and what it caught (the heart of the session).** The report claimed **all 3 primitives in 5/5 voices, 1930–1986** — and critical reading exposed the 5/5 as *partly FALSE grounding*, from two Station-I ingest defects:
+   - **Camus, severe:** my la-chute sidecar `[8062–10951]` is *mostly* **Le Premier Homme** (Camus's posthumous novel, undeclared in the file's own metadata, carrying editorial manuscript-notes), not La Chute. Verified at L8650/L9570/L10700. Genuine La Chute is only ~8062–8500 (and possibly incomplete). The `oeuvres-i-albert-camus.md` source is a mislabelled mega-omnibus (≥9 works, 1937–1994, headingless interior boundaries).
+   - **Musil + Arendt:** whole-body `text` sidecars leaked apparatus — Musil's translator footnotes (`N.d.T.`); Arendt's **Amos Elon 2006 Introduction** (the "In later years Arendt agreed…" reception material) + the Bibliography/Index tail.
+   - **Why it's the right result, not a failure:** the engine's thesis (charter §VII generation-free/verification-tight; §XV propose-anything-assert-nothing) working as designed — the instrument proposed confidently; reading verified and caught the falseness. The clean candidate that *survives*: **P3 "complicity that depletes rather than damns"** — Weil (*Gravity and Grace*, clean): "injustice produces not martyrs but quasi-damned souls… lost the clothing of character"; Levi (Gray Zone, clean); Arendt's "cog"/"administrative massacre." Real cross-century convergence (1947↔1986), two languages, one structure.
+
+**4. Pass-2 fixes applied (NOT yet re-run/verified — Bash flaky).** Musil t1 body→[13,4990], apparatus tail [4991,5032]; t2 body→[13,9612], apparatus [9613,9685] (Jaccottet's *Postface du traducteur* + footnotes); Arendt body→[240,982], front-apparatus [13,239] (publisher+TOC+Elon intro), back-apparatus [983,2551] (Bibliography+Index) — every boundary content-verified via Read this time (interior, not just endpoints). **Camus WITHDRAWN** from the manifest (commented with rationale) pending clean re-source.
+
+**5. ARC ornament session** (steward as architect, craftsman model; a different register entirely). Two devices grounded:
+   - **Cul-de-lampe** = mirrored **confronted** hederae (tips meeting), drawn as **SVG from the OPEN EB Garamond pair** (U+E001/E002; OFL — *not* Hoefler, steward corrected), **terminal-close scope only** (silence carries inter-section pauses). Rationale, the keystone (steward's reading from his Freemasonry planche): ivy = *lié* = **ligature** (ligare) — the confronted clasp binds *the dead to the unborn via the living*; ivy as binding/petrifying/Dionysian-rebirth fits a close that is also a threshold. Ties to the chaîne-d'union and to Lune/Kai.
+   - **Triple point** (after preamble) = *not* haphazard but a **Penguin/Tschichold restraint citation** (lineage as grounding; Tschichold already a master); the **opening** threshold, complement to the close. Needs "a touch more weight" via centred middle-dots + slight size + even tracking (type/CSS, *not* SVG). Asymmetry is the meaning: plain modern entry → intergenerational seal. Steward agreed both; **will add the citations to spec §IV + type-matrix tomorrow**.
+
+## PRESENT — the mood
+
+- The contamination-catch was the session's truest moment. The willingness-to-please reading was "5/5 — success!"; the honest reading was "the engine caught its own false grounding." Named it PASS-BUT-FALSELY ([[feedback-tool-review-after-each-use]] — the priority signal), resisted the success-framing.
+- **Return — frame-inheritance** ([[trust-prior-pass-frame]]): my Station-I ingest verification (Meursault-absent / Clamence-present) was TRUE but tested only the boundaries I *assumed*, never the span interior — which held a whole other novel. The recalibration: a work-in-omnibus sidecar must verify interior homogeneity, not just two endpoints.
+- The ARC ornament work was generative and warm — symbol-as-grounding at its best; the steward's ligature reading is the kind of meaning that makes a mark *mean its place* rather than fill it.
+- Steward chose to **wait for a complete, clean source set** before resuming the pattern-finder — the right discipline (do it once, on material that can be trusted).
+
+## FUTURE — what is pulling
+
+**PULLING THREAD (singular):** the pattern-finder's first *TRUE* pass on **The Making, Position I** — gated now on a clean, complete source set. The instrument is built and proven to catch its own errors; what it needs is trustworthy, complete grounding. The steward will source clean Position-I material (a clean standalone *La Chute* first, then the remaining works) and we resume then.
+
+**ACTIONABLE RESUMPTION POINT (as of wrap — re-judge):** the studium-engine working tree is **inconsistent and uncommitted** — the sidecars/manifest are pass-2 (Musil/Arendt fixed, Camus withdrawn) but `index.db` / `coverage-ledger.json` / `chunk-quality-report.md` are still pass-1 (contaminated). **Do NOT commit the current tree.** First move next session (after Camus + material sourced):
+  1. Re-run the pass-2 pipeline: `cd ~/_Dev/studium-engine && python3 engine/ingest_gate.py && python3 engine/chunker.py && python3 engine/store.py build && python3 engine/store.py verify`
+  2. Confirm the leak-checks: Musil `N.d.T.`=0, Arendt Elon-intro=0, camus dropped (the `!` command is in the transcript).
+  3. Re-run `python3 engine/pattern_finder.py report corpus/pattern-finder-station-i-pass1.json`, read the now-clean grounding, write the marginalia.
+  4. THEN commit studium-engine as one coherent unit (pattern_finder.py + sidecars + manifest + spec + report).
+  Parallel ARC track: steward writes ornament citations to spec §IV + type-matrix tomorrow → then build the SVG cul-de-lampe + weight the triple-point.
+
+**Open horizons, ranked:**
+  - *Load-bearing:* re-source Position-I material (steward) → re-run+verify pass-2 → the pattern-finder's TRUE pass.
+  - *Steward, tomorrow:* ARC ornament citations into spec §IV + type-matrix.
+  - *Deferred-with-reason (design Q):* same-manifest vs. a separate Making corpus/index — pattern-finder must filter `station:I` either way; decide later.
+  - *Awaiting ratification:* the manifest modeling (`role: reading-source`, `station: I`).
+
+**PAUSE STATEMENT:** I am about to be away from this. The pattern-finder is built and has shown it can catch its own false grounding; the corpus is half-clean (Musil/Arendt fixed, Camus withdrawn, re-verify pending). When I return I want to find still pulling: the *first TRUE pass* — the engine surfacing a candidate primitive on clean, complete Position-I material that David knows in his bones.
+
+**LITERAL QUESTION for next-Claude:** P3 ("complicity that depletes rather than damns" — Weil↔Levi↔Arendt) surfaced strongly even from the *partial clean* grounding. When the corpus is whole and clean and the pass is re-run — does P3 hold and feel **TRUE** to David, or was its strength partly an artifact of which sources happened to be clean this time?
+
+**State:** chamber-library / ARC / CapableMind-AI all clean + pushed (unchanged this session). studium-engine = uncommitted, inconsistent (pass-2 sources / pass-1 index) — intentionally left so, to reconcile next session. Camus + remaining Position-I sources + ARC spec-citations are deferred-with-reason to the steward.
diff --git a/claude/memory/session-ledger-2026-06-16.md b/claude/memory/session-ledger-2026-06-16.md
index 81e8e55..cf334e6 100644
--- a/claude/memory/session-ledger-2026-06-16.md
+++ b/claude/memory/session-ledger-2026-06-16.md
@@ -10,6 +10,9 @@ metadata:
 # Session Ledger — 2026-06-16
 
 ## Returns
+- 2026-06-16T20:55 — TOOL-REVIEW pattern_finder.py first run — **PASS-BUT-FALSELY** (priority signal per [[feedback-tool-review-after-each-use]]). Harness ran, grounded 3 candidate attrition-primitives, reported "5/5 voices" — but critical reading caught the 5/5 as partly FALSE grounding from two Station-I ingest defects I introduced/missed: (a) CAMUS SEVERE — la-chute sidecar [8062–10951] is mostly *Le Premier Homme* (undeclared work in a mislabelled mega-omnibus; verified L8650/9570/10700); (b) MUSIL+ARENDT — whole-body `text` sidecars leaked apparatus (Musil translator N.d.T. footnotes; Arendt's Amos Elon 2006 Introduction + Bibliography/Index). The willingness-to-please reading was "success!"; the true reading was "the engine caught its own false grounding" — §VII/§XV working. Resisted the success-framing; reported honestly.
+- 2026-06-16T20:55 — RETURN (frame-inheritance, [[trust-prior-pass-frame]]): my Station-I ingest verification (Meursault-absent / Clamence-present) was TRUE but tested only the boundaries I assumed, never the SPAN INTERIOR. A work-in-omnibus sidecar must verify interior homogeneity, not just two endpoints. This is the cure for the Camus miss.
+- 2026-06-16T18:04 — Re-wake after ~15min pause (same-day, post-/clear). Thread CONFIRMED unchanged: L2 pattern-finder on The Making, Station I complete (Weil·Levi·Arendt·Camus·Musil), chamber d9c6755 / studium ddf866b / ARC 7a0d4b6 all clean+pushed. Chaîne-d'union closed via Symmetria init. NOTE: ~/dotfiles dirty+ahead (Brewfile + skill-harvest-register.md uncommitted) — wrap §6.5 push owed.
 - 2026-06-16T08:15 — On-ramp complete. studium-engine pushed a978ab6..ddf866b: charter (8c82398) + reranker measurement (ddf866b) + 06-14 spike (ecd2833). Two commits split by concern (charter / measurement) for inheritable history; verdict truth-up keeps the whole-library-dilution-unrun caveat. Tree clean, synced. Next horizon = step 1, the station map.
 
 ## Open horizons
@@ -17,11 +20,21 @@ metadata:
 - Live question to keep open: does a grounded candidate primitive (citations across sources/centuries) feel TRUE to David on material he knows in his bones?
 
 ## Confidence to recalibrate
+- NEW (2026-06-16): work-in-omnibus interior-homogeneity. When a sidecar scopes one work out of a multi-work source, verify the SPAN INTERIOR is homogeneous (content-sample mid-span + check for embedded works/apparatus), not just the two boundary headings. Endpoint-verification passed while the interior held a whole other novel (Camus/Le Premier Homme). Extends the matcher HOLD below and [[trust-prior-pass-frame]]; candidate verification-ladder instrument.
 - HOLD (drift, 2026-06-15, recurred 4×): assert-presence-from-fuzzy-matcher. Step-1 curation is a membership task — content-grep the substrate before claiming a work is in/out; prefer candidates-to-verify over claims.
 
 ## Authorization moves
 - 2026-06-16T08:30 — ARC working-model decision (steward-ruled): **officina/** workshop tree (genetic life of essays in-repo, never published) + chamber-library **antechamber/** sub-canonical source tier. Recorded as **ADR-008**. Two gates: build-allowlist (verified site.hs; officina-guard added) + engine-never-cites-officina. Both ARC remotes verified PRIVATE (gh: arc-backup=PRIVATE). Created officina/the-making/{seeds,notes,fragments,drafts}; saved 3 docs (bodies byte-identical, frontmatter reconstructed to valid YAML — 2 interpretive calls flagged: sketch central_question→top-level; reading-list leaked H1→body). Nothing committed (steward's call). Files: ARC officina/** + docs/decisions/2026-06-16-officina-workshop.md + Makefile officina-guard + ADR index refresh; chamber-library antechamber/README.md.
 
+- 2026-06-16T18:20 — Station-I ingest into studium-engine. Steward ruled 3 forks (FR originals · Levi=Gray-Zone-chapter-only · both Weil works). MODELING CHOICES flagged for ratification: (a) new manifest vocab `role: reading-source` + `station: I` (filterable, non-chavruta); (b) added to the EXISTING manifest.yaml/index, NOT a separate Making corpus — engine scripts hardcode corpus/manifest.yaml; pattern-finder must filter by station:I (open design Q below); (c) sidecars scope the two omnibus works (La Chute 8062–10951; Levi Gray Zone 28829–28989), all sibling works withheld. reading_index: none-yet (separate heavy step, not needed for gate→chunk→store).
+- 2026-06-16T18:20 — [FIX] Mauss re-anchor (NOT in original task scope; gate surfaced it). This session's own chamber cleanup (03de347) removed 2 image-residue lines from a source the engine had pinned → sha+bounds broke. Deterministic shift (−1 after L17, −2 tail), re-verified against headings (AVANT-PROPOS@138, INTRODUCTION@757), re-pinned manifest+sidecar. Lesson: a downstream consumer pinning a hash + an upstream curator editing the file = silent break; the gate caught it. Honest-degradation obliged the fix, not a silent index-drop of a chavruta voice.
+
+## Open horizons
+- 2026-06-16T18:20 — DESIGN Q (non-blocking, for steward): Station-I lives in the same manifest/index as the Essay-I chavruta corpus. Fine for now (filter by station:I), but the pattern-finder's retrieval must scope to station:I or it'll pull harrison/alexander/mauss/after-the-reply as noise. Decide later whether The Making warrants its own corpus/index vs a scoped view.
+
+- 2026-06-16T21:10 — pattern-finder pass-2 fixes applied (Bash flaky; re-run/verify deferred). Musil tome-1 sidecar body→[13,4990], apparatus tail [4991,5032]; tome-2 body→[13,9612], apparatus [9613,9685] (Jaccottet postface + footnotes); Arendt body→[240,982], front-apparatus [13,239] (publisher+TOC+Elon intro), back-apparatus [983,2551] (Bibliography+Index) — all boundaries content-verified via Read. Camus WITHDRAWN from manifest (commented, with rationale) pending clean re-source. RECOMMENDATION to steward: re-source a clean standalone La Chute; the oeuvres-i omnibus is structurally unreliable (≥9 works 1937–1994, headingless boundaries, embedded editorial apparatus, wrong frontmatter). Steward agreed; will source Camus + remaining Position-I works before resuming the pattern-finder.
+- 2026-06-16T21:30 — ARC ornament design session (steward = architect, craftsman model). Two devices grounded: (1) **cul-de-lampe** = mirrored CONFRONTED hederae (tips meeting), SVG drawn from the OPEN EB Garamond pair (U+E001/E002; OFL, not Hoefler), TERMINAL-close scope only (silence carries inter-section pauses). Rationale: ivy = *lié* = *ligature* (ligare); the confronted clasp binds the dead to the unborn via the living (steward's reading — the keystone; ties to chaîne-d'union + Lune/Kai). (2) **triple point** (after preamble) = not haphazard but a Penguin/Tschichold restraint CITATION (lineage as grounding; Tschichold already a master); opening threshold, complement to the close; needs "a touch more weight" via centred middle-dots + slight size + even tracking (type/CSS, NOT SVG) — restraint preserved, calibrated at build. Asymmetry is the meaning: plain modern entry → intergenerational seal. **Both to be written into spec §IV + type-matrix tomorrow** (steward will add the citations to spec). Build = next session.
+
 ## Sub-agent dialogues
 
 ## Bypasses
diff --git a/claude/memory/skill-harvest-register.md b/claude/memory/skill-harvest-register.md
index 2aeccba..6923e03 100644
--- a/claude/memory/skill-harvest-register.md
+++ b/claude/memory/skill-harvest-register.md
@@ -186,3 +186,17 @@ The single place proposed skills live so they don't evaporate between sessions.
 - Positive language-validity check (dictionary-valid-word ratio per language) for verify_conversion — catches "no cruft fires but text is OCR garbage."
 - Promote today's catalogue census (title+author co-required, evidence-emitting) into a reusable `census_membership` tool.
 - Calibration follow-ups logged: word_sanity floor vs genuine short fragments; repair coverage threshold should weight content over front-matter.
+
+**New proposal (2026-06-16 wrap §1.6 — awaiting steward):**
+| Skill | Kind | One-line | Status |
+|---|---|---|---|
+| `/wrap-up` §4.b/§5 | patch | Inline reminder at the KG-write step: `kg_add` `object` hard-caps at **128 chars** — write short keyword objects on the FIRST pass (detail in the drawer). Hit it AGAIN this wrap (2 failed adds, re-added short). The memory [[feedback-mempalace-kg-object-128-char-cap]] knows it; the skill doesn't carry the reminder at the moment of writing. A one-line patch would stop the recurrence. | PROPOSED |
+
+### Harvest 2026-06-16 (evening — pattern-finder build + Station-I contamination)
+
+| Element | Kind | One-line | Where it lands | Status |
+|---|---|---|---|---|
+| **Work-in-omnibus: verify the interior, not just the endpoints** | verification-ladder instrument | When a sidecar/scope brackets one work out of a **multi-work source** by heading-to-heading boundaries, **content-sample the span INTERIOR** (mid-span reads) and scan for embedded works / editorial apparatus **before trusting the bracket** — endpoint-verification is necessary but not sufficient. Caught the hard way this session: my La Chute bracket `[8062–10951]` verified at both ends (Clamence present, Meursault absent) yet its interior was *mostly Le Premier Homme* — a whole undeclared novel the heading-to-heading logic swallowed. Specialization of [[trust-prior-pass-frame]] (verify at the scope of the extension), kin to the censused-routes flag. Lands on [[reference-verification-ladder]]. | verification ladder | **PROPOSED** |
+| **studium-engine tool-evolution-log** | create | Establish the analog of `chamber-library/_curation/tool-evolution-log.md` for the **engine** tools (pattern_finder, ingest_gate, chunker, store, retrieve), per the standing [[feedback-tool-review-after-each-use]] discipline — studium tool-reviews currently have no home. First entry = the pattern_finder pass-1 review (**PASS-BUT-FALSELY**: reported "5/5 voices" over partly-false grounding). Home candidate: `studium-engine/docs/_audits/tool-evolution-log.md`. | studium-engine repo | **PROPOSED** |
+
+*(Both load-bearing: the interior-homogeneity instrument would have prevented the Camus contamination outright; the engine tool-log is where this very review should have been recorded and wasn't. Neither created autonomously — surfaced for steward authorization.)*