diff --git a/components/chain-table.tsx b/components/chain-table.tsx
index 0ce7128..63c6a2a 100644
--- a/components/chain-table.tsx
+++ b/components/chain-table.tsx
@@ -272,7 +272,9 @@ export function ChainTableBody({
 			    statements pl-21 where a chain icon would sit).
 			    All three columns sort, small chevron on active. */}
 			<div className="min-h-0 flex-1 overflow-y-auto">
-				<div className="sticky top-0 z-10 flex items-center gap-1 px-3 py-1.5 text-[11px] uppercase tracking-widest text-muted-foreground/70 backdrop-blur">
+				{/* The TOP rule (owner round 5: the table is framed — a line under
+			    the header, a line under the last group, lines between groups). */}
+				<div className="sticky top-0 z-10 flex items-center gap-1 border-b border-border/60 bg-background/80 px-3 py-1.5 text-[11px] uppercase tracking-widest text-muted-foreground/70 backdrop-blur">
 					<button
 						type="button"
 						data-testid="sort-name"
@@ -308,17 +310,29 @@ export function ChainTableBody({
 					</div>
 				)}
 
-				{groups.map((group, groupIndex) => {
+				{groups.map((group) => {
 					const open = searching || expandedClusters.has(group.key);
 					const selectedCount = group.nodeIds.filter((id) => selection.has(id)).length;
 					return (
 						<div
 							key={group.key}
 							data-testid={`cluster-group-${group.key}`}
-							// The one line that survives (owner ruling): the separator
-							// BETWEEN cluster groups — never on individual rows.
-							className={groupIndex < groups.length - 1 ? "border-b border-border/60" : ""}
+							// The horizontal rules (owner rulings, rounds 2 + 5): between
+							// cluster groups AND under the last one — the table is framed
+							// top (header) and bottom. Never on individual rows.
+							className="relative border-b border-border/60"
 						>
+							{/* The TREE THREAD (owner round 5, trial): a vertical guide
+							    dropping from the cluster's expander arrow down through
+							    its chain rows — the arrow's column (center 37px: pl-3 +
+							    checkbox 14 + gap 4 + half the 14px expander), from the
+							    arrow's own height to the last row's middle. */}
+							{open && group.chains.length > 0 && (
+								<div
+									data-testid={`cluster-thread-${group.key}`}
+									className="pointer-events-none absolute bottom-3.5 left-[37px] top-4 w-px bg-border/50"
+								/>
+							)}
 							{/* The cluster row: birth-session group header. */}
 							<div className="flex items-center gap-1 px-3 py-1.5">
 								<RowCheckbox
@@ -353,7 +367,16 @@ export function ChainTableBody({
 									const chainOpen = expandedChains.has(chain.id);
 									const selectedInChain = memberIds.filter((id) => selection.has(id)).length;
 									return (
-										<div key={chain.id} data-testid={`chain-block-${chain.id}`}>
+										<div key={chain.id} data-testid={`chain-block-${chain.id}`} className="relative">
+												{/* The chain's own thread: from its expander arrow down
+												    through the statement rows (center 73px: pl-12 +
+												    checkbox 14 + gap 4 + half the expander). */}
+												{chainOpen && memberIds.length > 0 && (
+													<div
+														data-testid={`chain-thread-${chain.id}`}
+														className="pointer-events-none absolute bottom-3.5 left-[73px] top-3.5 w-px bg-border/50"
+													/>
+												)}
 											<div
 												data-testid={`chain-row-${chain.id}`}
 												aria-label={formatChainLine(chain, Object.fromEntries([...byId.entries()].map(([id, n]) => [id, { name: n.name }])))}
diff --git a/docs/architecture.md b/docs/architecture.md
index aa602ee..0c2248f 100644
--- a/docs/architecture.md
+++ b/docs/architecture.md
@@ -382,7 +382,7 @@ Practice turns the graph into problem sets the learner reasons through out loud.
 
 **The jobs pattern — registry, not JobRecords.** Practice runs are NOT Workbench jobs: no `JobStore` records, no steer messages, nothing ever wakes the teacher — a practice run isn't chat-scoped (a set outlives its birth session and has no lesson to re-enter; its result is itself, persisted in its own store and read by its own page). What it shares with the Workbench is the observability surface: a `globalThis` run registry (the module-instance idiom — `__lhPracticeRuns`) with PER-PROBLEM ROWS (owner ruling, 2026-09-29 — matching the extraction per-chunk pattern): `setId:partition` while the partitioner runs ("Partitioning — title"), then one `setId:gen:problemId` per problem ("queued"→"generating", label "Problem i/N — title"), and one `setId:grade:problemId` per graded answer ("queued"→"grading", "Grading i/N — title") — each row its OWN abort controller and its OWN usage accumulation; the row key IS the feed id (the dock dup-key law) and the abort target. ABORT ROUTES BY ROW SHAPE: a gen-row abort kills THAT problem only (siblings continue; the set still settles `ready` with one errored problem), a partition-row abort ends the whole run, a grade-row abort drops that grade (answer kept — one attempt, the problem stays done); set-level abort (delete) composes from row aborts plus the synchronous persisted marks. Model calls — partitioner included — run under the SHARED background-LLM semaphore (`bg-llm-semaphore.ts`): one process-wide pool of 3 permits for ALL background LLM jobs, extraction chunks included (total background LLM concurrency bounded at 3, whoever calls). A boot sweep marks orphaned `partitioning`/`generating` sets as `error "Interrupted (server restarted)"` — once per process, and only when the set's `updatedAt` heartbeat is >120s stale (a live run's 30s FILE watchdog keeps the heartbeat fresh even across dev's per-worker processes — the registry is not shared between them; the guard trio closed a live-stamping bug from the first live test) — the job runner's crash-honesty at set grain. Store writes serialize behind a per-set mutex; the pipeline is the sole writer of `data/practice`.
 
-**API + UI.** Creation is GRAPH-DOMAIN, not chat-domain (owner ruling 2026-09-29: "problem sets always take in nodes — graph objects"): `POST /api/practice/sets` takes `{nodeIds, params, title?, note?}` ONLY — no chatId, no clusterId; the UI resolves a picked cluster to its member node ids and passes the cluster's label as the title. The `note` is the learner's optional free-form creation input (owner, same evening: more learner control, no other purpose): trimmed, capped at 2000 characters, persisted on the set, and carried to the partitioner and every generator as `THE LEARNER'S NOTE` — an additional input the agents WEIGH but are never obliged to follow (the doctrine pins the stance: it never moves a problem off its rung or out of its band, and the partition laws always win). `GET /api/graph/clusters` serves the picker: the birth-session partition itself (`birthSessionPartition`, same ordering — size desc, ties by id) — EVERY group regardless of size (owner ruling, same evening: MIN_CLUSTER_SIZE is a hull-drawing rule; practice availability must not inherit a visual threshold), each as `{key, kind: "chat"|"notes", label, nodeIds, nodeCount}`; a cluster whose naming record is gone keeps `label: null` — honest, never guessed. 202 `{id}` with generation fire-and-forget; 400 empty selection (empty intersection with live nodes — the vanished-nodes edge) / bad params; 503 model/key refusal. `GET /api/practice/sets` · `GET/DELETE /api/practice/sets/[id]` (delete aborts live runs first) · `POST /api/practice/sets/[id]/answers` (202; 409 when not ready; 400 re-answering a submitted problem — one attempt). Surfaces: `/practice` (the list — THE shared **TopBar** (`components/top-bar.tsx`, standardization ruling 2026-09-29: one element, every surface — h-15 matching the sidebar's wordmark band, `lh-drag` with `lh-no-drag` carve-outs on interactive children, `select-none` so drags never highlight, the SidebarToggle reopen with its Electron traffic-light gutter, `px-3` + one border line; chat's rename mode passes `drag={false}` — the one escape hatch) with the New action at the right edge, status/counts rows, delete confirm, the new-set dialog with the named-level controls (a 5-detent ladder slider, a 3-detent cue slider, a Theory|Application toggle — each showing the level's name + definition), the cluster picker, and the optional free-form note textarea (typing-only). The SECOND creation entry (owner redesign, same evening): the graph view's **chain table** — a `Chains` button in the graph chrome (empty graph: hidden) opens a FULL-BLEED table overlay (absolute inside the graph area; the blur + translucent background IS the surface, no inner card; the top bar is THE shared TopBar — the standard title size and drag mechanics, the sidebar reopen, and the close X at the right edge, so home's window-drag mask never eats it (the overlay covers home's floating header while open, and the TopBar's own toggle is the reopen path); Esc or X dismiss). The hierarchy is three levels, one click each: CLUSTER (the birth session — a chat or a notes file; a chain lives where its CONCLUSION was born) → REASONING CHAIN → PROPOSITION STATEMENTS — expanding a chain reveals its members' statement text directly (names live in the chain line, never twice; a singleton reveals its one statement in a single click); everything starts collapsed. The table look (owner refinements, evening round 2): PROMINENT indentation carries the hierarchy (chains deep under their folder, statements deepest) and the ONLY lines left are the horizontal separators between cluster groups — per-row lines and the vertical column hairlines are gone ("anything else is better without them"); statement rows are PLAIN (the derivation already lives in the chain line — no conclusion marker). ALL THREE columns sort (owner: "consistency is key") — Name (A→Z default), Props (largest-first), Created (newest-first) — click toggles asc/desc, ordering both cluster rows and chains within; a SMALL chevron marks the active header. In the chain line every proposition name is a faint-outlined CHIP with the → a clearly separated connector between chips — a lonely chip reads as a standalone proposition (the "(no premises)" annotation is dead in the visible UI; it survives only as the row's aria-label); icons: ONLY the cluster FolderIcon (open-folder when expanded) survives (owner ruling, evening round 3: the chain waypoints icons are not needed — the chips + connectors already read as chains); chevrons rotate with expansion and share the checkbox's 14px width, which makes every column exactly 18px and the ALIGNMENT STAIRCASE exact (owner rulings, rounds 3–4): each level's checkbox sits under its parent's ICON slot — open a folder and the chain checkboxes line up with the FOLDER icon; the statements step to where a chain icon WOULD sit (the steps are TWO columns = 36px — "double the indentation"): clusters pl-3 → chains pl-12 → statements pl-21. Selection is checkboxes on whole clusters, chains, or single propositions — a flat node-id set (a node in several chains is one selection) — accumulated in a FLOATING BOTTOM BAR (the Google-Photos pattern: appears only when something is selected, never shifts the table layout) showing the agents' ACTUAL input (`chainsForSelection` over the selection: a premise picked without its conclusion counts as the singleton it becomes) as `N propositions → M chains`, with Clear and **Generate problem set** (the one action for now — the bar is a neutral action slot), which opens the SAME creation dialog in **custom mode**: no cluster picker — a selection summary (the derived shape + the dated default title the pipeline will stamp) — posting `{nodeIds, params, note}` with no title. A fuzzy search (names, aliases, statements) filters chains grouped + auto-expanded under their clusters (non-matching clusters hide; the collapsed view returns on clear). NO understanding-levels column (explicitly not built). The overlay is read-only toward the graph and `/practice/[id]` (the exam page — the shared TopBar, a levels + status line, numbered problems, markdown statements with math healing at the display seam, a textarea + focus-gated dictation bar per problem with its OWN Submit button — one attempt per problem, instant solution/rubric reveal, the score landing live). Entry: the sidebar's Practice nav row.
+**API + UI.** Creation is GRAPH-DOMAIN, not chat-domain (owner ruling 2026-09-29: "problem sets always take in nodes — graph objects"): `POST /api/practice/sets` takes `{nodeIds, params, title?, note?}` ONLY — no chatId, no clusterId; the UI resolves a picked cluster to its member node ids and passes the cluster's label as the title. The `note` is the learner's optional free-form creation input (owner, same evening: more learner control, no other purpose): trimmed, capped at 2000 characters, persisted on the set, and carried to the partitioner and every generator as `THE LEARNER'S NOTE` — an additional input the agents WEIGH but are never obliged to follow (the doctrine pins the stance: it never moves a problem off its rung or out of its band, and the partition laws always win). `GET /api/graph/clusters` serves the picker: the birth-session partition itself (`birthSessionPartition`, same ordering — size desc, ties by id) — EVERY group regardless of size (owner ruling, same evening: MIN_CLUSTER_SIZE is a hull-drawing rule; practice availability must not inherit a visual threshold), each as `{key, kind: "chat"|"notes", label, nodeIds, nodeCount}`; a cluster whose naming record is gone keeps `label: null` — honest, never guessed. 202 `{id}` with generation fire-and-forget; 400 empty selection (empty intersection with live nodes — the vanished-nodes edge) / bad params; 503 model/key refusal. `GET /api/practice/sets` · `GET/DELETE /api/practice/sets/[id]` (delete aborts live runs first) · `POST /api/practice/sets/[id]/answers` (202; 409 when not ready; 400 re-answering a submitted problem — one attempt). Surfaces: `/practice` (the list — THE shared **TopBar** (`components/top-bar.tsx`, standardization ruling 2026-09-29: one element, every surface — h-15 matching the sidebar's wordmark band, `lh-drag` with `lh-no-drag` carve-outs on interactive children, `select-none` so drags never highlight, the SidebarToggle reopen with its Electron traffic-light gutter, `px-3` + one border line; chat's rename mode passes `drag={false}` — the one escape hatch) with the New action at the right edge, status/counts rows, delete confirm, the new-set dialog with the named-level controls (a 5-detent ladder slider, a 3-detent cue slider, a Theory|Application toggle — each showing the level's name + definition), the cluster picker, and the optional free-form note textarea (typing-only). The SECOND creation entry (owner redesign, same evening): the graph view's **chain table** — a `Chains` button in the graph chrome (empty graph: hidden) opens a FULL-BLEED table overlay (absolute inside the graph area; the blur + translucent background IS the surface, no inner card; the top bar is THE shared TopBar — the standard title size and drag mechanics, the sidebar reopen, and the close X at the right edge, so home's window-drag mask never eats it (the overlay covers home's floating header while open, and the TopBar's own toggle is the reopen path); Esc or X dismiss). The hierarchy is three levels, one click each: CLUSTER (the birth session — a chat or a notes file; a chain lives where its CONCLUSION was born) → REASONING CHAIN → PROPOSITION STATEMENTS — expanding a chain reveals its members' statement text directly (names live in the chain line, never twice; a singleton reveals its one statement in a single click); everything starts collapsed. The table look (owner refinements, evening round 2): PROMINENT indentation carries the hierarchy (chains deep under their folder, statements deepest) and horizontal rules FRAME the table (owner round 5) — between cluster groups, under the header, and under the last group — with per-row lines and the vertical column hairlines still gone, and open folders/chains dropping a TREE THREAD (owner round 5, trial): a vertical guide from the expander arrow down the children's edge, at the arrow's column (37px clusters / 73px chains); statement rows are PLAIN (the derivation already lives in the chain line — no conclusion marker). ALL THREE columns sort (owner: "consistency is key") — Name (A→Z default), Props (largest-first), Created (newest-first) — click toggles asc/desc, ordering both cluster rows and chains within; a SMALL chevron marks the active header. In the chain line every proposition name is a faint-outlined CHIP with the → a clearly separated connector between chips — a lonely chip reads as a standalone proposition (the "(no premises)" annotation is dead in the visible UI; it survives only as the row's aria-label); icons: ONLY the cluster FolderIcon (open-folder when expanded) survives (owner ruling, evening round 3: the chain waypoints icons are not needed — the chips + connectors already read as chains); chevrons rotate with expansion and share the checkbox's 14px width, which makes every column exactly 18px and the ALIGNMENT STAIRCASE exact (owner rulings, rounds 3–4): each level's checkbox sits under its parent's ICON slot — open a folder and the chain checkboxes line up with the FOLDER icon; the statements step to where a chain icon WOULD sit (the steps are TWO columns = 36px — "double the indentation"): clusters pl-3 → chains pl-12 → statements pl-21. Selection is checkboxes on whole clusters, chains, or single propositions — a flat node-id set (a node in several chains is one selection) — accumulated in a FLOATING BOTTOM BAR (the Google-Photos pattern: appears only when something is selected, never shifts the table layout) showing the agents' ACTUAL input (`chainsForSelection` over the selection: a premise picked without its conclusion counts as the singleton it becomes) as `N propositions → M chains`, with Clear and **Generate problem set** (the one action for now — the bar is a neutral action slot), which opens the SAME creation dialog in **custom mode**: no cluster picker — a selection summary (the derived shape + the dated default title the pipeline will stamp) — posting `{nodeIds, params, note}` with no title. A fuzzy search (names, aliases, statements) filters chains grouped + auto-expanded under their clusters (non-matching clusters hide; the collapsed view returns on clear). NO understanding-levels column (explicitly not built). The overlay is read-only toward the graph and `/practice/[id]` (the exam page — the shared TopBar, a levels + status line, numbered problems, markdown statements with math healing at the display seam, a textarea + focus-gated dictation bar per problem with its OWN Submit button — one attempt per problem, instant solution/rubric reveal, the score landing live). Entry: the sidebar's Practice nav row.
 
 **Model wiring.** Three roles — `partitioner`, `problemGenerator`, `grader` — all pinned to the **strong** slot (quality-critical, one settings edit from cheap) and in `BACKGROUND_MODEL_ROLES` (per-role overrides). The stage agents are single-submit loops on the shared `runSubmitAgent` seam (`submit-agent.ts`, extracted from extraction for exactly this): a `validate` hook turns a bad submit into a tool-input error the model retries against, within `maxTurns 4`; exhaustion fails the stage honestly (the set says so, the chains re-run cleanly on a new set).
 
diff --git a/tests/render-chain-table.test.tsx b/tests/render-chain-table.test.tsx
index 1f2874c..7e29d00 100644
--- a/tests/render-chain-table.test.tsx
+++ b/tests/render-chain-table.test.tsx
@@ -133,16 +133,28 @@ describe("ChainTableBody — THE shared TopBar + the lines law (owner, round 2)"
 		assert.ok(html.includes('data-testid="chain-table-search"'), "the search input");
 	});
 
-	it("the ONLY lines are the separators between cluster groups — no per-row or vertical lines", () => {
+	it("the table is FRAMED: header rule + rules between groups + rule under the last group; rows stay lineless (owner round 5)", () => {
 		const html = render({ expandedClusters: new Set(["chat1"]) });
-		// group wrappers: every group but the last carries the separator
+		// every group wrapper carries the rule now — between groups AND under
+		// the last one; the header carries the TOP rule
 		const sep = html.split("border-b border-border/60").length - 1;
-		assert.ok(sep >= 2, "separators between groups");
+		assert.ok(sep >= 4, "header + 3 group rules");
 		// chain rows themselves carry NO border — their entire class list is
 		// the staircase law (pl-12 = aligned with the cluster's FOLDER icon);
 		// the chip outlines' border-border/70 is an outline, not a row line
 		assert.ok(html.includes('class="flex items-center gap-1 py-1 pl-12 pr-3"'), "the chain row: staircase indent, no border");
 	});
+
+	it("TREE THREADS: open folders/chains drop a vertical guide from their expander arrow down the children's edge (owner round 5)", () => {
+		const collapsed = render();
+		assert.ok(!collapsed.includes("cluster-thread-"), "no thread while collapsed");
+		const open = render({ expandedClusters: new Set(["chat1"]), expandedChains: new Set(["c1"]) });
+		assert.ok(open.includes('data-testid="cluster-thread-chat1"'), "the cluster thread");
+		assert.ok(open.includes('data-testid="chain-thread-c1"'), "the chain thread");
+		// the columns: the cluster arrow sits at 37px, the chain arrow at 73px
+		assert.ok(open.includes("left-[37px]"), "thread at the cluster arrow's column");
+		assert.ok(open.includes("left-[73px]"), "thread at the chain arrow's column");
+	});
 });
 
 describe("ChainTableBody — chain rows (chips + connectors)", () => {