diff --git a/docs/architecture-invariants.md b/docs/architecture-invariants.md
index 233e0fda..34f7c6fc 100644
--- a/docs/architecture-invariants.md
+++ b/docs/architecture-invariants.md
@@ -191,7 +191,7 @@ Further detail (current geometry and colors, superseding the dip numbers above w
 
 **Three owners, one setter.** `SessionState.nameSource` is `placeholder` | `auto` | `manual`. The constructor infers a missing value from the name (`isGeneratedSessionName()` = `w<n>-<case>` / `s<n>-<case>`, or no name at all, is a placeholder; anything else was a person's), the create routes pass none, the boot restore passes the persisted one. The `name` setter is the manual path and the ONLY thing that produces `manual` after construction; `applyAutoName()` is the only thing that produces `auto`, and it does so whether or not the string changed, which is what makes "first prompt" mean the first. A prompt whose title is null (`/clear`, `! npm test`, blank) never reaches it, so the session stays eligible: the first REAL prompt names the tab.
 
-**The origin gate defaults to "system".** `write()` / `writeViaMux()` take `SessionWriteOptions.fromUser`; only the browser WS path and `POST /api/sessions/:id/input` set it. Every other caller (Ralph, respawn, cron, approvals, the orchestrator's `/compact`, auto-ops, the trust-dialog keys) is system by omission, so a new user-input path that forgets the flag fails toward a tab that keeps its placeholder, never toward a tab named after a Ralph prompt. `_lastSubmitAt` is still stamped for every write; only the tracker feed is gated. The shell gate is `getCli(mode)?.capabilities.startMode !== 'shell'`, a capability rather than an id check (the no-id-branching guard), and the send-key route feeds `trackUserInput()` by hand because its `tmux send-keys -H` newline (a line feed unless the CLI declares another `capabilities.newline` chord) never passes through the session.
+**The origin gate defaults to "system".** `write()` / `writeViaMux()` take `SessionWriteOptions.fromUser`; only the browser WS path and `POST /api/sessions/:id/input` set it. Every other caller (Ralph, respawn, cron, approvals, the orchestrator's `/compact`, auto-ops, the trust-dialog keys) is system by omission, so a new user-input path that forgets the flag fails toward a tab that keeps its placeholder, never toward a tab named after a Ralph prompt. `_lastSubmitAt` is still stamped for every write; only the tracker feed is gated. The shell gate is `getCli(mode)?.capabilities.startMode !== 'shell'`, a capability rather than an id check (the no-id-branching guard), and the send-key route feeds `trackUserInput()` by hand because its `tmux send-keys -H` newline (always a line feed for Ctrl+Enter; for Shift+Enter a line feed unless the CLI declares another `capabilities.newline` chord) never passes through the session.
 
 **The tracker is a best-effort transcript with explicit per-key rules**, not a byte filter. Mirrored: printable text, backspace, Ctrl+W, Ctrl+U/Ctrl+C (composer emptied), `\n` and Alt+Enter (a newline IN the composer, joined with a space), bracketed paste (newlines inside it likewise). Ignored: cursor keys, Home/End/Delete, Shift+Tab, Tab, SGR mouse and focus reports, Alt chords, OSC/DCS, the rest of C0. Tainting: Up/Down (CSI and SS3), Ctrl+P/N/R, Ctrl+_, because the composer then holds a history line the tracker never saw and Enter must submit nothing rather than a fragment. A bare Esc is resolved at the END of the chunk it arrives in, since xterm hands each key's whole sequence to one write and the programmatic senders send Esc alone; a CSI split across chunks still resumes. The draft keeps its HEAD past 8192 code points (the title is the first sentence, so keeping the tail would title a long paste by its last line) and an escape sequence is abandoned past 64 bytes.
 
@@ -664,11 +664,13 @@ Matching semantics: a query containing `/` matches the relative path, otherwise
 
 **Command palette + shortcut registry** (COD-151/153/157/192, #146): `Ctrl/Cmd/Alt+K` opens the session palette (fuzzy search over live sessions; "Browse all sessions" → the Session Manager modal backed by `GET /api/sessions/unified`); the quick-start case `<select>` is fronted by a searchable picker (`buildCasePickerOptions`/`formatCasePickerLabel` — remote cases render `name @ hostId`). Shortcuts live in a rebindable registry (`DEFAULT_SHORTCUTS`/`getShortcutRegistry()`/`matchesShortcutEvent()` in app.js; overrides persist under `settings.shortcutOverrides` via `saveAppSettingsToStorage`); App Settings → Shortcuts renders capture/disable rows; `Ctrl+?` opens the registry-driven overlay (footer links to the full `#helpModal` reference). ⚠️ Palette-chord keys must ALSO be swallowed in `attachCustomKeyEventHandler` (terminal-ui.js) or xterm writes the control byte (0x0B) into the PTY. ⚠️ The global Escape handler (app.js) calls every overlay close method on every Escape, in the capture phase and so before the focused element's own Escape handler, which means a close method with side effects beyond hiding must return early when its overlay is not open: `closeCommandPalette()` and `closeSessionManager()` hand focus back on close (`_restoreOverlayFocus()`, panels-ui.js), and run unconditionally that restore stole focus from split Pane B and from any text field, and turned the inline tab rename's Escape into a commit. ⚠️ `saveAppSettings()` rebuilds settings from the DOM — keys edited elsewhere (`shortcutOverrides`, `showTokenCount`, `showCost`) need explicit `_prev` carry-over.
 
+**A key gate on a chord that can carry Shift alone must swallow the KEYPRESS too** (#520). xterm calls the custom handler from `attachCustomKeyEventHandler` for `keydown`, `keypress` AND `keyup`, and its own `_keyPress` drops a keypress carrying Ctrl, Alt or Meta but not one whose only modifier is Shift. So a gate that returns `false` for `ev.type === 'keydown'` alone is safe for a Ctrl or Alt chord (the palette's Ctrl+K, Alt+B, Ctrl+V, Ctrl+Z and the keydown-only smart copy below) and leaks for a Shift-only one. That is how Shift+Enter submitted the prompt from April to #520: its keypress fell through to `onData` as a bare `\r`, which goes out synchronously over the WebSocket ahead of the async `send-key` fetch, so the CR reached the PTY before the newline did. Ctrl+Enter worked all along only because Chrome fires no keypress for it. ⚠️ The Shift/Ctrl+Enter gate in terminal-ui.js and its twin in Pane B (terminal-split.js) return `false` for EVERY event type and key only the `send-key` POST on `keydown`; never narrow either back to keydown, and give any new Shift-only chord the same shape. Swallowing the keyup as well is harmless (xterm only skips a focus and cursor-style refresh for that one event). Tests: `test/shift-enter-keypress-swallowed.test.ts` (CI gate, a source guard on both files) and `test/shift-enter-keypress.browser.test.ts` (browser suite, which the gate does not run: real keys against both real handlers in the real app). Settings → Terminal & Input → **Key tester** shows what a device reports for each of the three events.
+
 **Terminal smart copy** (#211): `Ctrl+C` copies the selection when there is one and stays the interrupt when there isn't. Three rules keep that split honest, and breaking any of them silently costs the user their interrupt key:
 
 1. The branch lives in `attachCustomKeyEventHandler` (terminal-ui.js) and the **no-selection path returns `true` with no `preventDefault()`**, so xterm still evaluates `Ctrl+C` into `0x03`. Returning `false` alone does not cancel the event either way: xterm's `_keyDown` calls the custom handler *before* its own `cancel()`, which is exactly why the copy path calls `preventDefault()` explicitly (otherwise the browser also runs its native copy on top).
 2. `copyTerminalSelection` is a registry `action` **deliberately missing from `SHORTCUT_ACTIONS`** (same trick as `command-palette`): the entry stays rebindable and disableable in App Settings, while the generic document-capture loop, which `preventDefault()`s every match it dispatches, skips it and lets the terminal handler decide.
-3. The gate is keydown-only (the custom handler also runs for `keypress`/`keyup`), and `Ctrl+Shift+C` never falls through to the PTY: an "explicit copy" chord that interrupts a running agent because the selection happened to be empty is a footgun with no upside.
+3. The gate is keydown-only (the custom handler also runs for `keypress`/`keyup`; that is safe here only because xterm drops a Ctrl keypress itself, see the keypress rule above), and `Ctrl+Shift+C` never falls through to the PTY: an "explicit copy" chord that interrupts a running agent because the selection happened to be empty is a footgun with no upside.
 
 Copy goes through `_copyText()` (Clipboard API, then hidden-textarea + `execCommand`), not raw `navigator.clipboard`, because `install.sh`'s LAN option serves plain HTTP where `navigator.clipboard` is undefined; the fallback steals focus, so the terminal is refocused afterwards. Related: xterm registers its own `copy` listener on the terminal element gated on `hasSelection()`, which is why right-click → Copy has always worked. Selection itself is unavailable on touch devices by design (`user-select: none` on the terminal subtree), and in `shell`/`opencode`/`antigravity` tabs the TUI owns the mouse, so selecting there needs Shift+drag. Tests: `test/terminal-copy-selection.test.ts` (gate + wiring invariants), `test/terminal-copy-shortcut.test.ts` (browser, real key presses).
 
@@ -789,7 +791,7 @@ Further detail: with many sessions the horizontal strip stops being scannable, w
 
 ### Split-pane sessions
 
-**Split-pane sessions** (`showSplitButton`, header button, default OFF, per-device like `showFileViewerButton` — not in `SettingsUpdateSchema`, `displayKeys` in settings-ui.js): shows two live sessions side-by-side in one Codeman window. Pane A is the untouched, existing singleton terminal (`this.terminal`/`this._ws` in terminal-ui.js); Pane B is a new, independent `SplitTerminalPane` (terminal-split.js) with its own xterm instance and its own `/ws/sessions/:id/terminal` WebSocket. ⚠️ **Pane B is deliberately plainer than Pane A** — no local-echo overlay, no CJK IME, no touch/mobile handlers, no keyboard accessory bar — since this is a desktop-only feature (a split view needs a wide viewport) and those features exist for mobile/touch input; `.btn-split` is hard-hidden below 1180px regardless of the setting by the `@media (max-width: 1179px)` rule in styles.css (mobile.css only carries a comment pointing at it: that file loads up to 1023px, so it cannot cover the 1024-1179px tablet range the feature also needs to stay off), and the per-device setting means turning it on at a desk can never sync it onto a phone in the first place. No persistence: closing the browser tab or reloading always returns to the normal single-pane view; there is no localStorage key for split state. ⚠️ Splitting a session against itself is disallowed (the picker excludes the active session), **as is splitting against a popped-out (detached) session** — `buildSplitPickerSessions()` excludes `detachedSessions` because a detached session's own window is already claiming its PTY size, and `MAX_WS_PER_SESSION` needs no change since Pane A/B are always two different sessions. ⚠️ Either pane's session ending (deleted locally or from another client) auto-collapses the split — Pane A's session ending promotes Pane B to the new single pane via `selectSession(id, { auto: true })` (an app-driven selection, so it must not spend the session's idle alert — see the Approvals Inbox note above), never by trying to hot-swap the lightweight `SplitTerminalPane` object into the primary singleton state. ⚠️ Pane B refits on every window/sidebar/tab-rail resize via the SAME trailing-edge `ResizeObserver` callback that resizes Pane A (`throttledResize` in terminal-ui.js) — it only ever measured Pane A's own container, so without an explicit `this._splitPane?.fit()` call there Pane B silently kept its stale PTY size through every resize that did not happen to be a divider drag. ⚠️ A dropped WebSocket leaves Pane B visibly dead (a message written into its own xterm buffer) rather than silently swallowing keystrokes with nothing on screen to explain why — there is no reconnect logic for v1, matching the "deliberately plainer than Pane A" design. Related but distinct: `detachSession()` already opens one session in a separate OS-level browser window (`isSoloWindow`) — that is prior art for "two sessions visible at once" but not for one window with a draggable in-page divider, which is what this feature adds. ⚠️ Pane B installs its own `attachCustomKeyEventHandler` gating the same app-level chords Pane A's own handler gates (command palette, Alt+1-9/[/] tab nav, Alt+B sidebar toggle, Ctrl+Z suspend, Shift/Ctrl+Enter newline, smart-copy Ctrl+C) — without it the document capture-phase handler's `preventDefault()` (which never stops xterm) let each chord ALSO write its raw byte/escape sequence into Pane B's live PTY on top of whatever the app action did to Pane A (COD-153). Ctrl+Z is swallowed unless Pane B's own session is `mode === 'shell'`, mirroring terminal-ui.js's reasoning: in a plain shell it is the user's own job-control tool, everywhere else it silently suspends an unattended agent loop. Shift/Ctrl+Enter POSTs to `/api/sessions/:id/send-key` (`{key:'S-Enter'|'C-Enter'}`, tmux `send-keys -H` for a real 0x0a, or the CLI's declared `capabilities.newline` chord) targeting THIS pane's own `sessionId` rather than the primary pane's `activeSessionId` — without it xterm's plain `\r` would submit an incomplete prompt instead of adding a line to it. Smart-copy Ctrl+C/Ctrl+Shift+C is re-implemented against `this.terminal` (Pane B's own) rather than reusing `app.copyTerminalSelection()`, which reads Pane A's terminal and would copy the wrong pane's selection; Ctrl+Shift+C never falls through even with nothing to copy, mirroring terminal-ui.js's own `ev.shiftKey` branch. ⚠️ **This is a UX-parity fix, not an interrupt-safety one** — verified live in a real browser: xterm's `evaluateKeyboardEvent` routes a shifted ctrl-letter into a branch that assigns `c.key` only for two special cases (`_`→US, `@`→NUL), so it emits no data for Ctrl+Shift+C at all regardless of any application gate; a synthetic keydown with the gate removed produces zero WS frames, proving no accidental interrupt reaches the PTY either way. What gating the whole copy block on `hasSelection()` (an earlier draft) actually cost: with no selection, a selection-less Ctrl+Shift+C fell straight to `return true`, silently ceding the keystroke to the BROWSER's own handling (e.g. Chrome's Inspect-Element binding) with no feedback and no copy attempt — Pane A always intercepts it. Ctrl+V stays on xterm's own default paste, since Pane B has no image-paste trap to route it to. ⚠️ `buildSplitPickerSessions()` also excludes any session with `pid === null` (an exited CLI, a crash-looped session whose breaker tripped, a restore that never re-attached): Pane B has no equivalent of `selectSession()`'s auto re-attach POST, so a pane opened onto one has nothing reading its tmux pane — no `terminal` events ever arrive, and `Session.write()` silently drops every keystroke with no ack either way, so the loss is invisible behind a socket that reports healthy. Pane B's input frames deliberately carry no `cid`/`seq` (`ws-routes.ts` supports that), matching the no-overlay/no-IME "deliberately plainer" list above, since it has no exactly-once delivery layer to key them against. ⚠️ A SHELL Pane B pulls scrollback itself when the wheel goes up at the top of its buffer (`_maybeLoadMoreHistory`/`_pullHistory`): tmux repaints a burst of output instead of scrolling it, so Pane B's own xterm holds about one screen of scrollback while tmux holds every line, and it loaded history exactly once at connect and never again. It is the same bounded pull as Pane A's (`?full=1&tail=TERMINAL_TAIL_SIZE`, no rewrite when the window holds no more rows than the pane already has or the pane is at its `scrollback + rows` cap, and a 60 s back-off instead of 4 s when that skipped window was truncated or the pane is full, since each ask costs the server a whole-history `capture-pane`), against Pane B's OWN terminal rather than `app.terminal`, so it cannot share `_maybeRefetchFullHistory`. The wheel listener is capture-phase because xterm `stopPropagation()`s the events it consumes; the alternate-screen skip (nano, vim, less) only matters for a direct-PTY shell, since under tmux the browser xterm never enters the alternate buffer; skipped too for a detached session (mirrors `_sendResize()`'s own check and app.js's `_maybeRefetchFullHistory`), since its own window already owns its PTY size and scrollback. Live frames arriving from the response onward, a `{t:'c'}` clear frame included, are held with their arrival time (`_liveQueue`, opened right after `await fetch(...)` beside `capturedAt`; a frame from before it is replaced by the capture or written unchanged, so the pane keeps painting during the round trip) and replayed in order only if they arrived after the capture (the response's arrival stands in for the capture instant, as in `_finishBufferLoad`, so a frame inside that one round trip can be lost or doubled); the request uses the primary pane's budget (`CodemanFetchDeadline.terminalFetchDeadlineMs({ full: true })`, 45 s, 10 s only if that helper is absent) and the body read gets 10 s once the headers land, because from then on the pull holds the pane's live output. The request phase holds no live output, but it does hold the single-flight flag, so a coalesced `{t:'r'}` refresh and a marker owed by a close (below) wait for the response, at worst for that whole budget (accepted: a Codeman restart resets an in-flight request along with the socket, so that pull fails at once and stamps the marker). ⚠️ The "Pane B disconnected" marker must be the LAST thing on screen. A replay's own `\x1bc` would otherwise wipe a marker written before the pull and paint a fresh, current-looking history while `onData` keeps silently dropping every keystroke on the dead socket (a Codeman restart drops the socket while the tmux session, and so the HTTP pull, still succeeds), so `_pullHistory()` re-stamps it after the live-frame flush. A close DURING any load (a pull, a refresh; `connect()` awaits the initial load before it creates the socket, so no close lands in that one) writes nothing: `_onSocketClosed()` sets `_markerOwed` while `_bufferLoading` is true, since written there it would sit above the held frames the pull flushes after a skip, a downgrade, a failed fetch or the deadline, above a refresh's replay, or land mid-way through a chunked replay. Each load settles the marker in its OWN `finally` (`_stampMarkerIfOwed()`), after the queue flush, EXCEPT when a trailing refresh is pending: that refresh's `clear()` is synchronous while xterm parses a `write()` on a later tick, so a marker stamped just before it lands in the freshly cleared buffer ABOVE the refresh's replay (a second, stale copy; the default fake terminal in `test/split-pane-terminal-unit.test.ts` writes synchronously and cannot show it, so an async-parse fake there pins it). The marker stays owed instead, and the trailing refresh, which re-owes it on a closed socket anyway, writes the one copy below its own replay. Anything that wipes the terminal on a closed socket (a replay's `\x1bc`, a refresh's `clear()`) sets `_markerOwed` too, so the marker is rewritten whether or not the close landed during the load. Tracked via `_wsClosed`/`_markerOwed` rather than routed through `_onLiveOutput()`, since a close landing before the response is stamped before the cutoff and would be dropped with the rest of the pre-capture queue. There is no "Load full history" banner in Pane B, so a shell history past that 1 MiB window stays out of reach there. Non-shell Pane B is unchanged: it already loads `full=1`, and its history is out of scope for this pull (codex and Claude's inline renderer do grow tmux history; this just isn't how they recover it). Design: `docs/split-pane-sessions-plan.md`.
+**Split-pane sessions** (`showSplitButton`, header button, default OFF, per-device like `showFileViewerButton` — not in `SettingsUpdateSchema`, `displayKeys` in settings-ui.js): shows two live sessions side-by-side in one Codeman window. Pane A is the untouched, existing singleton terminal (`this.terminal`/`this._ws` in terminal-ui.js); Pane B is a new, independent `SplitTerminalPane` (terminal-split.js) with its own xterm instance and its own `/ws/sessions/:id/terminal` WebSocket. ⚠️ **Pane B is deliberately plainer than Pane A** — no local-echo overlay, no CJK IME, no touch/mobile handlers, no keyboard accessory bar — since this is a desktop-only feature (a split view needs a wide viewport) and those features exist for mobile/touch input; `.btn-split` is hard-hidden below 1180px regardless of the setting by the `@media (max-width: 1179px)` rule in styles.css (mobile.css only carries a comment pointing at it: that file loads up to 1023px, so it cannot cover the 1024-1179px tablet range the feature also needs to stay off), and the per-device setting means turning it on at a desk can never sync it onto a phone in the first place. No persistence: closing the browser tab or reloading always returns to the normal single-pane view; there is no localStorage key for split state. ⚠️ Splitting a session against itself is disallowed (the picker excludes the active session), **as is splitting against a popped-out (detached) session** — `buildSplitPickerSessions()` excludes `detachedSessions` because a detached session's own window is already claiming its PTY size, and `MAX_WS_PER_SESSION` needs no change since Pane A/B are always two different sessions. ⚠️ Either pane's session ending (deleted locally or from another client) auto-collapses the split — Pane A's session ending promotes Pane B to the new single pane via `selectSession(id, { auto: true })` (an app-driven selection, so it must not spend the session's idle alert — see the Approvals Inbox note above), never by trying to hot-swap the lightweight `SplitTerminalPane` object into the primary singleton state. ⚠️ Pane B refits on every window/sidebar/tab-rail resize via the SAME trailing-edge `ResizeObserver` callback that resizes Pane A (`throttledResize` in terminal-ui.js) — it only ever measured Pane A's own container, so without an explicit `this._splitPane?.fit()` call there Pane B silently kept its stale PTY size through every resize that did not happen to be a divider drag. ⚠️ A dropped WebSocket leaves Pane B visibly dead (a message written into its own xterm buffer) rather than silently swallowing keystrokes with nothing on screen to explain why — there is no reconnect logic for v1, matching the "deliberately plainer than Pane A" design. Related but distinct: `detachSession()` already opens one session in a separate OS-level browser window (`isSoloWindow`) — that is prior art for "two sessions visible at once" but not for one window with a draggable in-page divider, which is what this feature adds. ⚠️ Pane B installs its own `attachCustomKeyEventHandler` gating the same app-level chords Pane A's own handler gates (command palette, Alt+1-9/[/] tab nav, Alt+B sidebar toggle, Ctrl+Z suspend, Shift/Ctrl+Enter newline, smart-copy Ctrl+C) — without it the document capture-phase handler's `preventDefault()` (which never stops xterm) let each chord ALSO write its raw byte/escape sequence into Pane B's live PTY on top of whatever the app action did to Pane A (COD-153). Ctrl+Z is swallowed unless Pane B's own session is `mode === 'shell'`, mirroring terminal-ui.js's reasoning: in a plain shell it is the user's own job-control tool, everywhere else it silently suspends an unattended agent loop. Shift/Ctrl+Enter POSTs to `/api/sessions/:id/send-key` (`{key:'S-Enter'|'C-Enter'}`, typed by tmux `send-keys -H`: Ctrl+Enter is always a real 0x0a, Shift+Enter is the CLI's declared `capabilities.newline` chord, also 0x0a unless the CLI declares another; sent on keydown only, with the keypress and keyup swallowed too, see the keypress rule under Command palette and shortcut registry) targeting THIS pane's own `sessionId` rather than the primary pane's `activeSessionId` — without it xterm's plain `\r` would submit an incomplete prompt instead of adding a line to it. Smart-copy Ctrl+C/Ctrl+Shift+C is re-implemented against `this.terminal` (Pane B's own) rather than reusing `app.copyTerminalSelection()`, which reads Pane A's terminal and would copy the wrong pane's selection; Ctrl+Shift+C never falls through even with nothing to copy, mirroring terminal-ui.js's own `ev.shiftKey` branch. ⚠️ **This is a UX-parity fix, not an interrupt-safety one** — verified live in a real browser: xterm's `evaluateKeyboardEvent` routes a shifted ctrl-letter into a branch that assigns `c.key` only for two special cases (`_`→US, `@`→NUL), so it emits no data for Ctrl+Shift+C at all regardless of any application gate; a synthetic keydown with the gate removed produces zero WS frames, proving no accidental interrupt reaches the PTY either way. What gating the whole copy block on `hasSelection()` (an earlier draft) actually cost: with no selection, a selection-less Ctrl+Shift+C fell straight to `return true`, silently ceding the keystroke to the BROWSER's own handling (e.g. Chrome's Inspect-Element binding) with no feedback and no copy attempt — Pane A always intercepts it. Ctrl+V stays on xterm's own default paste, since Pane B has no image-paste trap to route it to. ⚠️ `buildSplitPickerSessions()` also excludes any session with `pid === null` (an exited CLI, a crash-looped session whose breaker tripped, a restore that never re-attached): Pane B has no equivalent of `selectSession()`'s auto re-attach POST, so a pane opened onto one has nothing reading its tmux pane — no `terminal` events ever arrive, and `Session.write()` silently drops every keystroke with no ack either way, so the loss is invisible behind a socket that reports healthy. Pane B's input frames deliberately carry no `cid`/`seq` (`ws-routes.ts` supports that), matching the no-overlay/no-IME "deliberately plainer" list above, since it has no exactly-once delivery layer to key them against. ⚠️ A SHELL Pane B pulls scrollback itself when the wheel goes up at the top of its buffer (`_maybeLoadMoreHistory`/`_pullHistory`): tmux repaints a burst of output instead of scrolling it, so Pane B's own xterm holds about one screen of scrollback while tmux holds every line, and it loaded history exactly once at connect and never again. It is the same bounded pull as Pane A's (`?full=1&tail=TERMINAL_TAIL_SIZE`, no rewrite when the window holds no more rows than the pane already has or the pane is at its `scrollback + rows` cap, and a 60 s back-off instead of 4 s when that skipped window was truncated or the pane is full, since each ask costs the server a whole-history `capture-pane`), against Pane B's OWN terminal rather than `app.terminal`, so it cannot share `_maybeRefetchFullHistory`. The wheel listener is capture-phase because xterm `stopPropagation()`s the events it consumes; the alternate-screen skip (nano, vim, less) only matters for a direct-PTY shell, since under tmux the browser xterm never enters the alternate buffer; skipped too for a detached session (mirrors `_sendResize()`'s own check and app.js's `_maybeRefetchFullHistory`), since its own window already owns its PTY size and scrollback. Live frames arriving from the response onward, a `{t:'c'}` clear frame included, are held with their arrival time (`_liveQueue`, opened right after `await fetch(...)` beside `capturedAt`; a frame from before it is replaced by the capture or written unchanged, so the pane keeps painting during the round trip) and replayed in order only if they arrived after the capture (the response's arrival stands in for the capture instant, as in `_finishBufferLoad`, so a frame inside that one round trip can be lost or doubled); the request uses the primary pane's budget (`CodemanFetchDeadline.terminalFetchDeadlineMs({ full: true })`, 45 s, 10 s only if that helper is absent) and the body read gets 10 s once the headers land, because from then on the pull holds the pane's live output. The request phase holds no live output, but it does hold the single-flight flag, so a coalesced `{t:'r'}` refresh and a marker owed by a close (below) wait for the response, at worst for that whole budget (accepted: a Codeman restart resets an in-flight request along with the socket, so that pull fails at once and stamps the marker). ⚠️ The "Pane B disconnected" marker must be the LAST thing on screen. A replay's own `\x1bc` would otherwise wipe a marker written before the pull and paint a fresh, current-looking history while `onData` keeps silently dropping every keystroke on the dead socket (a Codeman restart drops the socket while the tmux session, and so the HTTP pull, still succeeds), so `_pullHistory()` re-stamps it after the live-frame flush. A close DURING any load (a pull, a refresh; `connect()` awaits the initial load before it creates the socket, so no close lands in that one) writes nothing: `_onSocketClosed()` sets `_markerOwed` while `_bufferLoading` is true, since written there it would sit above the held frames the pull flushes after a skip, a downgrade, a failed fetch or the deadline, above a refresh's replay, or land mid-way through a chunked replay. Each load settles the marker in its OWN `finally` (`_stampMarkerIfOwed()`), after the queue flush, EXCEPT when a trailing refresh is pending: that refresh's `clear()` is synchronous while xterm parses a `write()` on a later tick, so a marker stamped just before it lands in the freshly cleared buffer ABOVE the refresh's replay (a second, stale copy; the default fake terminal in `test/split-pane-terminal-unit.test.ts` writes synchronously and cannot show it, so an async-parse fake there pins it). The marker stays owed instead, and the trailing refresh, which re-owes it on a closed socket anyway, writes the one copy below its own replay. Anything that wipes the terminal on a closed socket (a replay's `\x1bc`, a refresh's `clear()`) sets `_markerOwed` too, so the marker is rewritten whether or not the close landed during the load. Tracked via `_wsClosed`/`_markerOwed` rather than routed through `_onLiveOutput()`, since a close landing before the response is stamped before the cutoff and would be dropped with the rest of the pre-capture queue. There is no "Load full history" banner in Pane B, so a shell history past that 1 MiB window stays out of reach there. Non-shell Pane B is unchanged: it already loads `full=1`, and its history is out of scope for this pull (codex and Claude's inline renderer do grow tmux history; this just isn't how they recover it). Design: `docs/split-pane-sessions-plan.md`.
 
 ### Gesture control: the setting
 
diff --git a/docs/wiki/Settings-Reference.md b/docs/wiki/Settings-Reference.md
index 8d761abb..4e26616a 100644
--- a/docs/wiki/Settings-Reference.md
+++ b/docs/wiki/Settings-Reference.md
@@ -50,6 +50,7 @@ supervised by systemd or launchd; npm installs report as non-updatable. See
 | Normal / Bold font weight     | xterm defaults       | Per device, each slot from 100 to 900. The bundled JetBrains Mono renders every step, so a lighter normal weight makes Claude's bold headings stand out. Applies live to the terminal, both echo overlays and open team panes. |
 | WebGL Renderer                | On                   | With a GPU-stall watchdog that falls back to DOM rendering.            |
 | Gesture Control               | Off                  | Camera hand tracking. Also needs `CODEMAN_GESTURE=1` on the server.    |
+| Key tester                    | n/a                  | A diagnostic that stores nothing. Click the box and press keys to see what this browser reports (key, code, modifiers) for keydown, keypress and keyup, for when a chord such as Shift+Enter behaves differently on one device. Keys pressed there reach no session and trigger no shortcut. |
 
 ### Header & Panels
 
diff --git a/test/key-tester.browser.test.ts b/test/key-tester.browser.test.ts
index 5b1fcc3a..edc6eb5e 100644
--- a/test/key-tester.browser.test.ts
+++ b/test/key-tester.browser.test.ts
@@ -89,8 +89,29 @@ describe('Key tester in a real browser', () => {
   });
 
   it('keeps only the last 14 lines and never types into the field', async () => {
-    for (let i = 0; i < 8; i++) await page.keyboard.press('a');
-    expect((await log()).split('\n').length).toBeLessThanOrEqual(14);
+    // The test above blurred the field, so focus it again: without this the presses land on
+    // <body>, the log keeps whatever the earlier tests left, and the cap is never exercised.
+    await page.focus('#keyTesterInput');
+    expect(await page.evaluate(() => document.activeElement?.id)).toBe('keyTesterInput');
+    await page.evaluate(() => (document.getElementById('keyTesterLog')!.textContent = ''));
+    const lines = async () => (await log()).split('\n');
+
+    // A printable key fires keydown, keypress and keyup, so 2 presses are 6 lines: under the cap
+    // the log accumulates rather than showing only the latest event.
+    for (let i = 0; i < 2; i++) await page.keyboard.press('a');
+    expect(await lines()).toHaveLength(6);
+
+    // A different key, so eviction is visible: 4 x 3 = 12 more lines makes 18, capped to 14. The
+    // oldest 4 go (all of the first 'a' press and the second one's keydown), the newest stay in order.
+    for (let i = 0; i < 4; i++) await page.keyboard.press('b');
+    const capped = await lines();
+    expect(capped).toHaveLength(14);
+    expect(capped.filter((l) => l.includes('key="a"'))).toHaveLength(2);
+    expect(capped.filter((l) => l.includes('key="b"'))).toHaveLength(12);
+    expect(capped[0]).toMatch(/^keypress\s+key="a" code=KeyA mods=none charCode=97$/);
+    expect(capped[13]).toMatch(/^keyup\s+key="b" code=KeyB mods=none$/);
+
+    // Readonly: none of those presses typed anything into the field itself.
     expect(await page.inputValue('#keyTesterInput')).toBe('');
   });
 });
diff --git a/test/shift-enter-keypress.browser.test.ts b/test/shift-enter-keypress.browser.test.ts
index 77b5bd35..c13aead8 100644
--- a/test/shift-enter-keypress.browser.test.ts
+++ b/test/shift-enter-keypress.browser.test.ts
@@ -1,81 +1,217 @@
-// @vitest-environment node
-// Real xterm.js in real Chromium, real keystrokes. xterm runs the custom key handler for
-// keydown AND keypress, and drops a keypress that carries Ctrl/Alt but NOT a Shift-only one,
-// so a handler that returns false for keydown alone lets Shift+Enter fall through to a bare
-// \r (submit). That is why Ctrl+Enter and Alt+Enter inserted a newline while Shift+Enter
-// submitted. Self-contained: no server, only the xterm bundle.
+/**
+ * @fileoverview Shift/Ctrl+Enter against the SHIPPED key handlers, with real keystrokes in Chromium.
+ *
+ * xterm runs the custom key handler for keydown, keypress AND keyup, and drops a keypress that
+ * carries Ctrl/Alt but NOT a Shift-only one. A handler that returns false for keydown alone
+ * therefore lets Shift+Enter's keypress fall through to onData as a bare \r, which went out over
+ * the WebSocket ahead of the async send-key fetch and submitted the prompt (#520). Ctrl+Enter
+ * worked all along because Chromium fires no keypress for it.
+ *
+ * The page is the real app served by a real WebServer, so the handlers under test are the ones
+ * terminal-ui.js (the main pane, `app.terminal`) and terminal-split.js (Pane B, a real
+ * `SplitTerminalPane`) attach. Nothing restates their predicate. What stands in for the server is
+ * only the edge: a fetch wrapper records the send-key POSTs instead of letting them reach tmux, and
+ * no session exists behind the ids, so nothing is ever typed into a real pane.
+ *
+ * Browser-driven, so it is excluded from `npm run test:ci` like the other Playwright suites:
+ *   npm run test:browser -- test/shift-enter-keypress.browser.test.ts
+ * The CI gate's guard on the same handlers is the source check in
+ * test/shift-enter-keypress-swallowed.test.ts.
+ *
+ * Port: 3267 (per CLAUDE.md, ports 3150+ for tests)
+ */
 
-import { readFileSync } from 'node:fs';
-import { join } from 'node:path';
 import { afterAll, beforeAll, describe, expect, it } from 'vitest';
 import { chromium, type Browser, type Page } from 'playwright';
+import { WebServer } from '../src/web/server.js';
 
-const XTERM = readFileSync(join(process.cwd(), 'node_modules/@xterm/xterm/lib/xterm.js'), 'utf8');
+const PORT = 3267;
+const BASE_URL = `http://localhost:${PORT}`;
+const MAIN_ID = 'shift-enter-probe-main';
+const PANE_B_ID = 'shift-enter-probe-pane-b';
 
-declare global {
-  interface Window {
-    __t: Record<string, { term: { focus(): void }; sent: string[] }>;
-    Terminal: new () => {
-      open(el: HTMLElement): void;
-      focus(): void;
-      onData(cb: (d: string) => void): void;
-      attachCustomKeyEventHandler(cb: (ev: KeyboardEvent) => boolean): void;
-    };
-  }
+interface Pressed {
+  /** What xterm emitted through onData, i.e. what would have been written to the PTY. */
+  data: string[];
+  /** The send-key POSTs the handler made, as `<session id> <key>`. */
+  sendKeys: string[];
+  /** Whether xterm's own textarea held focus, so a key that emits nothing was really delivered. */
+  focused: boolean;
 }
 
-describe('Shift/Ctrl+Enter reach the PTY as a bare \\r only if the handler lets keypress through', () => {
+describe('Shift/Ctrl+Enter through the shipped key handlers (keypress must be swallowed too)', () => {
+  let server: WebServer;
   let browser: Browser;
   let page: Page;
 
   beforeAll(async () => {
+    server = new WebServer(PORT, false, true);
+    await server.start();
     browser = await chromium.launch({ headless: true });
     page = await browser.newPage();
-    await page.setContent('<body></body>');
-    await page.addScriptTag({ content: XTERM });
-    await page.evaluate(() => {
-      const mk = (handler: (ev: KeyboardEvent) => boolean) => {
-        const host = document.createElement('div');
-        document.body.appendChild(host);
-        const term = new window.Terminal();
-        term.open(host);
-        const sent: string[] = [];
-        term.onData((d) => sent.push(d));
-        term.attachCustomKeyEventHandler(handler);
-        return { term, sent };
-      };
-      // The shipped predicate (terminal-ui.js / terminal-split.js) and the one it replaced.
-      const shipped = (ev: KeyboardEvent) => !(ev.key === 'Enter' && (ev.shiftKey || ev.ctrlKey));
-      const keydownOnly = (ev: KeyboardEvent) =>
-        !(ev.key === 'Enter' && (ev.shiftKey || ev.ctrlKey) && ev.type === 'keydown');
-      window.__t = { shipped: mk(shipped), keydownOnly: mk(keydownOnly) };
-    });
-  });
-  afterAll(async () => {
-    await browser?.close();
-  });
+    await page.goto(BASE_URL, { waitUntil: 'domcontentloaded' });
+    await page.waitForFunction(() => (window as any).app?.terminal, null, { timeout: 30000 });
+    await page.evaluate(
+      ({ paneBId }) => {
+        const w = window as any;
+        // Record send-key instead of reaching the server (no tmux pane exists behind these ids).
+        w.__sendKeys = [] as string[];
+        const realFetch = window.fetch.bind(window);
+        window.fetch = (input: RequestInfo | URL, init?: RequestInit) => {
+          const url = typeof input === 'string' ? input : input instanceof URL ? input.href : input.url;
+          const m = /\/api\/sessions\/([^/]+)\/send-key$/.exec(url);
+          if (m) {
+            w.__sendKeys.push(`${m[1]} ${JSON.parse(String(init?.body ?? '{}')).key}`);
+            return Promise.resolve(
+              new Response('{}', { status: 200, headers: { 'Content-Type': 'application/json' } })
+            );
+          }
+          return realFetch(input, init);
+        };
 
-  async function typed(which: 'shipped' | 'keydownOnly', key: string): Promise<string[]> {
-    await page.evaluate((w) => {
-      window.__t[w].sent.length = 0;
-      window.__t[w].term.focus();
-    }, which);
+        // Pane B, built by the real class. connect() installs the key handler synchronously,
+        // before its first await; the buffer fetch and WebSocket for the missing session just fail.
+        const mount = document.createElement('div');
+        mount.style.cssText = 'position:fixed;left:0;top:0;width:400px;height:300px;';
+        document.body.appendChild(mount);
+        const pane = new w.SplitTerminalPane(paneBId, mount, { mode: 'claude' });
+        void pane.connect().catch(() => {});
+        // What xterm emits here is exactly what Pane B's own onData forwards to its WebSocket.
+        w.__paneBData = [] as string[];
+        pane.terminal.onData((d: string) => w.__paneBData.push(d));
+        w.__paneB = pane;
+
+        // The bare xterm the old keydown-only gate is reproduced on (see the first test).
+        const host = document.createElement('div');
+        host.style.cssText = 'position:fixed;left:420px;top:0;width:400px;height:300px;';
+        document.body.appendChild(host);
+        const old = new w.Terminal();
+        old.open(host);
+        w.__oldData = [] as string[];
+        old.onData((d: string) => w.__oldData.push(d));
+        old.attachCustomKeyEventHandler(
+          (ev: KeyboardEvent) => !(ev.key === 'Enter' && (ev.shiftKey || ev.ctrlKey) && ev.type === 'keydown')
+        );
+        w.__oldTerm = old;
+      },
+      { paneBId: PANE_B_ID }
+    );
+  }, 90000);
+
+  afterAll(async () => {
+    if (browser) await browser.close();
+    if (server) await server.stop();
+  }, 60000);
+
+  /**
+   * Press `key` in the main pane. `sessionId` null is the welcome screen; otherwise the app thinks
+   * that session is active, local echo is off (the desktop default) and its real send path,
+   * `_sendInputAsync`, is recorded instead of posting.
+   */
+  async function pressMain(key: string, sessionId: string | null): Promise<Pressed> {
+    await page.evaluate((id) => {
+      const w = window as any;
+      const app = w.app;
+      w.__mainSaved ??= { id: app.activeSessionId, echo: app._localEchoEnabled, send: app._sendInputAsync };
+      w.__mainData = [] as string[];
+      w.__sendKeys.length = 0;
+      app.activeSessionId = id;
+      app._localEchoEnabled = false;
+      app._pendingInput = '';
+      app._sendInputAsync = (_sid: string, chunk: string) => void w.__mainData.push(chunk);
+      app.terminal.focus();
+    }, sessionId);
+    const focused = await page.evaluate(() => document.activeElement === (window as any).app.terminal.textarea);
     await page.keyboard.press(key);
-    return page.evaluate((w) => [...window.__t[w].sent], which);
+    // The main pane batches onData through a short flush timer before _sendInputAsync.
+    await page.waitForTimeout(150);
+    return page.evaluate((f) => {
+      const w = window as any;
+      const app = w.app;
+      const out = { data: [...w.__mainData], sendKeys: [...w.__sendKeys], focused: f };
+      app.activeSessionId = w.__mainSaved.id;
+      app._localEchoEnabled = w.__mainSaved.echo;
+      app._sendInputAsync = w.__mainSaved.send;
+      app._pendingInput = '';
+      return out;
+    }, focused);
   }
 
-  it('reproduces the bug with the old keydown-only handler', async () => {
-    expect(await typed('keydownOnly', 'Shift+Enter')).toEqual(['\r']);
-    expect(await typed('keydownOnly', 'Control+Enter')).toEqual([]);
+  async function pressIn(which: 'paneB' | 'old', key: string): Promise<Pressed> {
+    await page.evaluate((k) => {
+      const w = window as any;
+      w[k === 'paneB' ? '__paneBData' : '__oldData'].length = 0;
+      w.__sendKeys.length = 0;
+      (k === 'paneB' ? w.__paneB.terminal : w.__oldTerm).focus();
+    }, which);
+    const focused = await page.evaluate((k) => {
+      const w = window as any;
+      return document.activeElement === (k === 'paneB' ? w.__paneB.terminal : w.__oldTerm).textarea;
+    }, which);
+    await page.keyboard.press(key);
+    await page.waitForTimeout(50);
+    return page.evaluate(
+      ({ k, f }) => {
+        const w = window as any;
+        return { data: [...w[k === 'paneB' ? '__paneBData' : '__oldData']], sendKeys: [...w.__sendKeys], focused: f };
+      },
+      { k: which, f: focused }
+    );
+  }
+
+  it('reproduces the leak on a bare xterm with the old keydown-only gate (the behaviour the fix depends on)', async () => {
+    // Not shipped code: this pins the xterm behaviour, so a future xterm that stops emitting the
+    // Shift-only keypress shows up here instead of silently turning the tests below vacuous.
+    const shift = await pressIn('old', 'Shift+Enter');
+    expect(shift.focused).toBe(true);
+    expect(shift.data).toEqual(['\r']);
+    expect((await pressIn('old', 'Control+Enter')).data).toEqual([]);
   });
 
-  it('sends nothing for Shift+Enter and Ctrl+Enter with the shipped handler (the send-key route supplies the newline)', async () => {
-    expect(await typed('shipped', 'Shift+Enter')).toEqual([]);
-    expect(await typed('shipped', 'Control+Enter')).toEqual([]);
+  it('main pane (terminal-ui.js): Shift+Enter and Ctrl+Enter write nothing and POST send-key exactly once', async () => {
+    const shift = await pressMain('Shift+Enter', MAIN_ID);
+    expect(shift.focused).toBe(true);
+    expect(shift.data).toEqual([]);
+    // Once, from the keydown: the swallowed keypress and keyup must not send again.
+    expect(shift.sendKeys).toEqual([`${MAIN_ID} S-Enter`]);
+
+    const ctrl = await pressMain('Control+Enter', MAIN_ID);
+    expect(ctrl.data).toEqual([]);
+    expect(ctrl.sendKeys).toEqual([`${MAIN_ID} C-Enter`]);
   });
 
-  it('leaves plain Enter and Alt+Enter alone', async () => {
-    expect(await typed('shipped', 'Enter')).toEqual(['\r']);
-    expect(await typed('shipped', 'Alt+Enter')).toEqual(['\x1b\r']);
+  it('main pane: plain Enter still reaches the send path as \\r, and Alt+Enter as ESC CR', async () => {
+    const enter = await pressMain('Enter', MAIN_ID);
+    expect(enter.focused).toBe(true);
+    expect(enter.data).toEqual(['\r']);
+    expect(enter.sendKeys).toEqual([]);
+
+    const alt = await pressMain('Alt+Enter', MAIN_ID);
+    expect(alt.data).toEqual(['\x1b\r']);
+    expect(alt.sendKeys).toEqual([]);
+  });
+
+  it('main pane on the welcome screen: Shift+Enter sends nothing and posts no send-key', async () => {
+    // With no active session the app's onData drops input anyway, so this pins the send-key
+    // guard (`this.activeSessionId`), not the keypress swallow; the tests above pin that.
+    const shift = await pressMain('Shift+Enter', null);
+    expect(shift.focused).toBe(true);
+    expect(shift.data).toEqual([]);
+    expect(shift.sendKeys).toEqual([]);
+  });
+
+  it("Pane B (terminal-split.js): Shift+Enter and Ctrl+Enter write nothing and POST send-key once for Pane B's own session", async () => {
+    const shift = await pressIn('paneB', 'Shift+Enter');
+    expect(shift.focused).toBe(true);
+    expect(shift.data).toEqual([]);
+    expect(shift.sendKeys).toEqual([`${PANE_B_ID} S-Enter`]);
+
+    const ctrl = await pressIn('paneB', 'Control+Enter');
+    expect(ctrl.data).toEqual([]);
+    expect(ctrl.sendKeys).toEqual([`${PANE_B_ID} C-Enter`]);
+
+    const enter = await pressIn('paneB', 'Enter');
+    expect(enter.data).toEqual(['\r']);
+    expect(enter.sendKeys).toEqual([]);
   });
 });