docs: describe right-to-left rendering in the manual and the invariants (release review)

The wiki gains right-to-left notes for Last Response, the phone's Compose
editor and the rendered Markdown view, plus a Troubleshooting entry for the
#715 report: the terminal is xterm.js, which has no right-to-left support
(xtermjs/xterm.js#701). src/web/CLAUDE.md and the invariants now state the dir
pins, the logical-property rule and that the direction pass runs last.

The changeset claimed every file path stays left to right. It now says only
what holds: the paths the viewer links on its own.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

.changeset/rtl-rendered-markdown.md

@@ -2,4 +2,4 @@
 'aicodeman': patch
 ---
 
-Persian, Arabic and Hebrew now read right to left in the Response Viewer and in the File Viewer's markdown view (#715). Each paragraph, heading, list, quote and table takes its direction from its own first letter, so a right-to-left answer is right-aligned with its punctuation, list bullets and quote bar on the right, while an English paragraph beside it stays left to right. Code blocks, inline code and file paths stay left to right and no longer decide a sentence's direction, so a Persian sentence that starts with `npm install` still reads right to left. The phone's Compose dialog lays a right-to-left prompt out the same way. The terminal pane is unchanged: xterm.js has no right-to-left support yet (xtermjs/xterm.js#701).
+Persian, Arabic and Hebrew now read right to left in the Response Viewer and in the File Viewer's markdown view (#715). Each paragraph, heading, list, quote and table takes its direction from its own first letter, so a right-to-left answer is right-aligned with its punctuation, list bullets and quote bar on the right, while an English paragraph beside it stays left to right. Code blocks, inline code and the file paths the viewer links on its own stay left to right and no longer decide a sentence's direction, so a Persian sentence that starts with `npm install` still reads right to left. The phone's Compose dialog lays a right-to-left prompt out the same way. The terminal pane is unchanged: xterm.js has no right-to-left support yet (xtermjs/xterm.js#701).

docs/architecture-invariants.md

@@ -403,7 +403,7 @@ A file path an agent prints is a link on both surfaces it can appear on, and cli
 
 **One pattern, two consumers.** `FILE_PATH_LINK_PATTERN` / `absoluteFilePathPattern()` live in `constants.js`; the xterm link provider (`registerFilePathLinkProvider`, terminal-ui.js) and the response viewer's `_linkifyFilePaths()` (app.js) both reach it through `findFilePathLinks()` (below), which builds a fresh instance per call. ⚠️ Fresh per call, never one shared object: `lastIndex` is per-object state on a `/g` regex. The pattern is anchored on a known absolute root and terminated by a known extension, so a fraction (`3/4`) or a date can't match and trailing punctuation stays out. Roots include `Users` and `mnt`, without which nothing was clickable on macOS or WSL. The linear-time guard and the "terminal-ui builds from the factory" structural check are in `test/link-provider-regex.test.ts`.
 
-**The chat linkifier walks text nodes.** `_linkifyFilePaths()` builds anchors with `createElement`/`textContent` on the rendered subtree, never by rebuilding sanitized markup as a string — the source is model output. Subtrees already inside an `<a>` are skipped (marked autolinks URLs; a nested anchor would swallow the click), and the anchor's text is the path verbatim so "copy code" still yields what the agent printed. Before the walk it also rewrites existing anchors whose href is a scheme-less, non-`//` path (Codex cites files as `[label](/abs/path.json)`, which otherwise opens as a same-origin URL) into the same `a.rv-path` shape, plus `rv-path-link` so the label keeps the prose font, on purpose without `FILE_PATH_LINK_PATTERN`'s root and extension filter: the href is a link target the agent chose. `test/response-viewer-file-links.test.ts` pins both properties and the anchor rewrite.
+**The chat linkifier walks text nodes.** `_linkifyFilePaths()` builds anchors with `createElement`/`textContent` on the rendered subtree, never by rebuilding sanitized markup as a string — the source is model output. Subtrees already inside an `<a>` are skipped (marked autolinks URLs; a nested anchor would swallow the click), and the anchor's text is the path verbatim so "copy code" still yields what the agent printed. Before the walk it also rewrites existing anchors whose href is a scheme-less, non-`//` path (Codex cites files as `[label](/abs/path.json)`, which otherwise opens as a same-origin URL) into the same `a.rv-path` shape, plus `rv-path-link` so the label keeps the prose font, on purpose without `FILE_PATH_LINK_PATTERN`'s root and extension filter: the href is a link target the agent chose. The anchors the walk builds carry `dir="ltr"`, so a path at the head of a right-to-left sentence does not decide that sentence's direction; a rewritten link keeps no `dir`, since its label is prose (#715). ⚠️ `_applyBlockDirection()` runs after the linkifier in both consumers, never before it: a block already carrying `dir="auto"` re-resolves its direction on every text node replaced under it. `test/response-viewer-file-links.test.ts` pins both properties and the anchor rewrite, `test/response-viewer-rtl.test.ts` the direction and the order.
 
 **Out-of-workspace paths go through the attachment routes, not the file routes.** `file-content`/`file-raw` resolve against `workingDir` and 404 anything that escapes it, which is correct and unchanged — but the paths agents most often print (a `/tmp` capture, Claude's own scratchpad, another checkout) are exactly that, so clicking one used to report "File not found" for a file sitting on disk. `openFilePreview()` now detects the case (`_isExternalPreviewPath`, a string compare for ROUTING only; the real decision stays server-side) and registers the path via `POST /api/sessions/:id/attachments` first, rendering by id. ⚠️ That registration passes `notify: false`, which suppresses ONLY the `attachment:detected` broadcast — the guard, the registry entry and the by-id routes are identical either way. Without it every click also popped an attachment card announcing the file already filling the screen. ⚠️ The click is an explicit user action on the **explicit, Origin-guarded** registration route, which is why it may cross the workspace boundary at all; the passive magic-link scanner stays force-confined. A type outside `SUPPORTED_ATTACHMENT_EXTENSIONS` (`.svg`, `.bmp`) is refused with a message naming what IS previewable, rather than the registry's own policy term.
 

docs/wiki/Mobile-Guide.md

@@ -66,8 +66,9 @@ A row of keys above the virtual keyboard, and what it contains depends on the se
 **Agent sessions** get quick actions: `/init`, `/clear`, `/compact`, a Compose key, `Esc`,
 a path picker, and 🧠 when Read My Mind is on. Compose opens a multiline editor with
 autocorrect: Enter adds a new line, and only Send delivers the text, as one paste followed
-by Enter, so your line breaks reach the agent intact. Anything already typed on the terminal
-prompt moves into the editor when it opens. Drafts are kept per session and in memory only,
+by Enter, so your line breaks reach the agent intact. A prompt written in Persian, Arabic or
+Hebrew lays out right to left. Anything already typed on the terminal prompt moves into the
+editor when it opens. Drafts are kept per session and in memory only,
 so switching tabs keeps them and a page reload forgets them; a dot on the key shows a draft
 is parked. The editor's Attach button attaches photos, videos or documents and puts their paths
 into the draft.
@@ -187,6 +188,7 @@ Setup in [Notifications And Approvals](Notifications-And-Approvals).
 
 The terminal viewport is small. **Last Response** (opt-in header button) renders the agent's
 last answer as scrollable text instead, with a **More** button for additional context.
+Persian, Arabic and Hebrew answers read right to left there.
 
 The [File Viewer](Working-With-Files) works on phones too, including edit mode, which is
 enough to fix a typo an agent introduced while you are away from your desk.

docs/wiki/Troubleshooting.md

@@ -246,6 +246,20 @@ If you are sending input over the API instead, your payload must end with `\r` o
 is ever sent. The request still succeeds and the text sits unsubmitted in the composer. See
 [Driving Codeman From An Agent](Driving-Codeman-From-An-Agent).
 
+### Persian, Arabic or Hebrew looks wrong in the terminal
+
+The terminal is xterm.js, which does not support right-to-left text yet
+([xtermjs/xterm.js#701](https://github.com/xtermjs/xterm.js/issues/701)). It puts each
+character in its own cell from left to right, so a sentence comes out reversed, and the
+default WebGL renderer also draws the letters unjoined. Codeman cannot fix that without
+replacing the terminal.
+
+Read the answer in **Last Response** instead, where it lays out right to left, or open a
+Markdown file in the [File Viewer](Working-With-Files). On desktop, turning off **WebGL
+Renderer** in **App Settings → Terminal & Input** and reloading joins the letters, so each
+word reads correctly, but the words of a sentence still run left to right. Phones already
+use that renderer.
+
 ## Mobile
 
 ### The keyboard covers the terminal, or scroll position jumps

docs/wiki/Watching-Agents-Work.md

@@ -77,7 +77,9 @@ Turning on either starts the watcher.
 
 **Last Response** (header button, opt-in) renders the agent's last answer as scrollable text
 rather than terminal output. It exists mostly for phones, where reading a long answer in a
-terminal viewport is painful. **More** loads additional context.
+terminal viewport is painful. **More** loads additional context. Persian, Arabic and Hebrew
+answers read right to left here, which the terminal cannot do, while code and the file paths
+it links stay left to right.
 
 ## After the fact
 

docs/wiki/Working-With-Files.md

@@ -14,7 +14,7 @@ It renders what it can:
 | Kind                     | Behaviour                                                                 |
 | ------------------------ | ------------------------------------------------------------------------- |
 | Text and code            | Plain preview with Lines (line numbers) and Wrap toggles in the header. Long files are truncated in plain preview. |
-| Markdown                 | Rendered by default: headings, tables, code blocks with copy buttons, images and links relative to the file (root-relative ones resolve from the workspace root, as on GitHub). Links to another heading of the same file (`[Install](#installation)`) scroll to it, with GitHub's heading names (lower-case, punctuation dropped, repeats numbered `-1`, `-2`), and never leave the page. Opened from an attachment card, where the file's folder is unknown, relative images show their alt text and relative links show as plain text. The MD pill in the header flips to source. |
+| Markdown                 | Rendered by default: headings, tables, code blocks with copy buttons, images and links relative to the file (root-relative ones resolve from the workspace root, as on GitHub). Links to another heading of the same file (`[Install](#installation)`) scroll to it, with GitHub's heading names (lower-case, punctuation dropped, repeats numbered `-1`, `-2`), and never leave the page. Opened from an attachment card, where the file's folder is unknown, relative images show their alt text and relative links show as plain text. Persian, Arabic and Hebrew read right to left, each block following its own first letter, while code stays left to right. The MD pill in the header flips to source. |
 | Images                   | Inline.                                                                    |
 | Audio and video          | Inline with a working scrub bar, because range requests are supported.     |
 | Spreadsheets (`.xlsx`)   | Read-only grid, parsed in your browser (never on the server), up to 10 MB. `.xls` and `.ods` are download only. |

src/web/CLAUDE.md

@@ -22,7 +22,7 @@
 
 **Ctrl+V paste trap** (`image-input.js`): `Ctrl+V` routes through `_handleImagePaste()`, which focuses a hidden `contenteditable` trap and reads the clipboard from the paste event landing there; images upload and their paths are typed in, plain text is a Codeman-built bracketed paste handed to that pane's terminal via `terminal.input(framed)` (strip nested 200/201 markers, newlines to `\r`, wrap CSI 200~/201~) because `terminal.paste()` drops those markers after a replay clears xterm's DECSET 2004 mirror (#716). `input()` fires onData as user input, so the main pane's bracketed-paste branch and a tile's `_onTerminalData` deliver the frame, including the local-echo flush (#219). The terminal is only refocused. ⚠️ **The trap must consume exactly ONE paste event** (Firefox delivers two per keypress: the `execCommand('paste')` event and the keydown's default action); the one-shot flag lives on the trap, never on a browser check. ⚠️ Do not remove the `execCommand('paste')` call: on some mobile engines it is the only route into the trap, and the trap is the only place image blobs are read. Tests: `test/image-paste-trap.test.ts`. → [architecture-invariants#terminal-paste-ctrlv](../../docs/architecture-invariants.md#terminal-paste-ctrlv)
 
-**File Viewer text view: rendered markdown + Lines/Wrap toggles** (`_renderFilePreviewText()` in panels-ui.js): a `.md`/`.markdown` opens RENDERED by default with an `MD` pill back to source; the plain-text view has `Lines` (CSS-counter gutter) and `Wrap` toggles. ⚠️ ONE markdown pipeline: the viewer calls `_renderMarkdown(text, { breaks: false })` (marked + the DOMPurify allowlist, the Response Viewer's; chat keeps the default `breaks: true`, a file must not turn every hard wrap into a `<br>`) and binds the Response Viewer's click delegate (`_bindResponseViewerInteractions`) on the preview body for code-copy buttons and path links; never a second parser or handler. ⚠️ The document is built inside a `<template>` (a detached div with `innerHTML` set starts fetching every `<img src>` before the rewrite), then `_rebaseFilePreviewMarkdownRefs()` points relative images at the workspace-confined `file-raw` under the document's directory and root-relative ones under the workspace root (never a widened route; a failed load degrades to alt text), after `decodeURIComponent`ing the ref and dropping `?query`/`#fragment` (marked percent-encodes destinations, and the route encodes again), and turns workspace links into `a.rv-path` carrying `data-session-id` for the delegate (the linkifier's absolute paths get it too), stripping the `target` marked gave them. ⚠️ A preview opened by attachment id under a bare file name (attachment cards, history drawer: the registry keeps no relative path) has no directory, so its workspace refs degrade (images to alt text, links to their text), never resolve against the workspace root. ⚠️ The container carries `data-i18n-skip` or the translator rewrites the document's prose. ⚠️ Toggles are per-device localStorage keys (`codeman:filePreview*`), never `SettingsUpdateSchema`; Lines/Wrap are class flips on the ONE `<pre>`, with rules scoped `.file-preview-body > pre.file-preview-text` so they never leak into the document's code blocks. Markdown fetches `lines=10000` (the route ceiling), other text keeps 500. ⚠️ `md`/`markdown` are IN `FILE_PREVIEW_EXTENSIONS` (owner decision 2026-10-13): a markdown path clicked or tapped in the TERMINAL (main pane, grid tiles and split Pane B share `registerFilePathLinkProvider`) opens this rendered view like every other surface, never the streaming log viewer; other in-workspace text keeps the tail viewer (live follow). The terminal pattern links `.markdown` and `.htm` too. Tests: `test/file-preview-markdown.test.ts`. → [architecture-invariants#file-viewer-text-view-rendered-markdown-and-text-toggles](../../docs/architecture-invariants.md#file-viewer-text-view-rendered-markdown-and-text-toggles)
+**File Viewer text view: rendered markdown + Lines/Wrap toggles** (`_renderFilePreviewText()` in panels-ui.js): a `.md`/`.markdown` opens RENDERED by default with an `MD` pill back to source; the plain-text view has `Lines` (CSS-counter gutter) and `Wrap` toggles. ⚠️ ONE markdown pipeline: the viewer calls `_renderMarkdown(text, { breaks: false })` (marked + the DOMPurify allowlist, the Response Viewer's; chat keeps the default `breaks: true`, a file must not turn every hard wrap into a `<br>`) and binds the Response Viewer's click delegate (`_bindResponseViewerInteractions`) on the preview body for code-copy buttons and path links; never a second parser or handler. ⚠️ The document is built inside a `<template>` (a detached div with `innerHTML` set starts fetching every `<img src>` before the rewrite), then `_rebaseFilePreviewMarkdownRefs()` points relative images at the workspace-confined `file-raw` under the document's directory and root-relative ones under the workspace root (never a widened route; a failed load degrades to alt text), after `decodeURIComponent`ing the ref and dropping `?query`/`#fragment` (marked percent-encodes destinations, and the route encodes again), and turns workspace links into `a.rv-path` carrying `data-session-id` for the delegate (the linkifier's absolute paths get it too), stripping the `target` marked gave them. ⚠️ A preview opened by attachment id under a bare file name (attachment cards, history drawer: the registry keeps no relative path) has no directory, so its workspace refs degrade (images to alt text, links to their text), never resolve against the workspace root. ⚠️ The container carries `data-i18n-skip` or the translator rewrites the document's prose. ⚠️ Toggles are per-device localStorage keys (`codeman:filePreview*`), never `SettingsUpdateSchema`; Lines/Wrap are class flips on the ONE `<pre>`, with rules scoped `.file-preview-body > pre.file-preview-text` so they never leak into the document's code blocks. Markdown fetches `lines=10000` (the route ceiling), other text keeps 500. ⚠️ `md`/`markdown` are IN `FILE_PREVIEW_EXTENSIONS` (owner decision 2026-10-13): a markdown path clicked or tapped in the TERMINAL (main pane, grid tiles and split Pane B share `registerFilePathLinkProvider`) opens this rendered view like every other surface, never the streaming log viewer; other in-workspace text keeps the tail viewer (live follow). The terminal pattern links `.markdown` and `.htm` too. ⚠️ **Right-to-left text** (#715) has three parts on that one pipeline: `_renderMarkdown` pins `pre, code` to `dir="ltr"`, `_linkifyFilePaths` pins the anchors it builds the same way (`dir="auto"` skips a descendant that carries a `dir`, so `npm install` or a path at the head of a Persian sentence does not decide its direction; a markdown link's label is prose and stays unpinned), and `_applyBlockDirection()` gives each OUTERMOST prose block `dir="auto"`, keeping a `dir` the agent wrote and what it covers. ⚠️ That pass runs LAST, on the DOM, in both consumers (`_buildResponseViewerMessage`, `_renderFilePreviewText`), never in `_renderMarkdown` (its string must hold no `dir="auto"`) and never inside the linkifier (which returns early on a stale constants.js): an element already carrying `dir="auto"` re-resolves its direction on every insertion under it, so a code-only or path-only document went quadratic and a 200 KB list froze the tab for 40 s. ⚠️ The `.rv-text` block rules use logical properties (`padding-inline-start`, `border-inline-start`, `border-inline-end`, `text-align: start`, logical corner radii), never physical left and right, or an RTL block keeps its bullets and quote bar on the wrong side. The terminal pane is out of reach: xterm.js has no RTL support (xtermjs/xterm.js#701). Tests: `test/file-preview-markdown.test.ts`, `test/response-viewer-rtl.test.ts`. → [architecture-invariants#file-viewer-text-view-rendered-markdown-and-text-toggles](../../docs/architecture-invariants.md#file-viewer-text-view-rendered-markdown-and-text-toggles)
 
 **Spreadsheet preview (XLSX)** (`spreadsheet-preview.js` + `spreadsheet-preview-worker.js`; moved from File-path links in the root file): ⚠️ **XLSX previews parse in the BROWSER**, never on the server: `spreadsheet-preview.js` fetches the raw route with `?preview=true` (413 above `MAX_XLSX_BROWSER_PREVIEW_BYTES`, 10 MB) and hands the bytes to `spreadsheet-preview-worker.js`, the only place the pinned `exceljs`/`fflate` vendor bundles load (never on page load). `admitXlsx()` caps the ZIP before ExcelJS runs, counting what ExcelJS will EXPAND as well as what it reads (a merge costs its area, a `<col>` past 16384 is refused, validations are never parsed via `ignoreNodes`, and defined names are never expanded: the worker stubs `_definedNames.model`). ⚠️ The INDEX a row or sheet claims is bounded too, since ExcelJS allocates and walks up to it: a `<row r>` outside 1-1048576 and an `xl/workbook.xml` `<sheet sheetId>` above `LIMITS.maxSheetId` (65535) are refused, and `sendTile` reuses the merges read at load (`mergesById`), never `sheet.model`, which rebuilds the whole sheet. ⚠️ The COLUMN index costs the same way (a row's cells live at `_cells[col - 1]`, so one `XFD` cell makes ExcelJS's `eachRow`/`eachCell`/`hasValues` visit 16,384 slots per row): `worksheetMetadata` builds each sheet's row and cell index from the keys that exist (`Object.keys` of `_rows` and `_cells`, skipping falsy and `Null`-type cells as `eachCell({ includeEmpty: false })` does), reads row heights in that pass and merges from `sheet._merges`, and `sendTile` reads cells from that index (`populatedRowsById`); never call ExcelJS's dense `eachRow`/`eachCell` there. `parseThemePalette` returns the default palette for a theme above 64 KB (64 * 1024 characters), since its patterns are quadratic on unclosed tags. ⚠️ Merges are capped per sheet (`LIMITS.maxMergesPerSheet`, 2,000) AND workbook-wide (`LIMITS.maxMerges`, 10,000), refused as `merge-limit` before ExcelJS loads, since ExcelJS's `_mergeCellsInternal` checks each merge against every earlier one on its sheet (quadratic). Every `<numFmt>` in `xl/styles.xml` is read with `readTagAttributes()` and refused (`number-format`) when its decoded `formatCode` is over 255 characters or has a `[` after its last `]`: ExcelJS's `isDateFmt` rescans to the end of the code per unclosed `[` once per numeric cell, and the code is echoed into the notice bar. ⚠️ `formatCellValue` caps every cell's display text at `LIMITS.maxCellTextChars` (1,000, ellipsis, never splitting a surrogate pair) for every value shape (rich text, hyperlink text, formula source, errors), since structured clone copies each tile cell's whole string to the page and the load timeout no longer covers tiles. `richTextPrefix` also visits at most `maxCellTextChars + 1` runs, not only keeps that much text: an empty `<r/>` adds nothing, so a shared string of a million empty runs was walked whole per cell, per tile (56 s a tile). ⚠️ Every part ExcelJS parses (all but the pure-bytes `xl/media/<name>.<ext>`) is also charged per START TAG against `LIMITS.maxElements` (2,000,000, refused as `element-limit`): ExcelJS builds an object per element in sharedStrings, styles, comments, drawings, VML and tables too, and empty stored blocks pad a deflate stream past the ratio cap, so 8.3M empty `<r/>` runs in a 375 KB file cost 570 MB of heap. ⚠️ Admission keys every entry on the name ExcelJS will SEE (`excelJsEntryName()`: JSZip's `.`/`..`/empty-segment resolution, then one leading `/` stripped), refuses two entries that land on one name, and treats anything matching ExcelJS's UNANCHORED `xl/worksheets/sheet<N>.xml` as a worksheet, so `/xl/...` or `xl/./...` cannot skip a counter, and ExcelJS parses only a STORE-only archive rebuilt from the entries admission inflated (`buildAdmittedArchive()`), never the fetched bytes; cell text goes through `textContent`, formulas are never evaluated. Bumping either package or editing the worker/core changes `SPREADSHEET_ASSET_VERSION`, which `npm run check:public-assets` pins. xls/ods stay download-only.
 
@@ -72,7 +72,7 @@ Frontend JS modules have `@fileoverview` with `@dependency`/`@loadorder` tags. L
 
 **Shell keyboard accessory bar + one-shot Ctrl** (`keyboard-accessory.js`): a shell-mode session swaps the mobile accessory bar for terminal controls; `setMode()` records `extendedKeyboardBar` as the base layout and `refreshForActiveSession()` resolves base-vs-shell. ⚠️ Ctrl is a one-shot modifier applied in `terminal.onData` (after `shouldSuppressTerminalQueryResponse`, before every send path), and must skip `isTerminalFocusOrMouseReport()` chunks. ⚠️ It must disarm on use, second tap, any other accessory key, session switch, keyboard dismissal and layout swap. ⚠️ `_handleCjkInput()` must apply it too (the CJK textarea bypasses onData). Mapping: `ctrlByteFor()`. ⚠️ mobile.css's light-skin repaint must keep excluding `.accessory-btn:not(.armed)` or the armed state is invisible. → [architecture-invariants#shell-keyboard-accessory-bar-and-one-shot-ctrl](../../docs/architecture-invariants.md#shell-keyboard-accessory-bar-and-one-shot-ctrl)
 
-**Mobile prompt composer** (`keyboard-accessory.js`): the agent bars' Paste key is **Compose**, a native multiline dialog where only **Send** submits (the shell bar keeps plain Paste). Opening it adopts the whole terminal prompt (`_takePendingLocalEcho`), erasing the flushed prefix with backspaces counted in code points. ⚠️ Drafts are per-session and in memory only (`_composerDrafts`), never persisted (prompts carry secrets). ⚠️ Delivery is a hand-built bracketed-paste frame via `_sendInputAsync` WITHOUT `useMux`, then a separate delayed Enter WITH it; never `terminal.paste()`, and the frame must never take the mux fallback (`TmuxManager.sendInput()` strips every `\r`, the frame's line separator, so the lines weld into one inside a second pair of markers and get submitted). ⚠️ `_composerMaxLength` must stay derived from `MAX_INPUT_LENGTH` minus the markers, or an oversized frame wedges the durable queue. ⚠️ The composer overlay needs its own gutter restatement after the fold rules. Test: `test/mobile-prompt-composer.test.ts`. → [architecture-invariants#mobile-prompt-composer](../../docs/architecture-invariants.md#mobile-prompt-composer)
+**Mobile prompt composer** (`keyboard-accessory.js`): the agent bars' Paste key is **Compose**, a native multiline dialog where only **Send** submits (the shell bar keeps plain Paste). Opening it adopts the whole terminal prompt (`_takePendingLocalEcho`), erasing the flushed prefix with backspaces counted in code points. ⚠️ Drafts are per-session and in memory only (`_composerDrafts`), never persisted (prompts carry secrets). ⚠️ Delivery is a hand-built bracketed-paste frame via `_sendInputAsync` WITHOUT `useMux`, then a separate delayed Enter WITH it; never `terminal.paste()`, and the frame must never take the mux fallback (`TmuxManager.sendInput()` strips every `\r`, the frame's line separator, so the lines weld into one inside a second pair of markers and get submitted). ⚠️ `_composerMaxLength` must stay derived from `MAX_INPUT_LENGTH` minus the markers, or an oversized frame wedges the durable queue. ⚠️ The composer overlay needs its own gutter restatement after the fold rules. The textarea carries `dir="auto"`, so a prompt written in Persian, Arabic or Hebrew lays out right to left from its first letter (#715). Test: `test/mobile-prompt-composer.test.ts`. → [architecture-invariants#mobile-prompt-composer](../../docs/architecture-invariants.md#mobile-prompt-composer)
 
 **Terminal resilience**: a replay clear is the queued in-stream `\x1bc` in `_resetTerminalForReplay()`, never `reset()`/`clear()` (queued bytes fuse into the snapshot); the renderer watchdog `_kickRenderer()` reads xterm privates, pinned by `test/xterm-private-api.test.ts` against the resolved lockfile version; every terminal capture fetch has a deadline that covers the BODY (`_fetchTerminalCapture`). → [architecture-invariants#terminal-resilience-replay-clears-renderer-liveness-fetch-deadlines](../../docs/architecture-invariants.md#terminal-resilience-replay-clears-renderer-liveness-fetch-deadlines)