@aragon-agent/cli 0.6.8 → 0.6.10

package/CHANGELOG.md

@@ -7,6 +7,15 @@ were written.
 
 ## Unreleased
 
+### Removed
+
+- Removed the inline terminal UI and its automatic downgrade rules. Interactive
+  sessions now always use the full-screen UI; small windows show a resize notice.
+- Removed `--fullscreen`, `--no-fullscreen`, `ARAGON_FULLSCREEN`, and the
+  `fullscreen` config setting. Old flags are rejected and old stored preferences
+  are ignored. Headless commands and piped input retain their existing behavior.
+- Interactive-only screens now reject non-TTY input/output with exit code 2.
+
 ### Added
 
 - **`system/init` now lists what this build can do.** A new optional
@@ -72,6 +81,16 @@ were written.
 
 ### Fixed
 
+- **The right-edge scrollbar no longer disappears while output streams.** The
+  differential renderer ended every repainted row with `CSI K` (erase to end of
+  line). The scrollbar fills the last column of every viewport row, and on
+  terminals that keep the cursor on the last column after writing it (Windows
+  console hosts) that erase blanked the cell that had just been drawn, so the
+  scrollbar vanished whenever a row was redrawn. Rows that already fill the
+  terminal width are now repainted without the erase; narrower rows are
+  unchanged. `--no-diff-render` was never affected. `/perf` gained a `scrollbar:`
+  line for diagnosing visibility and drag problems.
+
 - **`{"type":"interrupt"}` now really does settle the run — on every platform.**
   The stdin interrupt frame used to raise a SIGINT at our own process. On POSIX
   that reaches the runner's terminator and the run winds down properly: it
@@ -113,6 +132,14 @@ were written.
 
 ### Changed
 
+- **The run status row moved above the input box and carries the `steer` hint.**
+  While a run is active, the spinner, the phrase and `steer / interrupt / exit`
+  share one row directly above the input, in place of the idle hint row below it
+  (the footer keeps its height). It scrolls with the input; when it leaves the
+  viewport the fixed bottom row shows the animation instead, so there is never
+  more than one spinner. Terminals under 20 rows and `--no-hints` keep the old
+  fixed row. The update notice is also held back for the whole run.
+
 - **Auto-compaction no longer reports compactions that did not happen.** A long
   unattended run used to print a red `context not compacted #2: nothing_to_drop`
   card into the transcript. The card was accurate; nobody had asked for it.

package/README.md

@@ -362,8 +362,6 @@ below works there too, in either position — `aragon --model m exec "…"` and
 | `--hints` / `--no-hints` | Show or hide the composer hint row |
 | `--show-thinking` / `--no-show-thinking` | Draw the model's reasoning blocks in the transcript. Off by default — see [What you see while it runs](#what-you-see-while-it-runs) |
 | `--live-tool-output` / `--no-live-tool-output` | Show a bounded tail of a running command's output on its card. **On** by default — see [What you see while it runs](#what-you-see-while-it-runs) |
-| `--fullscreen` | Force the full-screen TUI, overriding the automatic downgrades |
-| `--no-fullscreen` | Force the inline renderer (the 0.2.0 behavior) |
 | `--no-exit-transcript` | Do not replay the session summary after exiting |
 | `--plan` / `--no-plan` | Start the session in PLAN mode (read-only research + review) or force BUILD |
 | `--team` / `--no-team` | Enable or disable team subagents. `--no-team` does not register the `task` tool at all |
@@ -434,17 +432,17 @@ completes, `Up` / `Down` moves the selection, `Esc` closes it.
 
 | Key | Action |
 | --- | --- |
-| `Enter` | Submit (idle) / queue a steering message (running) |
-| `Alt+Enter` / `Shift+Enter` | Insert a newline |
+| `Enter` | Submit (idle) / queue a steering message (running) — shown as `Queue: …` until the agent picks it up |
+| `Ctrl+J` / `Alt+Enter` / `Shift+Enter` | Insert a newline. `Ctrl+J` and `Alt+Enter` work everywhere; a distinct `Shift+Enter` needs a one-line terminal binding — see [`/terminal-setup`](#terminal-setup) |
 | `Shift+Tab` | Toggle **plan mode** (`BUILD` <-> `PLAN`). Equivalent: `/plan`. On Windows this needs a console flag Node < 22.17 does not set; aragon sets it at startup — see [Requirements](#requirements) |
 | `Ctrl+P` | The same toggle, over a channel no Windows console can swallow. Use it if `Shift+Tab` does nothing |
 | `Esc` | Close an overlay or popup; otherwise ask for interruption confirmation. While idle, cancel pending auto-continuation |
 | `Esc` ×2 | Interrupt the run when pressed within 1.5 seconds. Other keys or closing a menu cancel the first press |
 | `Esc` again after interruption | **Force-stop** a run that is still stopping, kill foreground shell commands and return to `idle`; background services remain running |
-| `Ctrl+C` | Stop the background **services** the agent started, when there are any. It does not arm exit |
+| `Ctrl+C` | **Copy the selection** when a drag-selection is waiting (release no longer copies by itself). Otherwise: stop the background **services** the agent started, when there are any. It does not arm exit |
 | `Ctrl+C` ×2 | Exit (first press warns), once no services are running |
-| `Ctrl+L` | Redraw the frame (clear the screen in inline mode) |
-| `Ctrl+T` | Show / hide thinking (**off by default**). In inline mode it applies to the live entry and everything drawn after it — settled entries are already in your terminal's scrollback and cannot be repainted |
+| `Ctrl+L` | Redraw the frame |
+| `Ctrl+T` | Show / hide thinking (**off by default**) for the entire session |
 | `Ctrl+O` | Expand / collapse the most recent tool, team, compaction or **service** card — including a diff or a service's log tail |
 | `PgUp` / `PgDn` | Scroll the transcript a page (full-screen mode) |
 | `Shift+↑` / `Shift+↓` | Scroll the transcript a line (full-screen mode) |
@@ -465,9 +463,7 @@ scrollbar occupies the terminal's last column (including 40-column windows). Dra
 to move through history, or click the track to move one page. Messages, the team panel
 and the input area scroll together; typing or editing returns to the input without
 losing the draft. The fixed status row shows `PgDn down` while reading history.
-`PgDn` moves one page down; editing or submitting returns directly to the end. In **inline mode**
-(`--no-fullscreen`) history lives in your terminal's native scrollback, printed
-once via Ink `<Static>`.
+`PgDn` moves one page down; editing or submitting returns directly to the end.
 
 **The mouse wheel is a viewport gesture and nothing else.** Wherever the pointer
 happens to be — including right on top of the composer — a notch scrolls the
@@ -478,10 +474,12 @@ open. It never touches your draft, and it never recalls prompt history: that is
 Two things follow from the app owning the mouse, and they are worth knowing
 before you meet them:
 
-- **Dragging text selects and copies it when mouse selection is enabled.**
-  Dragging the last-column scrollbar scrolls instead. `--no-mouse-select` disables
-  text selection while preserving scrollbar dragging. `Shift`+drag uses the
-  terminal's native selection on hosts that support that convention.
+- **Dragging selects; `Ctrl+C` copies.** Releasing the mouse keeps the
+  highlight (and freezes the rows under it) until you commit: `Ctrl+C` over a
+  pending selection copies it and never arms exit; any other key drops it.
+  Dragging the last-column scrollbar scrolls instead. `--no-mouse-select`
+  disables text selection while preserving scrollbar dragging. `Shift`+drag
+  uses the terminal's native selection on hosts that support that convention.
 - **`tmux` with `mouse on` keeps the wheel for itself**, so the app never sees
   it. `set -g mouse off` hands it back.
 - **A Windows console on Node below 22.17.0 cannot report the wheel at all**,
@@ -493,6 +491,48 @@ Turn the whole thing off with `--no-mouse`, `ARAGON_MOUSE=0`, or
 nothing in full-screen mode — it does not fall back to editing your draft, which
 is what it used to do. Inline mode is unaffected either way.
 
+## Terminal setup for Shift+Enter
+
+`Ctrl+J` and `Alt+Enter` insert a newline in every terminal with no setup. A
+dedicated `Shift+Enter` key needs the terminal to send the CSI-u sequence
+`\u001b[13;2u` instead of a plain `Enter` — run `/terminal-setup` inside aragon
+for copy-paste bindings per terminal:
+
+- **Windows Terminal** — `settings.json`, `actions`:
+  `{ "keys": "shift+enter", "command": { "action": "sendInput", "input": "\u001b[13;2u" } }`
+- **VS Code integrated terminal** — `keybindings.json`:
+  `{ "key": "shift+enter", "command": "workbench.action.terminal.sendSequence", "args": { "text": "\u001b[13;2u" } }`
+- **iTerm2** — Settings > Profiles > Keys > Key Mappings, `Shift+Enter` ->
+  *Send Escape Sequence* `[13;2u`
+- **kitty / WezTerm / foot** — native once keyboard-enhancement mode is on
+  (kitty: `map shift+enter send_text all \u001b[13;2u`)
+- **Alacritty** — `[keyboard]` bindings in `alacritty.toml`:
+  `{ key = "Enter", mods = "Shift", chars = "\u001b[13;2u" }`
+- **macOS Terminal.app / legacy conhost** — cannot rebind `Enter`; use
+  `Ctrl+J` or `Alt+Enter`
+
+aragon deliberately does not push kitty keyboard mode or `modifyOtherKeys=2`
+onto the terminal itself: mode 2 re-encodes *every* modified key as CSI-u and
+would blind this app's (and Ink's) key parsing for the whole session. One
+user-side binding reaches the same result for the one key that needs it.
+
+## Queued messages while the agent runs
+
+Sending a message while a run is in flight **queues** it for the agent's next
+turn. The transcript shows the queue honestly:
+
+```
+◷ Queue: also check the tests
+```
+
+The row stays until Core actually accepts the message into the conversation —
+the next `turn_start` — at which point it is rewritten **in place** into the
+user message it became, followed by the answer. Interrupting the run does not
+clear it: the message really is still queued, and it will be picked up when
+you continue the session. Messages still queued when the session is saved are
+restored as `Queued but never sent: …` warnings, because the in-memory queue
+does not survive a restart.
+
 ## Themes
 
 Three palettes plus `auto`:
@@ -538,19 +578,22 @@ session, `showThinking: true` (or `--show-thinking`, or `ARAGON_SHOW_THINKING=1`
 shows them permanently, and `/settings` has a **Show thinking** row next to
 **Thinking**.
 
-In **inline** mode (`--no-fullscreen`) the marker reads `✱ thought for 12s` with
-no key hint, and the toast on `Ctrl+T` says `Thinking shown for new output.`
-Settled entries there have already been printed into your terminal's own
-scrollback, which cannot be repainted — so the key applies to the live entry and
-to everything drawn after it, and the row does not offer what it cannot deliver.
-
-**There is something to look at while it thinks.** One row above the composer,
-for as long as a run is in flight:
+**There is something to look at while it thinks.** One row directly above the
+input box, for as long as a run is in flight, with the keys that act on the run
+beside it (it takes the place of the idle hint row below the box, so starting a
+run never moves the layout):
 
 ```
-  ⠋ Percolating…
+  ⠋ Percolating… · ⏎ queue · esc×2 interrupt · ctrl+c×2 exit            PLAN
 ```
 
+The row scrolls with the input box. On a narrow terminal whole clauses are dropped
+from the end (`exit` first) and only the phrase is shortened with an ellipsis, so
+`queue` and `interrupt` are never cut mid-word. When you scroll history until the
+input box is out of view, the animation moves to the fixed row at the bottom, so
+there is always exactly one spinner on screen. Terminals shorter than 20 rows, and
+`--no-hints`, keep the animation on that fixed row instead.
+
 The word rotates every four seconds. It carries no clock and no token count on
 purpose: the status bar one row below already has both, under the same
 condition, driven by the same ticker. Under `reducedMotion` the spinner is a
@@ -637,17 +680,28 @@ On exit the session is replayed into the normal buffer as plain text so the
 conversation survives leaving the screen (`--no-exit-transcript` opts out;
 `/save` still exports the full JSON).
 
-It downgrades to the inline renderer automatically when stdout is not a TTY,
-`TERM=dumb`, a CI environment variable is set, or the terminal is under 12 rows
-or 40 columns. `--fullscreen` overrides all of those except the non-TTY check —
-writing screen-control sequences into a pipe or a redirected file is never safe.
-`--no-fullscreen`, `ARAGON_FULLSCREEN=0`, or `aragon config set fullscreen false`
-opt out permanently.
+Interactive sessions always use the full-screen TUI, including under `TERM=dumb`
+or CI. Windows smaller than 40 columns or 12 rows show a size notice and recover
+when enlarged, preserving the session and draft.
+
+The layout switches `--fullscreen`, `--no-fullscreen`, `ARAGON_FULLSCREEN` and
+`fullscreen` configuration have been removed. Old command-line flags are errors;
+old config fields and environment variables are ignored.
+
+Interactive screens require both stdin and stdout to be TTYs. `aragon config`
+without a TTY exits with code 2; use `aragon config list` or `aragon config set`.
+`-p`, `exec`, piped input and prompt-based redirected output retain their
+headless behavior and never enter the alternate screen.
 
 The alternate screen uses an application scrollbar rather than the terminal's
 native scrollback. Windows Terminal and conhost have different mouse/VT capabilities;
 the PowerShell version alone does not determine dragging support. If the host cannot
 deliver SGR mouse reports, the scrollbar remains visible and keyboard scrolling works.
+The differential renderer never erases the last column: a row that already fills the
+terminal width is repainted without a trailing erase-to-end-of-line, which on Windows
+console hosts would otherwise wipe the scrollbar cell on every redrawn row. `/perf`
+prints a `scrollbar:` line (column, mouse state, whether a full frame was confirmed,
+thumb position, glyph tier) to tell "cannot see it" from "cannot drag it".
 `--no-mouse` provides that keyboard-only path; `--no-diff-render` disables output
 differencing without disabling the scrollbar. A mid-session resize below 40 columns
 or 12 rows displays a size hint and preserves the draft for restoration.
@@ -694,13 +748,6 @@ rung. `--no-render-governor` (or `renderGovernor: false`, or
 `ARAGON_RENDER_GOVERNOR=0`) turns the adaptation off entirely; the output is
 identical, just heavier under load.
 
-**Inline mode** (`--no-fullscreen`, `TERM=dumb`, CI, a very small terminal) gets
-one extra guarantee. Ink repaints the *entire session history* on every frame
-once a live entry grows taller than the terminal, which is a genuine freeze
-rather than a slowdown. The live region is now clamped to stay below that
-threshold, with a `... N earlier lines` marker naming what is deferred; the
-entry prints **in full** into your scrollback the moment it finishes.
-
 ## Plan mode
 
 `aragon` has two postures, and `Shift+Tab` switches between them. `/plan` does
@@ -758,8 +805,7 @@ still exits the session outright.
 ### Where the mode is shown
 
 The status bar always names a non-default mode (`PLAN`, or `PLAN → BUILD` while
-a switch is pending), including on short terminals, with `--no-hints`, and in
-inline mode. The composer additionally shows a `PLAN` chip on its hint row and
+a switch is pending), including on short terminals and with `--no-hints`. The composer additionally shows a `PLAN` chip on its hint row and
 tints its border, where there is room for it.
 
 ### Switching mid-run
@@ -830,8 +876,8 @@ for a question, or for a search with no work attached — and a one-item list
 against an empty plan is refused outright. A session that never plans has
 byte-identical layout to a build without this feature.
 
-**The list shows the current task's plan.** Starting a new task or explicitly
-clearing the list removes this live view while preserving conversation history:
+**The list shows the current task's plan.** Starting a new task or explicitly
+clearing the list removes this live view while preserving conversation history:
 
 - `/clear` and `/todo clear` take the plan off the screen. Neither touches the
   conversation, so the model still knows what it was doing and its next update
@@ -844,7 +890,7 @@ clearing the list removes this live view while preserving conversation history:
   when it starts successfully. The completed run keeps its final plan visible.
   `/todo continue` and automatic continuation preserve the plan and its timestamp;
   ordinary text such as "continue" starts a new task. Steering keeps the current
-  plan. Until a valid new `todo_write`, the rail, inline strip and count stay hidden.
+  plan. Until a valid new `todo_write`, the rail and count stay hidden.
   Startup failure or cancellation keeps the old plan. History cards remain available.
 - During a live run, interrupt before `/reset` or `/resume`. A conversation switch
   while startup is waiting cancels that pending request and returns the view to idle.
@@ -1810,7 +1856,7 @@ env / `.env` → CLI flags**.
   `ARAGON_CONTEXT_WINDOW` (a number, or `auto`; the denominator the context
   gauge measures against),
   `ARAGON_THEME`,
-  `ARAGON_FULLSCREEN`, `ARAGON_PLAN` (`1` starts in plan mode),
+  `ARAGON_PLAN` (`1` starts in plan mode),
   `ARAGON_TEAM` (`0` disables team subagents), `ARAGON_TEAM_MAX` (fan-out width),
   `ARAGON_TODO` (`0` disables todo planning entirely),
   `ARAGON_TODO_FOLLOW` (`notify` / `auto` / `off`),
@@ -1839,7 +1885,6 @@ env / `.env` → CLI flags**.
 | `contextWindow` | `null` | The window the context gauge measures against, clamped to `[8000, 5000000]`. `null` means **auto** — the built-in model table, or a fabricated 128000 for a model it has never seen. Set it when the gauge shows `?` on the denominator. See [The context gauge](#the-context-gauge). |
 | `showThinking` | `false` | Draw the reasoning the model returns. `thinkingLevel` is the effort the provider is asked to **spend**; this is whether the terminal **shows** it. Off by default: a settled turn that thought leaves one muted `thought for 12s` row in its place, so nothing is hidden silently. |
 | `liveToolOutput` | `true` | Draw up to eight sanitised rows of a **running** tool's output on its card, plus a `no output for Ns` row when the child goes quiet. On by default, unlike `showThinking`: this ADDS the information a long `bash` call otherwise hides, and its cost is bounded by construction — eight rows per call, sixteen calls, whatever the command emits. Turn it off and the card is a single `running` row again, with no store allocated and no recorder attached. |
-| `fullscreen` | `true` | Use the full-screen TUI (still subject to the automatic downgrades). Setting it to `false` opts out permanently; leaving it `true` is *not* a force — only `--fullscreen` / `ARAGON_FULLSCREEN=1` override the heuristics. |
 | `exitTranscript` | `true` | Replay a plain-text session summary after exiting (full-screen only). |
 | `transcriptWindow` | `1000` | How far back you can scroll, clamped to `[50, 20000]`. Entries beyond it collapse into one line. It is no longer a rendering budget: the viewport is virtualised, so off-screen entries are not laid out at all. |
 | `transcriptRetain` | `1000` | Entries kept in memory, clamped to `[200, 20000]` and raised to `transcriptWindow` if the two conflict — with a startup notice saying so, because retaining less than you can scroll to would otherwise make part of the horizon quietly unreachable. Older entries are dropped, counted, and reported by `/perf` and by the exit replay. |
@@ -1857,7 +1902,7 @@ env / `.env` → CLI flags**.
 | `team.dispatchTimeoutMs` | `900000` | The whole dispatch's wall clock, clamped to `[60000, 3600000]`. |
 | `team.maxTurnsPerSubagent` | `24` | Runaway-loop cap, clamped to `[4, 100]`. A subagent stopped here is reported as `stopped: turn cap reached`, with whatever it had produced. |
 | `todo.enabled` | `true` | Master switch for todo planning. `false` means the `todo_write` tool is never registered — `/todo on` in such a session saves the setting for next launch and says so, rather than advertising a tool that is not there. |
-| `todo.panel` | `true` | Render the right-hand rail (and, in inline mode, the one-row plan strip). Independent of `todo.enabled` on purpose: a screen-reader user wants the planning discipline without the column, and the system prompt varies one sentence accordingly. |
+| `todo.panel` | `true` | Render the right-hand rail. Independent of `todo.enabled` on purpose: a screen-reader user wants the planning discipline without the column, and the system prompt varies one sentence accordingly. |
 | `todo.followThrough` | `"notify"` | What happens when a run ends with steps left. `notify` says so and stops; `auto` continues the plan after a 3-second grace window, bounded by two structural limits (one fruitless attempt, 25 continuations per plan) that are deliberately **not** settings; `off` is silent. An unrecognized value falls back to `notify`. |
 | `retry.enabled` | `true` | Retry a failed provider call at all. Unlike `team.enabled` / `todo.enabled` this decides nothing at construction — the policy lives on the provider registry, so `/retry on` / `/retry off` takes effect **in the running session**, subagents included. |
 | `retry.maxRetries` | `10` | Retries **after** the first attempt, clamped to `[0, 20]`. `0` is the kill switch and is a real value here, not a typo for the default. The 20 is a hard ceiling enforced twice — on the config value and again inside the engine — so hand-editing this file cannot raise it. |

package/dist/agent/controller.d.ts

@@ -61,7 +61,7 @@ export interface ControllerDeps {
     humanInput?: HumanInputGate;
     /**
      * Whether this session could EVER render the todo rail (todo-plan-execution
-     * §3.7 / P1-6). Full-screen and interactive; `-p` and inline mode are false.
+     * §3.7 / P1-6). Full-screen and interactive; headless calls are false.
      *
      * It varies ONE SENTENCE of the `<todo_planning>` block, and nothing else.
      * Telling the model "the user sees this list in a panel beside the

package/dist/agent/reducer.d.ts

@@ -36,6 +36,21 @@ export type Entry = {
     id: string;
     kind: 'user';
     text: string;
+}
+/**
+ * A message submitted while a run was in flight (tui-shift-enter-copy-queue
+ * 5.2). It renders as `Queue: <first line>` and is rewritten IN PLACE, same
+ * id, into the `user` entry above by the `turnStart` that proves Core
+ * accepted it -- so the id-based height cache re-measures rather than
+ * freezes, and the settled boundary (positional, monotonic) never moves.
+ *
+ * NOT A LIVE TARGET: no id pointer in `ViewState` names it, so the trim
+ * ring and the `appendEntry` assertions need no new member.
+ */
+ | {
+    id: string;
+    kind: 'queued';
+    text: string;
 } | {
     id: string;
     kind: 'assistant';
@@ -115,20 +130,7 @@ export type Entry = {
     kind: 'notice';
     level: NoticeLevel;
     text: string;
-}
-/**
- * A team dispatch (team-subagents §5.3). JSON-SERIALIZABLE by construction, so
- * `/save` and `/resume` need no format change.
- *
- * `active: true` is normalized away ON LOAD, in `session/persist.ts`, not here
- * (P1-5). A session saved mid-dispatch would otherwise resume with a card that
- * claims to be running while nothing is: it spins forever, and because
- * `Transcript`'s settled boundary is MONOTONIC an entry that never settles
- * never reaches `<Static>` and is re-rendered on every frame for the rest of
- * the session. The reducer never legitimately sees a stale-active entry — only
- * the file does — which is why the fix belongs in the load path.
- */
- | {
+} | {
     id: string;
     kind: 'team';
     dispatchId: string;
@@ -137,20 +139,7 @@ export type Entry = {
     aborted: boolean;
     durationMs?: number;
     active: boolean;
-}
-/**
- * The turn's todo checklist (todo-plan-execution §3.8). ONE per user turn,
- * rewritten in place — a 7-item list updated twice per item would otherwise be
- * 14 near-identical tool cards in the one surface the user reads most.
- *
- * JSON-SERIALIZABLE by construction, so `/save` needs no format change.
- *
- * `live: true` is normalized away ON LOAD, in `session/persist.ts`, for the
- * same reason the team entry's `active` is (C-5): `Transcript`'s settled
- * boundary is MONOTONIC, so an entry that never settles never reaches
- * `<Static>` and is re-rendered on every frame for the rest of the session.
- */
- | {
+} | {
     id: string;
     kind: 'todo';
     items: TodoItem[];
@@ -165,21 +154,7 @@ export type Entry = {
      * card this session produced.
      */
     interrupted?: boolean;
-}
-/**
- * The turn's retry story (llm-api-retry-backoff §6.4). ONE per turn, REWRITTEN
- * IN PLACE — ten retries rendered as ten notices would bury the transcript in
- * the one moment the user most needs it legible.
- *
- * JSON-SERIALIZABLE by construction, so `/save` needs no format change.
- *
- * `phase: 'waiting' | 'retrying'` is normalized away ON LOAD, in
- * `session/persist.ts`, exactly as the team entry's `active` and the todo
- * card's `live` are (R-10): `Transcript`'s settled boundary is MONOTONIC, so a
- * card that never settles never reaches `<Static>` and is re-rendered on every
- * frame for the rest of the session.
- */
- | {
+} | {
     id: string;
     kind: 'retry';
     /** 1-based retry index of the LATEST attempt this card has seen. */
@@ -195,22 +170,7 @@ export type Entry = {
     /** Filled on settle. */
     totalRetries?: number;
     elapsedMs?: number;
-}
-/**
- * One fast-model review (fast-model-tier §5.4). ONE CARD PER REVIEW, never
- * rewritten across reviews — a critique is a single event, unlike the todo
- * card's per-turn checklist.
- *
- * JSON-SERIALIZABLE by construction, so `/save` needs no format change.
- *
- * `live: true` is normalized away ON LOAD, in `session/persist.ts`, next to
- * `team.active`, `todo.live` and the retry card's phase, and for the identical
- * reason (C-5): `Transcript`'s settled boundary is MONOTONIC, so an entry that
- * never settles never reaches `<Static>` and is re-rendered on every frame for
- * the rest of the session. The reducer never legitimately sees a stale-live
- * entry — only the file does.
- */
- | {
+} | {
     id: string;
     kind: 'fast';
     /** 1-based within the session, so the card can name itself. */
@@ -224,27 +184,7 @@ export type Entry = {
     turn: number;
     durationMs?: number;
     live: boolean;
-}
-/**
- * One context compaction (context-auto-compaction §5.5). ONE CARD PER
- * COMPACTION, rewritten once when the call returns — like the fast card, and
- * unlike the todo card's per-turn checklist.
- *
- * JSON-SERIALIZABLE by construction, so `/save` and `/resume` need no format
- * change.
- *
- * `live: true` is normalized away ON LOAD, in `session/persist.ts`, next to
- * `team.active`, `todo.live`, the retry card's phase and `fast.live`, and for
- * the identical reason (C-8): `Transcript`'s settled boundary is MONOTONIC, so
- * an entry that never settles never reaches `<Static>` and is re-rendered on
- * every frame for the rest of the session.
- *
- * EVERY STRING ON THIS CARD DESCRIBES THE HISTORY, NEVER THE SCREEN (P2-9).
- * `/clear` empties the transcript view but not `messages`, so a compaction can
- * legitimately fire against a visually empty transcript and report
- * "112 -> 9 messages" with nine visible rows.
- */
- | {
+} | {
     id: string;
     kind: 'compaction';
     /** 1-based within the session, so the card can name itself. */
@@ -275,29 +215,7 @@ export type Entry = {
         charsRemoved: number;
     };
     live: boolean;
-}
-/**
- * One supervised background service (background-service-supervision §3.10).
- *
- * JSON-SERIALIZABLE by construction, so `/save` needs no format change.
- *
- * IT IS THE ONE LIVE CARD THAT DOES **NOT** BLOCK THE SETTLED BOUNDARY ONCE IT
- * IS UP (D-11 / P0-4). Every other live clause in
- * `Transcript.tsx::computeSettledCount` is bounded by an OPERATION — a
- * dispatch, a turn, one LLM call, one countdown — and each carries a comment
- * saying that an entry which never settles is re-rendered on every frame for
- * the rest of the session. A service is bounded by the USER'S INTENT: a dev
- * server left up for an hour is the normal case. So only `starting` blocks
- * (bounded by `readyTimeoutMs`), a `ready`/`running` card commits its body to
- * `<Static>`, and a later terminal transition appends a NEW one-row entry
- * rather than trying to rewrite a card `<Static>` has already printed.
- *
- * `terminal: true` marks that second entry. `session/persist.ts` rewrites any
- * restored non-terminal card to `stopped`, next to `team.active`, `todo.live`,
- * the retry phase, `fast.live` and `compaction.live`, and for the identical
- * reason.
- */
- | {
+} | {
     id: string;
     kind: 'service';
     /** `s1`, `s2`, ... — the id the model quotes back. */
@@ -476,6 +394,15 @@ export declare function initialViewState(seed?: ViewStateSeed): ViewState;
 export type ViewAction = {
     type: 'submit';
     text: string;
+}
+/**
+ * `controller.steer` was accepted; append the persistent queue marker
+ * (5.2.2). Dispatched from exactly one place -- `App.submitMessage`'s
+ * running branch -- beside the `controller.steer` call it mirrors.
+ */
+ | {
+    type: 'steerQueued';
+    text: string;
 } | {
     type: 'runStart';
 } | {
@@ -677,16 +604,7 @@ export type ViewAction = {
 } | {
     type: 'serviceUpdate';
     service: ServiceSnapshot;
-}
-/**
- * A service reached a terminal state.
- *
- * IT APPENDS A NEW ONE-ROW ENTRY rather than rewriting the card (D-11). The
- * card may already have been printed into `<Static>`, and `<Static>` cannot
- * un-print — so any design that rewrites a printed card is wrong whatever it
- * claims. A second event at a second time is also simply more honest.
- */
- | {
+} | {
     type: 'serviceEnd';
     service: ServiceSnapshot;
 };