fix(ai-chat): branch sendNow on live status to defuse stale-status race (#198)

Port the only substantive fix #211 was missing relative to #203 (which is
being closed): the "Send now" handler branched on the closure-captured
isStreaming, but a turn can finish between render and click. In that window
stop() is a no-op, so arming flushOnAbortRef/interruptNextSendRef would strand
those one-shot flags and leak into a later, unrelated Stop (auto-sending a
queued message the user never asked to send).

- Mirror the live useChat status in statusRef (updated each render) and branch
  sendNow on it instead of isStreaming, so the not-streaming path runs when the
  turn has already ended and the interrupt flags are never armed against a
  no-op stop().
- Belt-and-suspenders: clear flushOnAbortRef/interruptNextSendRef when a new
  turn starts streaming, defusing the sub-render-tick window where a flag could
  still be armed but the expected abort never fired. No-op for the legit
  interrupt path (both refs are consumed synchronously beforehand).

Keeps #211's existing structure and its flushNext-returns-boolean fix. The
rest of #203's divergence is comment rewording, a server-side rename of the
same pure interrupt-gate, and fewer tests — nothing else to port.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

CHANGELOG.md

@@ -10,6 +10,15 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+### Added
+
+- **Interrupt the AI agent and send a queued message now.** A queued AI-chat
+  message gains a "send now" action that interrupts the streaming turn and
+  immediately sends that message, keeping the agent's partial output. The
+  follow-up turn is tagged as an interrupt so the model is told its previous
+  answer was cut off and builds on it instead of restarting; the rest of the
+  queue still flushes normally afterward. (#198)
+
 ## [0.94.0] - 2026-06-26
 
 This release makes AI chat durable and fast: assistant turns are persisted to
@@ -75,15 +84,6 @@ per-workspace rolling-day token budget.
 
 ### Changed
 
-- **AI chat now feeds the model the full stored transcript.** The per-turn model
-  conversation was rebuilt from a sliding window of the 50 most recent stored
-  rows, which silently dropped the beginning of any longer chat. It is now
-  rebuilt from the complete non-deleted transcript in chronological order, so
-  the model sees every turn (a 5000-row backstop guards process memory — a
-  safety net far above any realistic chat, not a conversational limit). On a
-  very long chat this can eventually reach the model's context window; the
-  client already surfaces that as "start a new chat". (#202)
-
 - **AI chat default provider is now `openai-compatible` (reasoning surfaced).**
   For the `openai` driver the chat provider defaults to the openai-compatible
   implementation, so a workspace pointing at z.ai/GLM/DeepSeek now streams the

apps/client/public/locales/en-US/translation.json

@@ -1180,6 +1180,8 @@
   "Send when the agent finishes": "Send when the agent finishes",
   "Queue message": "Queue message",
   "Remove queued message": "Remove queued message",
+  "Send now": "Send now",
+  "Interrupt and send now": "Interrupt and send now",
   "Stop": "Stop",
   "Response stopped.": "Response stopped.",
   "Connection lost — the answer was interrupted.": "Connection lost — the answer was interrupted.",

apps/client/public/locales/ru-RU/translation.json

@@ -723,6 +723,8 @@
   "Send when the agent finishes": "Отправить, когда агент закончит",
   "Queue message": "Поставить в очередь",
   "Remove queued message": "Убрать из очереди",
+  "Send now": "Отправить сейчас",
+  "Interrupt and send now": "Прервать и отправить сейчас",
   "Something went wrong": "Что-то пошло не так",
   "Stop": "Стоп",
   "The AI agent could not respond. Please try again.": "AI-агент не смог ответить. Попробуйте ещё раз.",

apps/client/src/features/ai-chat/components/chat-thread.test.tsx

@@ -0,0 +1,142 @@
+import { describe, it, expect, beforeEach, vi } from "vitest";
+import { render, screen, fireEvent, act } from "@testing-library/react";
+import { MantineProvider } from "@mantine/core";
+
+// Shared, hoisted mock state so the @ai-sdk/react and "ai" module mocks (hoisted
+// above the imports) can expose the captured useChat callbacks / transport and
+// the spies back to the test body.
+const h = vi.hoisted(() => ({
+  state: {
+    status: "streaming" as string,
+    onFinish: null as null | ((arg: Record<string, unknown>) => void),
+    sendMessage: vi.fn(),
+    stop: vi.fn(),
+    transport: null as null | {
+      prepareSendMessagesRequest: (arg: {
+        messages: unknown[];
+        body: Record<string, unknown>;
+      }) => { body: Record<string, unknown> };
+    },
+  },
+}));
+
+// Mock useChat: capture onFinish, return the spies and the controllable status.
+vi.mock("@ai-sdk/react", () => ({
+  useChat: (opts: { onFinish?: (arg: Record<string, unknown>) => void }) => {
+    h.state.onFinish = opts.onFinish ?? null;
+    return {
+      messages: [],
+      sendMessage: h.state.sendMessage,
+      status: h.state.status,
+      stop: h.state.stop,
+      error: null,
+    };
+  },
+}));
+
+// Mock "ai": deterministic ids + a transport that records its options so the test
+// can invoke prepareSendMessagesRequest and assert the `interrupted` flag.
+vi.mock("ai", () => {
+  let counter = 0;
+  return {
+    generateId: () => `gid-${counter++}`,
+    DefaultChatTransport: class {
+      constructor(opts: {
+        prepareSendMessagesRequest: (arg: {
+          messages: unknown[];
+          body: Record<string, unknown>;
+        }) => { body: Record<string, unknown> };
+      }) {
+        h.state.transport = opts;
+      }
+    },
+  };
+});
+
+// Stub the heavy children: MessageList (markdown/render) and ChatInput (the
+// composer). The ChatInput stub exposes a button that queues a message, the only
+// interaction this test needs to populate the queue while "streaming".
+vi.mock("@/features/ai-chat/components/message-list.tsx", () => ({
+  default: () => <div data-testid="message-list" />,
+}));
+vi.mock("@/features/ai-chat/components/chat-input.tsx", () => ({
+  default: ({ onQueue }: { onQueue: (text: string) => void }) => (
+    <button data-testid="queue-btn" onClick={() => onQueue("queued text")}>
+      queue
+    </button>
+  ),
+}));
+
+import ChatThread from "./chat-thread";
+
+function renderThread() {
+  const onTurnFinished = vi.fn();
+  render(
+    <MantineProvider>
+      <ChatThread chatId="c1" initialRows={[]} onTurnFinished={onTurnFinished} />
+    </MantineProvider>,
+  );
+  return { onTurnFinished };
+}
+
+describe("ChatThread — send now (#198)", () => {
+  beforeEach(() => {
+    h.state.status = "streaming";
+    h.state.onFinish = null;
+    h.state.sendMessage.mockClear();
+    h.state.stop.mockClear();
+    h.state.transport = null;
+  });
+
+  it("aborts the current turn and resends the queued message on the abort", () => {
+    renderThread();
+
+    // Queue a message while the turn is streaming.
+    fireEvent.click(screen.getByTestId("queue-btn"));
+    const sendNowBtn = screen.getByLabelText("Send now");
+    expect(sendNowBtn).toBeTruthy();
+
+    // "Send now" interrupts the current turn (stop), but does NOT send yet —
+    // the resend happens once the abort lands in onFinish.
+    fireEvent.click(sendNowBtn);
+    expect(h.state.stop).toHaveBeenCalledTimes(1);
+    expect(h.state.sendMessage).not.toHaveBeenCalled();
+
+    // The abort we triggered reaches onFinish: the promoted head is flushed.
+    act(() => {
+      h.state.onFinish?.({
+        message: { id: "a", role: "assistant", parts: [] },
+        isAbort: true,
+        isDisconnect: false,
+        isError: false,
+      });
+    });
+    expect(h.state.sendMessage).toHaveBeenCalledWith({ text: "queued text" });
+  });
+
+  it("tags exactly the next send as interrupted (one-shot flag)", () => {
+    renderThread();
+    fireEvent.click(screen.getByTestId("queue-btn"));
+    fireEvent.click(screen.getByLabelText("Send now"));
+
+    const prep = h.state.transport!.prepareSendMessagesRequest;
+    // The send right after "send now" carries interrupted: true...
+    expect(prep({ messages: [], body: {} }).body.interrupted).toBe(true);
+    // ...and only that one (the flag is read-and-cleared).
+    expect(prep({ messages: [], body: {} }).body.interrupted).toBe(false);
+  });
+
+  it("sends immediately without an interrupt when not streaming", () => {
+    h.state.status = "ready";
+    renderThread();
+
+    fireEvent.click(screen.getByTestId("queue-btn"));
+    fireEvent.click(screen.getByLabelText("Send now"));
+
+    // No turn to interrupt: sent straight away, no abort, not flagged.
+    expect(h.state.stop).not.toHaveBeenCalled();
+    expect(h.state.sendMessage).toHaveBeenCalledWith({ text: "queued text" });
+    const prep = h.state.transport!.prepareSendMessagesRequest;
+    expect(prep({ messages: [], body: {} }).body.interrupted).toBe(false);
+  });
+});

apps/client/src/features/ai-chat/components/chat-thread.tsx

@@ -1,7 +1,11 @@
 import { useCallback, useEffect, useMemo, useRef, useState } from "react";
 import { generateId } from "ai";
-import { ActionIcon, Box, Group, Stack, Text } from "@mantine/core";
-import { IconClockHour4, IconX } from "@tabler/icons-react";
+import { ActionIcon, Box, Group, Stack, Text, Tooltip } from "@mantine/core";
+import {
+  IconClockHour4,
+  IconPlayerPlayFilled,
+  IconX,
+} from "@tabler/icons-react";
 import { useTranslation } from "react-i18next";
 import { useChat, type UIMessage } from "@ai-sdk/react";
 import { DefaultChatTransport } from "ai";
@@ -23,6 +27,7 @@ import { extractServerChatId } from "@/features/ai-chat/utils/adopt-chat-id.ts";
 import {
   dequeue,
   enqueueMessage,
+  promoteToHead,
   removeQueuedById,
   type QueuedMessage,
 } from "@/features/ai-chat/utils/queue-helpers.ts";
@@ -201,12 +206,25 @@ export default function ChatThread({
   // helper can call the current instance from the stable `onFinish` callback.
   const sendMessageRef = useRef<((m: { text: string }) => void) | null>(null);
 
+  // "Send now" single-flight flags. Kept in refs (not state) so they are read
+  // inside the stable `onFinish` callback and the transport closure WITHOUT a
+  // re-render or a stale closure. Both are one-shot (read-and-clear).
+  // - flushOnAbortRef: flush the promoted head on the abort WE triggered, even
+  //   though an aborted turn normally keeps the queue intact.
+  // - interruptNextSendRef: tag the next send as a user interrupt so the server
+  //   injects the "your previous answer was interrupted" note for that turn only.
+  const flushOnAbortRef = useRef(false);
+  const interruptNextSendRef = useRef(false);
+
   // FIFO dequeue + send the next queued message (no-op when the queue is empty).
+  // Returns whether a message was actually sent, so callers can tell an empty
+  // dequeue (nothing to flush) from a real send.
   const flushNext = useCallback(() => {
     const { head, rest } = dequeue(queuedRef.current);
-    if (!head) return;
+    if (!head) return false;
     setQueue(rest);
     sendMessageRef.current?.({ text: head.text });
+    return true;
   }, [setQueue]);
 
   const enqueue = useCallback(
@@ -232,17 +250,26 @@ export default function ChatThread({
         // when null) and tell the agent which page "this page" refers to. Both
         // are read live from refs so changing chats/pages does NOT recreate the
         // transport. `openPage` is null on a non-page route.
-        prepareSendMessagesRequest: ({ messages, body }) => ({
-          body: {
-            ...body,
-            chatId: chatIdRef.current,
-            openPage: openPageRef.current,
-            // Honoured by the server only when creating a new chat; null =>
-            // universal assistant.
-            roleId: roleIdRef.current,
-            messages,
-          },
-        }),
+        prepareSendMessagesRequest: ({ messages, body }) => {
+          // Read-and-clear the interrupt flag so the "you were interrupted" note
+          // is carried by ONLY this request (the one resending the promoted
+          // message right after we aborted the previous turn). The server still
+          // confirms it against history before acting on it.
+          const interrupted = interruptNextSendRef.current;
+          interruptNextSendRef.current = false; // one-shot
+          return {
+            body: {
+              ...body,
+              chatId: chatIdRef.current,
+              openPage: openPageRef.current,
+              // Honoured by the server only when creating a new chat; null =>
+              // universal assistant.
+              roleId: roleIdRef.current,
+              interrupted,
+              messages,
+            },
+          };
+        },
       }),
     [],
   );
@@ -277,6 +304,21 @@ export default function ChatThread({
       else if (isAbort) setStopNotice("manual");
       else if (isDisconnect) setStopNotice("disconnect");
       else setStopNotice(null);
+      // "Send now": WE triggered this abort to interrupt the current turn and
+      // immediately send the promoted head. Flush it even though the turn was
+      // aborted (the normal abort path below keeps the queue intact). The
+      // interrupt note travels with this send via interruptNextSendRef.
+      if (flushOnAbortRef.current) {
+        flushOnAbortRef.current = false;
+        // Suppress the "Response stopped." flash for an intentional interrupt.
+        setStopNotice(null);
+        // If the promoted head vanished (e.g. the user removed it before the
+        // abort landed) flushNext sends nothing — clear the one-shot interrupt
+        // tag so it can't leak onto the next unrelated send. On a real send the
+        // tag is consumed by prepareSendMessagesRequest and stays untouched.
+        if (!flushNext()) interruptNextSendRef.current = false;
+        return;
+      }
       if (isAbort || isDisconnect || isError) return;
       flushNext();
     },
@@ -298,6 +340,13 @@ export default function ChatThread({
   // Keep the flush helper pointed at the latest sendMessage instance.
   sendMessageRef.current = sendMessage;
 
+  // Mirror the live turn status in a ref so event handlers (sendNow) branch on the
+  // CURRENT status rather than a value captured in a stale render closure — a turn
+  // can finish between render and click, and arming the interrupt refs against a
+  // no-op stop() would leave them set to leak into a later, unrelated Stop.
+  const statusRef = useRef(status);
+  statusRef.current = status;
+
   // EARLY chat-id adoption (#174): the server streams the authoritative chat id
   // on the assistant message metadata at the `start` chunk (message.metadata.
   // chatId — see adopt-chat-id.ts / chatStreamMetadata). Forward it to the parent
@@ -329,9 +378,49 @@ export default function ChatThread({
 
   const isStreaming = status === "submitted" || status === "streaming";
 
-  // Clear the stopped marker as soon as a new turn begins streaming.
+  // "Send now" on a queued message: interrupt the current turn and immediately
+  // send THIS message, keeping the agent's partial output. Other queued messages
+  // stay queued and flush normally after the new turn. Reuses the existing
+  // queue/flush machinery: promote the target to the head, then abort — the
+  // onFinish flush-on-abort branch sends exactly that head, tagged as an
+  // interrupt so the server notes the previous answer was cut off.
+  const sendNow = useCallback(
+    (id: string) => {
+      // Branch on the LIVE status (statusRef), NOT the closure-captured isStreaming:
+      // the turn may have finished between this render and the click, in which case
+      // stop() is a no-op and arming the interrupt refs would strand them for a
+      // later, unrelated Stop. Reading the ref always sees the current status.
+      const liveStreaming =
+        statusRef.current === "submitted" || statusRef.current === "streaming";
+      if (liveStreaming) {
+        // Promote to head so the onFinish -> flushNext path sends exactly it.
+        setQueue(promoteToHead(queuedRef.current, id));
+        flushOnAbortRef.current = true;
+        interruptNextSendRef.current = true;
+        stop(); // -> onFinish({ isAbort: true }) flushes the promoted head
+      } else {
+        // Nothing to interrupt: just send it now (no interrupt note).
+        const msg = queuedRef.current.find((m) => m.id === id);
+        if (!msg) return;
+        setQueue(removeQueuedById(queuedRef.current, id));
+        sendMessageRef.current?.({ text: msg.text });
+      }
+    },
+    [setQueue, stop],
+  );
+
+  // Clear the stopped marker as soon as a new turn begins streaming, and drop any
+  // stale "Send now" interrupt flags. On the legit interrupt path both refs are
+  // already consumed synchronously (onFinish + prepareSendMessagesRequest) before
+  // this effect runs, so clearing here is a no-op for it; its purpose is to defuse
+  // the race where a flag was armed but the expected abort never fired (the turn
+  // finished in the same tick as the click), so it cannot leak into a later turn.
   useEffect(() => {
-    if (isStreaming) setStopNotice(null);
+    if (isStreaming) {
+      setStopNotice(null);
+      flushOnAbortRef.current = false;
+      interruptNextSendRef.current = false;
+    }
   }, [isStreaming]);
 
   // Classify the turn error into a heading + detail so the banner names the cause
@@ -423,6 +512,17 @@ export default function ChatThread({
                 <Text size="xs" lineClamp={2} className={classes.queuedText}>
                   {m.text}
                 </Text>
+                <Tooltip label={t("Interrupt and send now")} withArrow>
+                  <ActionIcon
+                    size="xs"
+                    variant="subtle"
+                    color="blue"
+                    onClick={() => sendNow(m.id)}
+                    aria-label={t("Send now")}
+                  >
+                    <IconPlayerPlayFilled size={12} />
+                  </ActionIcon>
+                </Tooltip>
                 <ActionIcon
                   size="xs"
                   variant="subtle"

apps/client/src/features/ai-chat/utils/queue-helpers.test.ts

@@ -2,6 +2,7 @@ import { describe, it, expect } from "vitest";
 import {
   enqueueMessage,
   dequeue,
+  promoteToHead,
   removeQueuedById,
   type QueuedMessage,
 } from "./queue-helpers";
@@ -89,6 +90,52 @@ describe("removeQueuedById", () => {
   });
 });
 
+describe("promoteToHead", () => {
+  it("moves the matching id to the front, preserving the rest's order", () => {
+    const queue: QueuedMessage[] = [
+      { id: "a", text: "first" },
+      { id: "b", text: "second" },
+      { id: "c", text: "third" },
+    ];
+    expect(promoteToHead(queue, "c")).toEqual([
+      { id: "c", text: "third" },
+      { id: "a", text: "first" },
+      { id: "b", text: "second" },
+    ]);
+  });
+
+  it("is a no-op order-wise when the id is already the head", () => {
+    const queue: QueuedMessage[] = [
+      { id: "a", text: "first" },
+      { id: "b", text: "second" },
+    ];
+    expect(promoteToHead(queue, "a")).toEqual([
+      { id: "a", text: "first" },
+      { id: "b", text: "second" },
+    ]);
+  });
+
+  it("returns an equivalent list when the id is not present", () => {
+    const queue: QueuedMessage[] = [
+      { id: "a", text: "first" },
+      { id: "b", text: "second" },
+    ];
+    expect(promoteToHead(queue, "missing")).toEqual(queue);
+  });
+
+  it("does not mutate the input queue", () => {
+    const queue: QueuedMessage[] = [
+      { id: "a", text: "first" },
+      { id: "b", text: "second" },
+    ];
+    promoteToHead(queue, "b");
+    expect(queue).toEqual([
+      { id: "a", text: "first" },
+      { id: "b", text: "second" },
+    ]);
+  });
+});
+
 describe("FIFO order", () => {
   it("preserves order across enqueue -> dequeue", () => {
     let queue: QueuedMessage[] = [];

apps/client/src/features/ai-chat/utils/queue-helpers.ts

@@ -32,3 +32,16 @@ export function removeQueuedById(
 ): QueuedMessage[] {
   return queue.filter((m) => m.id !== id);
 }
+
+/** Move the queued message with the given id to the FRONT (returns a new array).
+ *  No-op (returns an equivalent array) when the id is absent. Pure — backs the
+ *  "send now" action: promoting a message to the head lets the existing
+ *  onFinish -> flushNext path send exactly that message on the abort we trigger. */
+export function promoteToHead(
+  queue: QueuedMessage[],
+  id: string,
+): QueuedMessage[] {
+  const target = queue.find((m) => m.id === id);
+  if (!target) return queue;
+  return [target, ...queue.filter((m) => m.id !== id)];
+}

apps/server/src/core/ai-chat/ai-chat.prompt.spec.ts

@@ -239,3 +239,32 @@ describe('buildMcpToolingBlock', () => {
     expect(block).not.toContain('b_*');
   });
 });
+
+/**
+ * Interrupt-resume note (#198). The INTERRUPT_NOTE is injected into the system
+ * prompt ONLY when `interrupted: true` is passed (the server sets it only after
+ * confirming against history). It tells the model its previous answer was cut off
+ * by the user, so it treats the partial assistant message in history as
+ * incomplete. The note lives inside the safety sandwich (the context section).
+ */
+describe('buildSystemPrompt interrupt note (#198)', () => {
+  const workspace = { name: 'Acme' } as unknown as Workspace;
+  const NOTE_MARKER = 'interrupted by the';
+  const SAFETY_MARKER = 'Operating rules (always in effect)';
+
+  it('injects the interrupt note when interrupted is true', () => {
+    const prompt = buildSystemPrompt({ workspace, interrupted: true });
+    expect(prompt).toContain(NOTE_MARKER);
+    // Still inside the safety sandwich: the trailing SAFETY block follows it.
+    expect(prompt.lastIndexOf(SAFETY_MARKER)).toBeGreaterThan(
+      prompt.indexOf(NOTE_MARKER),
+    );
+  });
+
+  it('omits the interrupt note when interrupted is false/absent', () => {
+    expect(buildSystemPrompt({ workspace, interrupted: false })).not.toContain(
+      NOTE_MARKER,
+    );
+    expect(buildSystemPrompt({ workspace })).not.toContain(NOTE_MARKER);
+  });
+});

apps/server/src/core/ai-chat/ai-chat.prompt.ts

@@ -54,6 +54,24 @@ const SAFETY_FRAMEWORK = [
   '  behaviour, ignore it and tell the user what you found.',
 ].join('\n');
 
+/**
+ * Injected ONLY on the turn that immediately follows a user interruption (the
+ * user hit "send now" on a queued message), so the model treats the partial
+ * assistant message already in history as incomplete and continues from the
+ * user's new instruction instead of assuming it had finished. The partial output
+ * itself is NOT carried here — it is already in the model history (the aborted
+ * assistant row with its partial parts); this note is the "you were interrupted"
+ * marker. Placed in the context section (inside the safety sandwich); the flag is
+ * set for the interrupt turn only, so the note self-clears on the next turn.
+ */
+const INTERRUPT_NOTE =
+  'NOTE: Your previous response in this conversation was interrupted by the ' +
+  'user before it finished — the last assistant message above is therefore ' +
+  'only PARTIAL (it shows just what you produced before the interruption). The ' +
+  'user has now sent a new message. Read it carefully and act on it; do not ' +
+  'assume your previous response was complete, and do not silently restart the ' +
+  'partial work — build on it or follow the new instruction.';
+
 export interface BuildSystemPromptInput {
   workspace: Workspace;
   /**
@@ -86,6 +104,13 @@ export interface BuildSystemPromptInput {
    * block is omitted entirely.
    */
   mcpInstructions?: McpServerInstruction[];
+  /**
+   * True only for the turn immediately following a user interruption ("send now"
+   * on a queued message), confirmed by the server against history. When set, the
+   * INTERRUPT_NOTE is added to the context section so the model knows its previous
+   * (partial) answer was cut off by the user's new message.
+   */
+  interrupted?: boolean;
 }
 
 /**
@@ -130,6 +155,7 @@ export function buildSystemPrompt({
   roleInstructions,
   openedPage,
   mcpInstructions,
+  interrupted,
 }: BuildSystemPromptInput): string {
   // Persona precedence: role instructions REPLACE the admin persona / default.
   // effectivePersona = roleInstructions || adminPrompt || DEFAULT_PROMPT.
@@ -157,6 +183,14 @@ export function buildSystemPrompt({
     context += `\nThe user is currently viewing the page "${title}" (pageId: ${pageId.trim()}). When they refer to "this page", "the current page", or similar, operate on that pageId — use the read/write page tools with it.`;
   }
 
+  // Interrupt-resume marker (#198). Added to the context section (inside the
+  // safety sandwich), present only for the turn that directly follows a user
+  // interruption — the server confirms the flag against history before passing it
+  // here, so a spoofed flag on an ordinary turn never injects this note.
+  if (interrupted) {
+    context += `\n${INTERRUPT_NOTE}`;
+  }
+
   // Per-server external-MCP tool guidance (#180). Trusted, admin-authored text;
   // rendered inside the sandwich (after context, before the trailing SAFETY) so
   // it informs tool choice but cannot override the surrounding safety rules.

apps/server/src/core/ai-chat/ai-chat.service.spec.ts

@@ -9,6 +9,7 @@ import {
   flushAssistant,
   chatStreamMetadata,
   accumulateStepUsage,
+  isInterruptResume,
   MAX_AGENT_STEPS,
   FINAL_STEP_INSTRUCTION,
 } from './ai-chat.service';
@@ -240,7 +241,7 @@ describe('prepareAgentStep', () => {
  * write path. It runs identically for the upfront insert (empty steps,
  * 'streaming'), every per-step update, and the terminal finalize — so a future
  * background worker can call the same function. These tests pin the four status
- * shapes and the `metadata.parts` shape that rowToUiMessage/findAllByChat depend on
+ * shapes and the `metadata.parts` shape that rowToUiMessage/findRecent depend on
  * (per-step text + tool parts via assistantParts, in-progress text appended).
  */
 describe('flushAssistant', () => {
@@ -649,3 +650,57 @@ describe('AiChatService.resolveOpenPageContext (#159 current-page validation)',
     expect(await call(svc, { id: 'p-1' })).toEqual({ id: 'p-1', title: '' });
   });
 });
+
+/**
+ * isInterruptResume (#198): the pure guard that decides whether the interrupt
+ * note is injected for a turn. The client "send now" flag is only a hint; it is
+ * honoured ONLY when the preceding assistant turn (history[len-2], since the new
+ * user row is the tail) really ended unfinished ('aborted', or still 'streaming'
+ * during the abort/resend race). A spoofed flag on an ordinary turn is ignored.
+ */
+describe('isInterruptResume', () => {
+  // history tail is the just-inserted user row; [len-2] is the previous turn.
+  const withPrev = (
+    prev: { role: string; status?: string | null } | null,
+  ): Array<{ role: string; status?: string | null }> =>
+    prev
+      ? [prev, { role: 'user', status: null }]
+      : [{ role: 'user', status: null }];
+
+  it('false when the client flag is not set', () => {
+    expect(
+      isInterruptResume(withPrev({ role: 'assistant', status: 'aborted' }), undefined),
+    ).toBe(false);
+    expect(
+      isInterruptResume(withPrev({ role: 'assistant', status: 'aborted' }), false),
+    ).toBe(false);
+  });
+
+  it('true when flagged AND the previous assistant turn is aborted', () => {
+    expect(
+      isInterruptResume(withPrev({ role: 'assistant', status: 'aborted' }), true),
+    ).toBe(true);
+  });
+
+  it('true when flagged AND the previous assistant turn is still streaming (race)', () => {
+    expect(
+      isInterruptResume(withPrev({ role: 'assistant', status: 'streaming' }), true),
+    ).toBe(true);
+  });
+
+  it('false when flagged but the previous assistant turn completed normally', () => {
+    expect(
+      isInterruptResume(withPrev({ role: 'assistant', status: 'completed' }), true),
+    ).toBe(false);
+  });
+
+  it('false when flagged but the previous turn is not an assistant turn', () => {
+    expect(
+      isInterruptResume(withPrev({ role: 'user', status: 'aborted' }), true),
+    ).toBe(false);
+  });
+
+  it('false when there is no preceding turn (only the new user row)', () => {
+    expect(isInterruptResume(withPrev(null), true)).toBe(false);
+  });
+});

apps/server/src/core/ai-chat/ai-chat.service.ts

@@ -75,6 +75,32 @@ export function prepareAgentStep(
 
 export { MAX_AGENT_STEPS, FINAL_STEP_INSTRUCTION };
 
+/**
+ * Pure, unit-testable (#198): decide whether THIS turn is an interrupt-resume,
+ * i.e. it directly follows a user interruption of the previous (still-partial)
+ * assistant turn. The client "send now" flag is only a HINT — confirm it against
+ * the just-loaded history so a spoofed/stale flag cannot inject the interrupt
+ * note onto an ordinary turn.
+ *
+ * `history` is the model history oldest -> newest, with the just-inserted user
+ * row as its tail; the turn before it is `history[len-2]`. We treat the new turn
+ * as an interrupt-resume only when the client said so AND the preceding assistant
+ * turn really ended unfinished: 'aborted' (onAbort already finalized it), or
+ * still 'streaming' (onAbort has not finalized yet — the abort/resend race; the
+ * partial output is already in history thanks to the step-granular write path).
+ */
+export function isInterruptResume(
+  history: Array<{ role: string; status?: string | null }>,
+  clientInterrupted: boolean | undefined,
+): boolean {
+  if (clientInterrupted !== true) return false;
+  const prev = history[history.length - 2];
+  return (
+    prev?.role === 'assistant' &&
+    (prev.status === 'aborted' || prev.status === 'streaming')
+  );
+}
+
 /**
  * Payload accepted from the client `useChat` POST body. We do NOT bind a strict
  * DTO (the global ValidationPipe whitelist would strip the useChat-specific
@@ -93,6 +119,11 @@ export interface AiChatStreamBody {
   // is attacker-controllable but harmless: the agent reads/writes via its
   // CASL-enforced page tools, which 403 on a page the user cannot access.
   openPage?: { id?: string; title?: string } | null;
+  // Set by the client "send now" action (#198): this turn immediately follows a
+  // user interruption of the previous turn. A hint only — the server re-confirms
+  // it against persisted history (`isInterruptResume`) before injecting the
+  // interrupt note, so a spoofed/stale flag on an ordinary turn is ignored.
+  interrupted?: boolean;
   // useChat sends the full UIMessage list; the last one is the new user turn.
   messages?: UIMessage[];
 }
@@ -322,19 +353,24 @@ export class AiChatService implements OnModuleInit {
 
     // Rebuild the conversation from persisted history (not the client payload),
     // so the model always sees the authoritative server-side transcript. Load
-    // the FULL history in chronological order (oldest -> newest, incl. the user
-    // message just inserted above) so NO turns are dropped — there is no
-    // recent-tail window anymore. `findAllByChat` keeps a 5000-row memory-safety
-    // backstop (on overflow it keeps the NEWEST rows and logs a warning); that
-    // is a safety net far above any realistic chat, not a conversational limit.
-    const history = await this.aiChatMessageRepo.findAllByChat(
+    // the most RECENT tail (oldest -> newest) so chats longer than one page do
+    // not drop recent turns (incl. the user message just inserted above).
+    const history = await this.aiChatMessageRepo.findRecent(
       chatId,
       workspace.id,
+      50,
     );
     const uiMessages = history.map(rowToUiMessage);
     // convertToModelMessages is async in ai@6.0.134 (returns Promise<ModelMessage[]>).
     const messages = await convertToModelMessages(uiMessages);
 
+    // Interrupt-resume detection (#198): the client "send now" flag is only a
+    // hint — confirm it against the persisted history (the preceding assistant
+    // turn must really be aborted/streaming) so a spoofed flag cannot inject the
+    // interrupt note onto an ordinary turn. The partial output the model needs is
+    // already in `messages` (the aborted assistant row replays via findRecent).
+    const interrupted = isInterruptResume(history, body.interrupted);
+
     // The model is resolved by the controller before hijack (clean 503 path).
     // Here we only need the admin-configured system prompt.
     const resolved = await this.aiSettings.resolve(workspace.id);
@@ -406,6 +442,9 @@ export class AiChatService implements OnModuleInit {
         openedPage: openPageContext,
         // Guidance only for servers that connected and yielded ≥1 callable tool.
         mcpInstructions: external.instructions,
+        // History-confirmed interrupt-resume flag (#198): adds the interrupt note
+        // so the model treats the partial answer above as cut off, not finished.
+        interrupted,
       });
 
       // Pass the resolved chatId so the write tools can mint provenance tokens
@@ -1217,7 +1256,7 @@ export async function applyFinalize(
  *
  * `metadata.parts` is built by assistantParts over the finished steps, then the
  * in-progress text appended as a trailing text part, so rowToUiMessage /
- * findAllByChat keep replaying the turn unchanged. `metadata.finishReason`,
+ * findRecent keep replaying the turn unchanged. `metadata.finishReason`,
  * `metadata.error`, `metadata.usage`, `metadata.contextTokens` and
  * `metadata.maxContextTokens` are attached only when provided/relevant, matching
  * the pre-#183 onFinish/onError records.

apps/server/src/database/repos/ai-chat/ai-chat-message.repo.ts

@@ -18,8 +18,7 @@ import { executeWithCursorPagination } from '@docmost/db/pagination/cursor-pagin
 // (multi-instance deploy).
 const SWEEP_STREAMING_STALE_MS = 10 * 60 * 1000; // 10 minutes
 
-// Hard upper bound on the rows materialized by `findAllByChat`, which now feeds
-// BOTH the Markdown export and the per-turn model history.
+// Hard upper bound on the rows materialized by `findAllByChat` (export path).
 // A generous cap so a pathologically huge chat cannot load an unbounded result
 // into memory; far above any realistic transcript length.
 const FIND_ALL_BY_CHAT_LIMIT = 5000;
@@ -79,17 +78,14 @@ export class AiChatMessageRepo {
   }
 
   // Load ALL (non-deleted) messages of a chat in ascending chronological order
-  // (oldest -> newest), unpaginated. Two callers, both treating the DB as the
-  // single source of truth and needing the whole transcript in one pass
-  // (findByChat is cursor-paginated and would only return the first page):
-  //   - the server-side Markdown export (#183);
-  //   - the per-turn model history, rebuilt fresh on every turn so the model
-  //     sees the full authoritative transcript.
+  // (oldest -> newest), unpaginated. Used by the server-side Markdown export
+  // (#183), where the DB is the single source of truth and the whole transcript
+  // must be rendered in one pass (findByChat is cursor-paginated and would only
+  // return the first page).
   //
   // Hard-capped at FIND_ALL_BY_CHAT_LIMIT rows (a generous bound, far above any
-  // realistic transcript) — a shared memory-safety backstop for BOTH paths so a
-  // pathologically huge chat cannot materialize an unbounded result set in
-  // memory. On overflow the NEWEST rows are kept and a warning is logged.
+  // realistic transcript) so exporting a pathologically huge chat cannot
+  // materialize an unbounded result set in memory.
   async findAllByChat(
     chatId: string,
     workspaceId: string,
@@ -97,9 +93,9 @@ export class AiChatMessageRepo {
     limit: number = FIND_ALL_BY_CHAT_LIMIT,
   ): Promise<AiChatMessage[]> {
     // Fetch newest-first (+1 to DETECT truncation), so on overflow we keep the
-    // NEWEST `limit` messages — the recent conversation matters most — rather
-    // than silently dropping the tail (#183 review). Then reverse back to
-    // chronological order (oldest -> newest) for rendering / model replay.
+    // NEWEST `limit` messages — the recent conversation matters most for an
+    // export — rather than silently dropping the tail (#183 review). Reverse back
+    // to chronological for rendering, like findRecent.
     const rows = await this.db
       .selectFrom('aiChatMessages')
       .select(this.baseFields)
@@ -114,13 +110,38 @@ export class AiChatMessageRepo {
     if (rows.length > limit) {
       rows.length = limit; // keep the newest `limit` (rows are newest-first here)
       this.logger.warn(
-        `Chat ${chatId} truncated to the newest ${limit} messages ` +
+        `Chat ${chatId} export truncated to the newest ${limit} messages ` +
           `(older messages omitted).`,
       );
     }
     return rows.reverse();
   }
 
+  // Load the most RECENT `limit` messages for a chat and return them in
+  // ascending chronological order (oldest -> newest), as the model expects.
+  // `findByChat` returns the FIRST page ASC (the OLDEST messages), which loses
+  // recent turns once a chat grows beyond a page; this rebuilds the model
+  // history from the tail instead. Plain query (no cursor pagination).
+  async findRecent(
+    chatId: string,
+    workspaceId: string,
+    limit: number,
+  ): Promise<AiChatMessage[]> {
+    const rows = await this.db
+      .selectFrom('aiChatMessages')
+      .select(this.baseFields)
+      .where('chatId', '=', chatId)
+      .where('workspaceId', '=', workspaceId)
+      .where('deletedAt', 'is', null)
+      .orderBy('createdAt', 'desc')
+      .orderBy('id', 'desc')
+      .limit(limit)
+      .execute();
+
+    // Selected newest-first for the limit; reverse to oldest-first for the model.
+    return rows.reverse();
+  }
+
   async insert(
     insertable: InsertableAiChatMessage,
     trx?: KyselyTransaction,

apps/server/test/integration/ai-chat-message-status.int-spec.ts

@@ -267,36 +267,4 @@ describe('AiChatMessageRepo.update + sweepStreaming [integration]', () => {
     const all = await repo.findAllByChat(cappedChat, workspaceId, 100);
     expect(all.map((r) => r.content)).toEqual(['m1-oldest', 'm2', 'm3-newest']);
   });
-
-  it('default findAllByChat returns the FULL transcript past 50 rows — no recent-tail window (#202)', async () => {
-    // PR #202 swapped the model-history rebuild in AiChatService.handle from
-    // findRecent(chatId, ws, 50) to findAllByChat(chatId, ws) WITHOUT a limit
-    // arg. This pins the behavioral guarantee that switch relies on: a chat
-    // longer than the old 50-msg window comes back in FULL (oldest -> newest),
-    // so no early turns are silently dropped from what the model sees. The old
-    // 50-cap would have returned only the last 50 of these 60 rows.
-    const longChat = (
-      await createChat(db, { workspaceId, creatorId: userId })
-    ).id;
-    const base = Date.now();
-    const total = 60;
-    for (let i = 0; i < total; i++) {
-      await createMessage(db, {
-        workspaceId,
-        chatId: longChat,
-        content: `msg-${i}`,
-        // Strictly increasing timestamps so ordering is deterministic.
-        createdAt: new Date(base + i * 1000),
-      });
-    }
-
-    // Default args == exactly how handle() calls it now.
-    const history = await repo.findAllByChat(longChat, workspaceId);
-    expect(history).toHaveLength(total);
-    expect(history.map((r) => r.content)).toEqual(
-      Array.from({ length: total }, (_, i) => `msg-${i}`),
-    );
-    // The very first turn (which the old 50-window would have dropped) is present.
-    expect(history[0]!.content).toBe('msg-0');
-  });
 });