core/cli: level-1 cancel of in-flight turns via Ctrl+C

Run turns as asyncio tasks; SIGINT cancels the task; agent rolls back
and keeps pending_retry_text for /retry.

CLAUDE.md

@@ -74,7 +74,7 @@ Click app + readline REPL. Entry: `plyngent` / `python -m plyngent`.
 
 - **`plyngent chat`**: provider/model selection (flags or interactive), SQLite sessions via config `[database]` (file DB under user data if unset/`:memory:`), resumes latest session by default (`--new` / `--session`).
 - Slash: `/history`, `/sessions`, `/resume`, `/rounds`, `/retry`, …
-- Failed turns: not written to DB; auto-retry 10s/20s/30s (Ctrl+C cancels wait); `/retry` manual.
+- Failed/cancelled turns: not written to DB; Ctrl+C cancels the in-flight turn task; auto-retry 10s/20s/30s (Ctrl+C cancels wait); `/retry` manual.
 - **`plyngent providers`**: list config providers.
 - **`plyngent config path|edit`**: print or open config in `$EDITOR` (`shlex`-split, e.g. `codium --wait`).
 - If no providers and `$EDITOR` is set, chat/providers prompt to edit config then reload.

src/plyngent/agent/chat.py

@@ -78,6 +78,10 @@ class ChatAgent:
         if self.memory is not None and self.session_id is not None:
             _ = await self.memory.append_message(self.session_id, message)
 
+    def _rollback_turn(self, pre_len: int, user_text: str) -> None:
+        del self.messages[pre_len:]
+        self.pending_retry_text = user_text
+
     async def _run_from_user_message(self, user_msg: UserChatMessage) -> AsyncIterator[AgentEvent]:
         """Run the tool loop for an already-appended user message; persist only on success."""
         pre_len = len(self.messages) - 1
@@ -85,6 +89,7 @@ class ChatAgent:
             pre_len = len(self.messages)
             self.messages.append(user_msg)
 
+        completed = False
         try:
             async for event in run_chat_loop(
                 self.client,
@@ -96,10 +101,11 @@ class ChatAgent:
                 on_limit=self.on_limit,
             ):
                 yield event
-        except Exception:
-            # Roll back the whole turn so a failed user message is not left half-applied.
-            del self.messages[pre_len:]
-            self.pending_retry_text = user_msg.content
+            completed = True
+        except BaseException:
+            # Includes CancelledError / KeyboardInterrupt paths via task cancel.
+            if not completed:
+                self._rollback_turn(pre_len, user_msg.content)
             raise
 
         for message in self.messages[pre_len:]:

src/plyngent/cli/repl.py

@@ -37,8 +37,9 @@ Commands:
   /rounds [n]        Show or set max tool-loop rounds
   /retry             Retry the last failed user turn (after errors)
 
-On network/API errors, the user turn is not saved to the DB. Auto-retry
-waits 10s, 20s, then 30s (Ctrl+C cancels waits; use /retry later).
+On network/API errors or Ctrl+C during a turn, the turn is not saved to
+the DB. Auto-retry waits 10s, 20s, then 30s (Ctrl+C cancels waits or the
+in-flight turn; use /retry later).
 
 Tab completes slash commands and some arguments (provider, model, tools).
 Use --session ID or /resume to continue a prior chat after restart.

src/plyngent/cli/retry.py

@@ -1,6 +1,8 @@
 from __future__ import annotations
 
 import asyncio
+import contextlib
+import signal
 import time
 from typing import TYPE_CHECKING
 
@@ -9,7 +11,7 @@ import click
 from plyngent.cli.display import render_events
 
 if TYPE_CHECKING:
-    from collections.abc import AsyncIterator, Callable
+    from collections.abc import AsyncIterator, Callable, Coroutine
 
     from plyngent.agent import AgentEvent
     from plyngent.agent.chat import ChatAgent
@@ -34,6 +36,43 @@ async def sleep_cancellable(seconds: float) -> bool:
         return False
 
 
+async def run_cancellable[T](coro: Coroutine[object, object, T]) -> T:
+    """Await ``coro`` as a task; Ctrl+C / SIGINT cancels the task.
+
+    Raises:
+        asyncio.CancelledError: If the task was cancelled (including via SIGINT).
+    """
+    task = asyncio.create_task(coro)
+    loop = asyncio.get_running_loop()
+    installed = False
+
+    def _on_sigint() -> None:
+        if not task.done():
+            _ = task.cancel()
+
+    try:
+        loop.add_signal_handler(signal.SIGINT, _on_sigint)
+        installed = True
+    except (NotImplementedError, RuntimeError, ValueError):
+        installed = False
+
+    try:
+        return await task
+    except KeyboardInterrupt:
+        if not task.done():
+            _ = task.cancel()
+            with contextlib.suppress(asyncio.CancelledError):
+                await task
+        raise asyncio.CancelledError from None
+    finally:
+        if installed:
+            _ = loop.remove_signal_handler(signal.SIGINT)
+        if not task.done():
+            _ = task.cancel()
+            with contextlib.suppress(asyncio.CancelledError):
+                await task
+
+
 async def run_turn_with_retries(
     agent: ChatAgent,
     *,
@@ -44,15 +83,27 @@ async def run_turn_with_retries(
 
     ``starter`` produces the event stream for the current attempt (``agent.run``
     or ``agent.retry``). Returns True if the turn completed successfully.
+
+    Ctrl+C during a model/tool turn cancels the in-flight task (level-1 cancel).
+    The agent rolls back the turn and keeps ``pending_retry_text`` for ``/retry``.
     """
     max_retries = len(delays)
     attempt = 0
     while True:
         try:
-            await render_events(starter())
+            await run_cancellable(render_events(starter()))
+        except asyncio.CancelledError:
+            click.echo()
+            click.secho(
+                "cancelled (turn not saved); use /retry to try again",
+                fg="yellow",
+            )
+            click.echo()
+            return False
         except KeyboardInterrupt:
             click.echo()
             click.secho("interrupted", fg="yellow")
+            click.echo()
             return False
         except Exception as exc:  # noqa: BLE001 — surface and optionally retry
             click.secho(f"error: {exc}", fg="red")

tests/test_cli/test_retry.py

@@ -2,6 +2,8 @@ from __future__ import annotations
 
 from typing import TYPE_CHECKING, Literal, overload
 
+import pytest
+
 from plyngent.agent import ChatAgent
 from plyngent.cli.retry import retry_pending_with_retries, run_turn_with_retries, sleep_cancellable
 from plyngent.config.models import DatabaseConfig
@@ -18,8 +20,6 @@ from plyngent.memory import MemoryStore
 if TYPE_CHECKING:
     from collections.abc import AsyncIterator
 
-    import pytest
-
 
 def _response(message: AssistantChatMessage) -> ChatCompletionResponse:
     return ChatCompletionResponse(
@@ -105,6 +105,64 @@ async def test_auto_retry_eventually_succeeds(monkeypatch: pytest.MonkeyPatch) -
     await store.close()
 
 
+async def test_run_cancellable_cancels_task() -> None:
+    import asyncio
+
+    from plyngent.cli.retry import run_cancellable
+
+    started = asyncio.Event()
+
+    async def hang() -> None:
+        started.set()
+        await asyncio.sleep(60)
+
+    task = asyncio.create_task(run_cancellable(hang()))
+    _ = await started.wait()
+    _ = task.cancel()
+    with pytest.raises(asyncio.CancelledError):
+        await task
+
+
+async def test_cancel_turn_rolls_back_and_sets_pending() -> None:
+    import asyncio
+
+    from plyngent.cli.display import render_events
+    from plyngent.cli.retry import run_cancellable
+
+    store = await MemoryStore.open(DatabaseConfig())
+    session = await store.create_session(name="t")
+
+    class HangClient:
+        @overload
+        async def chat_completions(
+            self, param: ChatCompletionsParam, *, stream: Literal[False] = False
+        ) -> ChatCompletionResponse: ...
+
+        @overload
+        async def chat_completions(
+            self, param: ChatCompletionsParam, *, stream: Literal[True]
+        ) -> AsyncIterator[ChatCompletionChunk]: ...
+
+        async def chat_completions(
+            self, param: ChatCompletionsParam, *, stream: bool = False
+        ) -> ChatCompletionResponse | AsyncIterator[ChatCompletionChunk]:
+            del param, stream
+            await asyncio.sleep(60)
+            return _response(AssistantChatMessage(content="never"))
+
+    agent = ChatAgent(HangClient(), model="m", memory=store, session_id=session.sid)
+    turn = asyncio.create_task(run_cancellable(render_events(agent.run("cancel-me"))))
+    await asyncio.sleep(0.05)
+    _ = turn.cancel()
+    with pytest.raises(asyncio.CancelledError):
+        await turn
+
+    assert agent.pending_retry_text == "cancel-me"
+    assert agent.messages == []
+    assert await store.list_messages(session.sid) == []
+    await store.close()
+
+
 async def test_manual_retry_after_exhausted(monkeypatch: pytest.MonkeyPatch) -> None:
     async def no_wait(_seconds: float) -> bool:
         return True