Files
hermes-agent/agent/interrupt_scope.py
Teknium ac087e6ada fix(console): cancel/timeout interrupts the command's agent and waits for the worker
Hermes Console ran each command on a ThreadPoolExecutor and cancelled only
the asyncio waiter. A command that forks an AIAgent (`curator run
--consolidate`) kept its worker thread and its in-flight provider request
alive after the prompt said "cancelled" — a llama.cpp generation kept
decoding for 30+ minutes and held the inference slot (#106179).

Root cause: the host owning the thread never knew about the agent created
deep inside the synchronous command, so it could not call the existing
cooperative `interrupt()` path that closes the request sockets.

Fix: `agent/interrupt_scope.py` gives the host an `InterruptScope`; the
console binds it around the worker (ContextVar), and every
`AIAgent.run_conversation()` registers itself with the bound scope for the
turn. On cancel, timeout and disconnect the console calls `scope.cancel()`
(hard-interrupts every registered agent; an agent registering after the
cancel is interrupted on entry so the cancel cannot lose the race with a
turn that has not started) and awaits the worker Future with a 10s bound
before reporting cancelled/timeout. Queued-but-unstarted work is dropped via
`Future.cancel()` alone.

Live repro (fake OpenAI-compatible provider blocking like llama.cpp, real
/api/console, real curator dispatch, real AIAgent + direct request path):
origin/main at the "cancelled" frame -> request_exited=false,
worker_exited=false; with this change -> both true, provider saw the peer
close, prompt reported cancelled 0.27s after the frame.

Salvage of #106197 by @kyssta-exe (executor-future handle, cancel/timeout/
disconnect propagation) and #106320 by @Xixiartemis (deterministic
lifecycle regression: terminal(cancelled) => no owned request or worker
remains live; interrupt-on-late-registration). Both rebuilt slimmer: #106197 keyed its fallback on Future.cancel()
returning False, but the handle it held was run_in_executor's asyncio
wrapper, whose cancel() returns True while the thread keeps running, so its
thread-name abort registry was never consulted; #106320's command-scoped
ownership model is folded into one small module hooked at the turn facade
instead of a per-caller `bind_agent`.

Co-authored-by: kyssta-exe <218078013+kyssta-exe@users.noreply.github.com>
Co-authored-by: Xixiartemis <182932319+Xixiartemis@users.noreply.github.com>
2026-09-09 10:58:16 -07:00

66 lines
2.4 KiB
Python

"""Host-owned cancellation for agents created deep inside synchronous work.
A host that runs a blocking command on a worker thread (Hermes Console) never sees
the ``AIAgent`` a CLI subcommand forks inside it, so it cannot call ``interrupt()``
when the user cancels. The host binds an :class:`InterruptScope` around the work;
every ``run_conversation()`` under that scope registers its agent, and
``scope.cancel()`` hard-interrupts them from any thread. Agents registering after
the cancel are interrupted immediately, so a cancel never loses the race with a
turn that has not started yet (#106179).
"""
from __future__ import annotations
import threading
from contextlib import contextmanager, nullcontext
from contextvars import ContextVar
from typing import Any, Iterator, Optional
from agent.interrupt_compat import request_hard_interrupt
_ACTIVE_SCOPE: ContextVar[Optional["InterruptScope"]] = ContextVar("hermes_interrupt_scope", default=None)
class InterruptScope:
def __init__(self) -> None:
self._lock = threading.Lock()
self._agents: list[Any] = []
self.reason: Optional[str] = None
def cancel(self, reason: str) -> None:
"""Latch ``reason`` and hard-interrupt every agent running under this scope."""
with self._lock:
self.reason = reason
agents = list(self._agents)
for agent in agents:
request_hard_interrupt(agent, reason, tool_reason="host cancelled the command")
@contextmanager
def track(self, agent: Any) -> Iterator[None]:
with self._lock:
self._agents.append(agent)
reason = self.reason
if reason is not None:
request_hard_interrupt(agent, reason, tool_reason="host cancelled the command")
try:
yield
finally:
with self._lock:
self._agents.remove(agent)
@contextmanager
def bind_interrupt_scope(scope: Optional[InterruptScope]) -> Iterator[None]:
"""Make ``scope`` the owner of every agent turn started in this context."""
token = _ACTIVE_SCOPE.set(scope)
try:
yield
finally:
_ACTIVE_SCOPE.reset(token)
def track_in_interrupt_scope(agent: Any):
"""Register ``agent`` with the bound scope for the duration of its turn (no-op without one)."""
scope = _ACTIVE_SCOPE.get()
return nullcontext() if scope is None else scope.track(agent)