hermes-agent/tests/run_agent
Teknium 20f2258f34
fix(interrupt): propagate to concurrent-tool workers + opt-in debug trace (#11907)
* fix(interrupt): propagate to concurrent-tool workers + opt-in debug trace

interrupt() previously only flagged the agent's _execution_thread_id.
Tools running inside _execute_tool_calls_concurrent execute on
ThreadPoolExecutor worker threads whose tids are distinct from the
agent's, so is_interrupted() inside those tools returned False no matter
how many times the gateway called .interrupt() — hung ssh / curl / long
make-builds ran to their own timeout.

Changes:
- run_agent.py: track concurrent-tool worker tids in a per-agent set,
  fan interrupt()/clear_interrupt() out to them, and handle the
  register-after-interrupt race at _run_tool entry.  getattr fallback
  for the tracker so test stubs built via object.__new__ keep working.
- tools/environments/base.py: opt-in _wait_for_process trace (ENTER,
  per-30s HEARTBEAT with interrupt+activity-cb state, INTERRUPT
  DETECTED, TIMEOUT, EXIT) behind HERMES_DEBUG_INTERRUPT=1.
- tools/interrupt.py: opt-in set_interrupt() trace (caller tid, target
  tid, set snapshot) behind the same env flag.
- tests: new regression test runs a polling tool on a concurrent worker
  and asserts is_interrupted() flips to True within ~1s of interrupt().
  Second new test guards clear_interrupt() clearing tracked worker bits.

Validation: tests/run_agent/ all 762 pass; tests/tools/ interrupt+env
subset 216 pass.

* fix(interrupt-debug): bypass quiet_mode logger filter so trace reaches agent.log

AIAgent.__init__ sets logging.getLogger('tools').setLevel(ERROR) when
quiet_mode=True (the CLI default). This would silently swallow every
INFO-level trace line from the HERMES_DEBUG_INTERRUPT=1 instrumentation
added in the parent commit — confirmed by running hermes chat -q with
the flag and finding zero trace lines in agent.log even though
_wait_for_process was clearly executing (subprocess pid existed).

Fix: when HERMES_DEBUG_INTERRUPT=1, each traced module explicitly sets
its own logger level to INFO at import time, overriding the 'tools'
parent-level filter. Scoped to the opt-in case only, so production
(quiet_mode default) logs stay quiet as designed.

Validation: hermes chat -q with HERMES_DEBUG_INTERRUPT=1 now writes
'_wait_for_process ENTER/EXIT' lines to agent.log as expected.

* fix(cli): SIGTERM/SIGHUP no longer orphans tool subprocesses

Tool subprocesses spawned by the local environment backend use
os.setsid so they run in their own process group. Before this fix,
SIGTERM/SIGHUP to the hermes CLI killed the main thread via
KeyboardInterrupt but the worker thread running _wait_for_process
never got a chance to call _kill_process — Python exited, the child
was reparented to init (PPID=1), and the subprocess ran to its
natural end (confirmed live: sleep 300 survived 4+ min after SIGTERM
to the agent until manual cleanup).

Changes:
- cli.py _signal_handler (interactive) + _signal_handler_q (-q mode):
  route SIGTERM/SIGHUP through agent.interrupt() so the worker's poll
  loop sees the per-thread interrupt flag and calls _kill_process
  (os.killpg) on the subprocess group. HERMES_SIGTERM_GRACE (default
  1.5s) gives the worker time to complete its SIGTERM+SIGKILL
  escalation before KeyboardInterrupt unwinds main.
- tools/environments/base.py _wait_for_process: wrap the poll loop in
  try/except (KeyboardInterrupt, SystemExit) so the cleanup fires
  even on paths the signal handlers don't cover (direct sys.exit,
  unhandled KI from nested code, etc.). Emits EXCEPTION_EXIT trace
  line when HERMES_DEBUG_INTERRUPT=1.
- New regression test: injects KeyboardInterrupt into a running
  _wait_for_process via PyThreadState_SetAsyncExc, verifies the
  subprocess process group is dead within 3s of the exception and
  that KeyboardInterrupt re-raises cleanly afterward.

Validation:
| Before                                                  | After              |
|---------------------------------------------------------|--------------------|
| sleep 300 survives 4+ min as PPID=1 orphan after SIGTERM | dies within 2 s   |
| No INTERRUPT DETECTED in trace                          | INTERRUPT DETECTED fires + killing process group |
| tests/tools/test_local_interrupt_cleanup                | 1/1 pass          |
| tests/run_agent/test_concurrent_interrupt               | 4/4 pass          |
2026-04-17 20:39:25 -07:00
..
__init__.py refactor(tests): re-architect tests + fix CI failures (#5946) 2026-04-07 17:19:07 -07:00
conftest.py test: speed up slow tests (backoff + subprocess + IMDS network) (#11797) 2026-04-17 14:21:22 -07:00
test_413_compression.py test: speed up slow tests (backoff + subprocess + IMDS network) (#11797) 2026-04-17 14:21:22 -07:00
test_860_dedup.py fix(tests): make AIAgent constructor calls self-contained (#11755) 2026-04-17 12:32:03 -07:00
test_1630_context_overflow_loop.py fix(tests): make AIAgent constructor calls self-contained (#11755) 2026-04-17 12:32:03 -07:00
test_agent_guardrails.py fix(delegate): make max_concurrent_children configurable + error on excess 2026-04-10 13:38:14 -07:00
test_agent_loop.py refactor(tests): re-architect tests + fix CI failures (#5946) 2026-04-07 17:19:07 -07:00
test_agent_loop_tool_calling.py refactor(tests): re-architect tests + fix CI failures (#5946) 2026-04-07 17:19:07 -07:00
test_agent_loop_vllm.py refactor(tests): re-architect tests + fix CI failures (#5946) 2026-04-07 17:19:07 -07:00
test_anthropic_error_handling.py test: speed up slow tests (backoff + subprocess + IMDS network) (#11797) 2026-04-17 14:21:22 -07:00
test_async_httpx_del_neuter.py fix: bound auxiliary client cache to prevent fd exhaustion in long-running gateways (#10200) (#10470) 2026-04-15 13:16:28 -07:00
test_compression_boundary.py refactor(tests): re-architect tests + fix CI failures (#5946) 2026-04-07 17:19:07 -07:00
test_compression_feasibility.py test: add tests for compression config_context_length passthrough 2026-04-12 17:52:34 -07:00
test_compression_persistence.py fix(tests): make AIAgent constructor calls self-contained (#11755) 2026-04-17 12:32:03 -07:00
test_compressor_fallback_update.py refactor(tests): re-architect tests + fix CI failures (#5946) 2026-04-07 17:19:07 -07:00
test_concurrent_interrupt.py fix(interrupt): propagate to concurrent-tool workers + opt-in debug trace (#11907) 2026-04-17 20:39:25 -07:00
test_context_token_tracking.py fix(tests): resolve 12 CI failures + 10 errors across 6 root causes (#11040) 2026-04-16 06:49:36 -07:00
test_create_openai_client_kwargs_isolation.py fix(tests): make AIAgent constructor calls self-contained (#11755) 2026-04-17 12:32:03 -07:00
test_create_openai_client_reuse.py fix(tests): make AIAgent constructor calls self-contained (#11755) 2026-04-17 12:32:03 -07:00
test_dict_tool_call_args.py fix(tests): fix 78 CI test failures and remove dead test (#9036) 2026-04-13 10:50:24 -07:00
test_exit_cleanup_interrupt.py test: speed up slow tests (backoff + subprocess + IMDS network) (#11797) 2026-04-17 14:21:22 -07:00
test_fallback_model.py test: speed up slow tests (backoff + subprocess + IMDS network) (#11797) 2026-04-17 14:21:22 -07:00
test_flush_memories_codex.py fix(agent): respect config timeout for flush_memories instead of hardcoded 30s 2026-04-08 18:55:33 -07:00
test_interactive_interrupt.py refactor(tests): re-architect tests + fix CI failures (#5946) 2026-04-07 17:19:07 -07:00
test_interrupt_propagation.py fix(interrupt): preserve pre-start terminal interrupts 2026-04-15 13:29:57 -07:00
test_invalid_context_length_warning.py fix(tests): resolve CI test failures — pool auto-seeding, stale assertions, mock isolation 2026-04-15 22:05:21 -07:00
test_long_context_tier_429.py refactor(tests): re-architect tests + fix CI failures (#5946) 2026-04-07 17:19:07 -07:00
test_openai_client_lifecycle.py refactor(tests): re-architect tests + fix CI failures (#5946) 2026-04-07 17:19:07 -07:00
test_percentage_clamp.py fix: update 6 test files broken by dead code removal 2026-04-10 03:44:43 -07:00
test_plugin_context_engine_init.py fix(tests): make AIAgent constructor calls self-contained (#11755) 2026-04-17 12:32:03 -07:00
test_primary_runtime_restore.py fix(run_agent): recover primary client on openai transport errors 2026-04-10 03:21:24 -07:00
test_provider_fallback.py fix(tests): make AIAgent constructor calls self-contained (#11755) 2026-04-17 12:32:03 -07:00
test_provider_parity.py fix(kimi): force 0.6 on main chat path 2026-04-17 18:47:01 -07:00
test_real_interrupt_subagent.py fix(tests): fix 78 CI test failures and remove dead test (#9036) 2026-04-13 10:50:24 -07:00
test_redirect_stdout_issue.py refactor(tests): re-architect tests + fix CI failures (#5946) 2026-04-07 17:19:07 -07:00
test_run_agent.py fix(tests): make AIAgent constructor calls self-contained (#11755) 2026-04-17 12:32:03 -07:00
test_run_agent_codex_responses.py test: speed up slow tests (backoff + subprocess + IMDS network) (#11797) 2026-04-17 14:21:22 -07:00
test_sequential_chats_live.py test: regression guards for the keepalive/transport bug class (#10933) (#11266) 2026-04-16 16:36:33 -07:00
test_session_meta_filtering.py refactor(tests): re-architect tests + fix CI failures (#5946) 2026-04-07 17:19:07 -07:00
test_session_reset_fix.py refactor(tests): re-architect tests + fix CI failures (#5946) 2026-04-07 17:19:07 -07:00
test_streaming.py fix(tests): make AIAgent constructor calls self-contained (#11755) 2026-04-17 12:32:03 -07:00
test_strict_api_validation.py refactor(tests): re-architect tests + fix CI failures (#5946) 2026-04-07 17:19:07 -07:00
test_switch_model_context.py fix: pass config_context_length to switch_model context compressor 2026-04-10 05:52:45 -07:00
test_token_persistence_non_cli.py fix(tests): make AIAgent constructor calls self-contained (#11755) 2026-04-17 12:32:03 -07:00
test_tool_arg_coercion.py refactor(tests): re-architect tests + fix CI failures (#5946) 2026-04-07 17:19:07 -07:00
test_unicode_ascii_codec.py fix: always retry on ASCII codec UnicodeEncodeError — don't gate on per-component sanitization 2026-04-15 15:03:28 -07:00