tool/actor_runtime: add failing test proving router path in _execute_tool_call swallows exceptions without recording ToolCallRecord #10383

New issue

Open

opened 2026-04-18 09:21:54 +00:00 by HAL9000 · 0 comments

HAL9000 commented

2026-04-18 09:21:54 +00:00

Owner

Metadata

Commit Message: test(tool/actor_runtime): add failing test for router path exception swallowing in _execute_tool_call
Branch: test/actor-runtime-router-error-tdd

Background and Context

The ToolCallingRuntime._execute_tool_call() method has asymmetric error handling: the runner path has proper try/except blocks, but the router path (when self._router is not None) has none. This means any exception from self._router.route(payload) propagates uncaught, leaving the tool call unrecorded in the actor context and the ACTOR_ERRORED event unemitted. This TDD issue captures a failing test that proves the bug exists before the fix is applied.

Expected Behavior

A failing test exists at tests/unit/tool/test_actor_runtime_router_error.py that:

Is tagged @tdd_expected_fail
Fails with current code (proving the bug exists)
Passes after the corresponding bug fix is applied

Acceptance Criteria

Test file exists at tests/unit/tool/test_actor_runtime_router_error.py
Test is tagged @tdd_expected_fail and fails before the fix
Test passes after the bug fix in the corresponding Bug issue

Subtasks

Write the failing test
Verify it fails with current code
Link to the Bug issue

Definition of Done

This issue is complete when:

The test file exists and is tagged @tdd_expected_fail
The test fails with current code (proving the bug exists)
The corresponding Bug issue fix makes this test pass

TDD Test Specification

Test Description

Add a failing BDD/unit test that proves ToolCallingRuntime._execute_tool_call() does not record a ToolCallRecord or emit ACTOR_ERRORED when self._router.route(payload) raises an exception.

Expected Failing Test

# tests/unit/tool/test_actor_runtime_router_error.py
import pytest
from unittest.mock import MagicMock, patch
from cleveragents.tool.actor_runtime import ToolCallingRuntime, LLMToolCall
from cleveragents.tool.actor_context import ToolActorContext

def test_router_path_records_tool_call_on_exception():
    """When router.route() raises, _execute_tool_call should still record a ToolCallRecord."""
    registry = MagicMock()
    runner = MagicMock()
    llm_caller = MagicMock()
    router = MagicMock()
    router.route.side_effect = RuntimeError("Router failure")

    runtime = ToolCallingRuntime(
        registry=registry,
        runner=runner,
        llm_caller=llm_caller,
        router=router,
    )
    context = ToolActorContext(plan_id="test-plan", phase="execute")
    tool_call = LLMToolCall(name="my_tool", arguments={"key": "val"})

    # This should NOT raise - it should catch the error and record it
    result = runtime._execute_tool_call(tool_call, context, iteration=1)

    # Should have recorded a failed ToolCallRecord
    assert len(context.tool_call_history) == 1
    record = context.tool_call_history[0]
    assert record.success is False
    assert record.error is not None
    assert "Router failure" in record.error

Why This Test Currently Fails

In src/cleveragents/tool/actor_runtime.py, the router path (lines ~350-362) has no try/except:

if self._router is not None:
    payload = {"name": tool_call.name, "args": enriched_inputs, "type": "tool_call"}
    routed_result = self._router.route(payload)  # No error handling!
    elapsed_ms = (time.monotonic() - start) * 1000.0
    success = routed_result.result.success
    ...

If self._router.route(payload) raises, the exception propagates without:

Recording a ToolCallRecord in the actor context
Emitting ACTOR_ERRORED event
Returning an error result to the LLM

The runner path (lines ~363-383) has proper try/except for both ToolError and Exception.

Automated by CleverAgents Bot
Agent: new-issue-creator

## Metadata - **Commit Message**: `test(tool/actor_runtime): add failing test for router path exception swallowing in _execute_tool_call` - **Branch**: `test/actor-runtime-router-error-tdd` ## Background and Context The `ToolCallingRuntime._execute_tool_call()` method has asymmetric error handling: the runner path has proper try/except blocks, but the router path (when `self._router is not None`) has none. This means any exception from `self._router.route(payload)` propagates uncaught, leaving the tool call unrecorded in the actor context and the `ACTOR_ERRORED` event unemitted. This TDD issue captures a failing test that proves the bug exists before the fix is applied. ## Expected Behavior A failing test exists at `tests/unit/tool/test_actor_runtime_router_error.py` that: - Is tagged `@tdd_expected_fail` - Fails with current code (proving the bug exists) - Passes after the corresponding bug fix is applied ## Acceptance Criteria - [ ] Test file exists at `tests/unit/tool/test_actor_runtime_router_error.py` - [ ] Test is tagged `@tdd_expected_fail` and fails before the fix - [ ] Test passes after the bug fix in the corresponding Bug issue ## Subtasks - [ ] Write the failing test - [ ] Verify it fails with current code - [ ] Link to the Bug issue ## Definition of Done This issue is complete when: - The test file exists and is tagged `@tdd_expected_fail` - The test fails with current code (proving the bug exists) - The corresponding Bug issue fix makes this test pass --- ## TDD Test Specification ### Tags @tdd_issue @tdd_issue_1 @tdd_expected_fail ### Test Description Add a failing BDD/unit test that proves `ToolCallingRuntime._execute_tool_call()` does not record a `ToolCallRecord` or emit `ACTOR_ERRORED` when `self._router.route(payload)` raises an exception. ### Expected Failing Test ```python # tests/unit/tool/test_actor_runtime_router_error.py import pytest from unittest.mock import MagicMock, patch from cleveragents.tool.actor_runtime import ToolCallingRuntime, LLMToolCall from cleveragents.tool.actor_context import ToolActorContext def test_router_path_records_tool_call_on_exception(): """When router.route() raises, _execute_tool_call should still record a ToolCallRecord.""" registry = MagicMock() runner = MagicMock() llm_caller = MagicMock() router = MagicMock() router.route.side_effect = RuntimeError("Router failure") runtime = ToolCallingRuntime( registry=registry, runner=runner, llm_caller=llm_caller, router=router, ) context = ToolActorContext(plan_id="test-plan", phase="execute") tool_call = LLMToolCall(name="my_tool", arguments={"key": "val"}) # This should NOT raise - it should catch the error and record it result = runtime._execute_tool_call(tool_call, context, iteration=1) # Should have recorded a failed ToolCallRecord assert len(context.tool_call_history) == 1 record = context.tool_call_history[0] assert record.success is False assert record.error is not None assert "Router failure" in record.error ``` ### Why This Test Currently Fails In `src/cleveragents/tool/actor_runtime.py`, the router path (lines ~350-362) has no try/except: ```python if self._router is not None: payload = {"name": tool_call.name, "args": enriched_inputs, "type": "tool_call"} routed_result = self._router.route(payload) # No error handling! elapsed_ms = (time.monotonic() - start) * 1000.0 success = routed_result.result.success ... ``` If `self._router.route(payload)` raises, the exception propagates without: 1. Recording a `ToolCallRecord` in the actor context 2. Emitting `ACTOR_ERRORED` event 3. Returning an error result to the LLM The runner path (lines ~363-383) has proper try/except for both `ToolError` and `Exception`. --- **Automated by CleverAgents Bot** Agent: new-issue-creator

HAL9000 referenced this issue

2026-04-18 09:22:18 +00:00

tool/actor_runtime: ToolCallingRuntime._execute_tool_call router path lacks error handling — exceptions not recorded in ToolCallRecord #10386

HAL9000 added the