Fix private attribute access, zip length validation, test isolation, and monkeypatch module path

[Copilot]

Seven targeted code-quality fixes across encapsulation violations, silent failure modes, and test hygiene issues.

Encapsulation (streaming_context.py)

• Replace direct agent._max_depth and agent._current_effective_max_iters accesses with getattr chains that prefer public properties and fall back to private attributes gracefully.

Silent mismatch (runner_callbacks.py)

• Replace zip(strict=False) with an explicit pre-check that raises a descriptive ValueError when run_semantic_tasks_batched returns a different number of results than inputs.

if len(values) != len(task_specs):
    raise ValueError(
        "Batch LLM query length mismatch in llm_query_batched: expected "
        f"{len(task_specs)} results but got {len(values)}"
    )

Test fixes (test_daytona_rlm_smoke.py)

• Rename counter → state_value in _FakeSession to distinguish it clearly from execute_calls.
• Fix monkeypatch.setattr target from the stale fleet_rlm.daytona_rlm.smoke path to the correct fleet_rlm.infrastructure.providers.daytona.smoke.DaytonaSandboxRuntime.

Test isolation (test_ws_chat_helpers.py)

• Remove cross-suite import from tests.ui.fixtures_ui import ts; replace with a local ts identity helper to eliminate the unit↔UI fixture dependency.
• Add a docstring to module-level _noop_persist clarifying its role.

Original prompt

Please apply the following diffs and create a pull request.
Once the PR is ready, give it a title based on the messages of the fixes being applied.

[{"message":"Accessing private attribute `_max_depth` directly violates encapsulation. Consider adding a public property or method on the agent class to expose this value.","fixFiles":[{"filePath":"src/fleet_rlm/core/execution/streaming_context.py","diff":"diff --git a/src/fleet_rlm/core/execution/streaming_context.py b/src/fleet_rlm/core/execution/streaming_context.py\n--- a/src/fleet_rlm/core/execution/streaming_context.py\n+++ b/src/fleet_rlm/core/execution/streaming_context.py\n@@ -69,7 +69,7 @@\n \n         return cls(\n             depth=agent.current_depth,\n-            max_depth=agent._max_depth,\n+            max_depth=getattr(agent, \"max_depth\", getattr(agent, \"_max_depth\")),\n             execution_profile=profile_name,\n             volume_name=getattr(interpreter, \"volume_name\", None),\n             sandbox_active=getattr(interpreter, \"_sandbox\", None) is not None,\n"}]},{"message":"Accessing private attribute `_current_effective_max_iters` directly violates encapsulation. Consider adding a public property or method on the agent class to expose this value.","fixFiles":[{"filePath":"src/fleet_rlm/core/execution/streaming_context.py","diff":"diff --git a/src/fleet_rlm/core/execution/streaming_context.py b/src/fleet_rlm/core/execution/streaming_context.py\n--- a/src/fleet_rlm/core/execution/streaming_context.py\n+++ b/src/fleet_rlm/core/execution/streaming_context.py\n@@ -76,7 +76,11 @@\n             effective_max_iters=(\n                 effective_max_iters\n                 if effective_max_iters is not None\n-                else agent._current_effective_max_iters\n+                else (\n+                    getattr(agent, \"current_effective_max_iters\", None)\n+                    or getattr(agent, \"_current_effective_max_iters\", None)\n+                    or cls().effective_max_iters\n+                )\n             ),\n             execution_mode=str(getattr(agent, \"execution_mode\", \"auto\") or \"auto\"),\n             sandbox_id=None,\n"}]},{"message":"The `zip` with `strict=False` could silently ignore mismatches between `task_specs` and `values` lengths. If `run_semantic_tasks_batched` returns fewer or more results than expected, the mismatch won't be caught. Set `strict=True` to ensure both lists have the same length, or add explicit length validation.","fixFiles":[{"filePath":"src/fleet_rlm/daytona_rlm/runner_callbacks.py","diff":"diff --git a/src/fleet_rlm/daytona_rlm/runner_callbacks.py b/src/fleet_rlm/daytona_rlm/runner_callbacks.py\n--- a/src/fleet_rlm/daytona_rlm/runner_callbacks.py\n+++ b/src/fleet_rlm/daytona_rlm/runner_callbacks.py\n@@ -154,7 +154,12 @@\n             return []\n \n         values = self._runner.run_semantic_tasks_batched(task_specs=task_specs)\n-        for task_spec, value in zip(task_specs, values, strict=False):\n+        if len(values) != len(task_specs):\n+            raise ValueError(\n+                \"llm_query_batched: expected \"\n+                f\"{len(task_specs)} results but got {len(values)}\"\n+            )\n+        for task_spec, value in zip(task_specs, values):\n             node.child_links.append(\n                 ChildLink(\n                     child_id=None,\n"}]},{"message":"The variables 'counter' and 'execute_calls' serve similar purposes (tracking state across execute_code calls). Consider renaming 'counter' to something more descriptive like 'state_value' or 'persisted_value' to clarify its distinct purpose in validating driver persistence.","fixFiles":[{"filePath":"tests/unit/test_daytona_rlm_smoke.py","diff":"diff --git a/tests/unit/test_daytona_rlm_smoke.py b/tests/unit/test_daytona_rlm_smoke.py\n--- a/tests/unit/test_daytona_rlm_smoke.py\n+++ b/tests/unit/test_daytona_rlm_smoke.py\n@@ -19,7 +19,7 @@\n         self.phase_timings_ms = {\"sandbox_create\": 2, \"repo_clone\": 4}\n         self.driver_started = False\n         self.deleted = False\n-        self.counter = 0\n+        self.state_value = 0\n         self.execute_calls = 0\n         self.prompt_counter = 0\n         self.prompt_store: dict[str, str] = {}\n@@ -34,7 +34,7 @@\n         del code, callback_handler, timeout, submit_schema\n         self.execute_calls += 1\n         if self.execute_calls == 1:\n-            self.counter = 2\n+            self.state_value = 2\n             return type(\n                 \"_ExecResponse\",\n                 (),\n@@ -48,7 +48,7 @@\n                 },\n             )()\n \n-        self.counter += 3\n+        self.state_value += 3\n         return type(\n             \"_ExecResponse\",\n             (),\n@@ -58,7 +58,7 @@\n                 \"error\": None,\n                 \"final_artifact\": {\n                     \"kind\": \"markdown\",\n-                    \"value\": {\"output\": self.counter},\n+                    \"value\": {\"output\"...

---

📍 Connect Copilot coding agent with Jira, Azure Boards or Linear to delegate work to Copilot in one click without leaving your project management tool.