From 387724c35592fa1ef058c54dabd8e7a19436df6a Mon Sep 17 00:00:00 2001
From: Xubin Ren <xubinrencs@gmail.com>
Date: Sat, 16 May 2026 11:14:56 +0000
Subject: [PATCH] test(agent): add tests to ensure goal state does not leak
 across sessions

---
 tests/agent/test_context_builder.py | 25 ++++++++++++++
 tests/agent/test_loop_save_turn.py  | 53 +++++++++++++++++++++++++++++
 2 files changed, 78 insertions(+)

diff --git a/tests/agent/test_context_builder.py b/tests/agent/test_context_builder.py
index 93ce9cb46..0206d0986 100644
--- a/tests/agent/test_context_builder.py
+++ b/tests/agent/test_context_builder.py
@@ -299,6 +299,31 @@ class TestBuildMessages:
         assert "Goal (active):" in user_msg
         assert "Finish docs migration." in user_msg
 
+    def test_goal_state_does_not_leak_without_session_metadata(self, tmp_path):
+        builder = _builder(tmp_path)
+        other_session_meta = {
+            GOAL_STATE_KEY: {"status": "active", "objective": "Other chat goal."},
+        }
+
+        with_goal = builder.build_messages(
+            [],
+            "hi",
+            channel="websocket",
+            chat_id="chat-a",
+            session_metadata=other_session_meta,
+        )
+        without_goal = builder.build_messages(
+            [],
+            "hi",
+            channel="websocket",
+            chat_id="chat-b",
+            session_metadata={},
+        )
+
+        assert "Other chat goal." in str(with_goal[-1]["content"])
+        assert "Other chat goal." not in str(without_goal[-1]["content"])
+        assert "Goal (active):" not in str(without_goal[-1]["content"])
+
     def test_consecutive_same_role_merged(self, tmp_path):
         builder = _builder(tmp_path)
         history = [{"role": "user", "content": "previous user message"}]
diff --git a/tests/agent/test_loop_save_turn.py b/tests/agent/test_loop_save_turn.py
index c33ecf422..ed78e7192 100644
--- a/tests/agent/test_loop_save_turn.py
+++ b/tests/agent/test_loop_save_turn.py
@@ -9,6 +9,7 @@ from nanobot.agent.loop import AgentLoop
 from nanobot.bus.events import InboundMessage
 from nanobot.bus.queue import MessageBus
 from nanobot.providers.base import LLMResponse
+from nanobot.session.goal_state import GOAL_STATE_KEY
 from nanobot.session.manager import Session
 from nanobot.utils.webui_titles import (
     WEBUI_SESSION_METADATA_KEY,
@@ -493,6 +494,58 @@ async def test_process_message_uses_context_chat_id_for_runtime_prompt(tmp_path:
     assert loop._run_agent_loop.call_args.kwargs["chat_id"] == "thread-777"
 
 
+@pytest.mark.asyncio
+async def test_process_message_uses_explicit_session_metadata_for_goal_context(
+    tmp_path: Path,
+) -> None:
+    loop = _make_full_loop(tmp_path)
+    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
+    chat_session = loop.sessions.get_or_create("websocket:chat-with-goal")
+    chat_session.metadata[GOAL_STATE_KEY] = {
+        "status": "active",
+        "objective": "This chat goal must not leak into heartbeat.",
+    }
+    loop.sessions.save(chat_session)
+    system_session = loop.sessions.get_or_create("heartbeat")
+    system_session.metadata = {}
+    loop.sessions.save(system_session)
+
+    loop.context.build_messages = MagicMock(  # type: ignore[method-assign]
+        return_value=[
+            {"role": "system", "content": "system"},
+            {"role": "user", "content": "runtime + heartbeat"},
+        ]
+    )
+    loop._run_agent_loop = AsyncMock(return_value=(  # type: ignore[method-assign]
+        "ok",
+        [],
+        [
+            {"role": "system", "content": "system"},
+            {"role": "user", "content": "runtime + heartbeat"},
+            {"role": "assistant", "content": "ok"},
+        ],
+        "stop",
+        False,
+    ))
+
+    result = await loop._process_message(
+        InboundMessage(
+            channel="websocket",
+            sender_id="heartbeat",
+            chat_id="chat-with-goal",
+            content="heartbeat work",
+        ),
+        session_key="heartbeat",
+    )
+
+    assert result is not None
+    assert result.content == "ok"
+    kwargs = loop.context.build_messages.call_args.kwargs
+    assert kwargs["chat_id"] == "chat-with-goal"
+    assert kwargs["session_metadata"] is system_session.metadata
+    assert GOAL_STATE_KEY not in kwargs["session_metadata"]
+
+
 def test_set_tool_context_uses_effective_key_for_spawn_tool(tmp_path: Path) -> None:
     loop = _make_full_loop(tmp_path)
     spawn_tool = loop.tools.get("spawn")