diff --git a/gptme/util/cli.py b/gptme/util/cli.py
index 4a018c0f..3bdc52b3 100644
--- a/gptme/util/cli.py
+++ b/gptme/util/cli.py
@@ -89,7 +89,7 @@ def tokens_count(text: str | None, model: str, file: str | None):
     except KeyError:
         print(f"Error: Model '{model}' not supported by tiktoken.")
         print("Supported models include: gpt-4, gpt-3.5-turbo, text-davinci-003")
-        return 1
+        sys.exit(1)
 
     # Count tokens
     tokens = enc.encode(text)
@@ -167,7 +167,7 @@ def tools_info(tool_name: str):
         print(f"Tool '{tool_name}' not found. Available tools:")
         for t in loaded_tools:
             print(f"- {t.name}")
-        return
+        sys.exit(1)
 
     print(f"Tool: {tool.name}")
     print(f"Description: {tool.desc}")
diff --git a/poetry.lock b/poetry.lock
index 19bc564e..03d87082 100644
--- a/poetry.lock
+++ b/poetry.lock
@@ -2497,6 +2497,23 @@ files = [
 pytest = ">=5.0.0"
 python-dotenv = ">=0.9.1"
 
+[[package]]
+name = "pytest-mock"
+version = "3.14.0"
+description = "Thin-wrapper around the mock package for easier use with pytest"
+optional = false
+python-versions = ">=3.8"
+files = [
+    {file = "pytest-mock-3.14.0.tar.gz", hash = "sha256:2719255a1efeceadbc056d6bf3df3d1c5015530fb40cf347c0f9afac88410bd0"},
+    {file = "pytest_mock-3.14.0-py3-none-any.whl", hash = "sha256:0b72c38033392a5f4621342fe11e9219ac11ec9d375f8e2a0c164539e0d70f6f"},
+]
+
+[package.dependencies]
+pytest = ">=6.2.5"
+
+[package.extras]
+dev = ["pre-commit", "pytest-asyncio", "tox"]
+
 [[package]]
 name = "pytest-profiling"
 version = "1.7.0"
@@ -3531,4 +3548,4 @@ server = ["flask", "flask-cors"]
 [metadata]
 lock-version = "2.0"
 python-versions = "^3.10"
-content-hash = "80aab1874f6f188bce645324628fd642236c7fff2859b6acbd242b790ab561e6"
+content-hash = "fc6e6adb9c74703fa2abf164f432eb9006fc2da17b9ab1179836395439492823"
diff --git a/pyproject.toml b/pyproject.toml
index 79012b3f..8fddbe64 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -70,6 +70,7 @@ pytest-profiling = "^1.7.0"
 pytest-dotenv = "^0.5.2"
 pytest-timeout = "^2.2.0"
 pytest-retry = "^1.6.3"
+pytest-mock = "*"
 greenlet = "*"  # dependency of playwright, but needed for coverage
 
 # docs
diff --git a/tests/test_util_cli.py b/tests/test_util_cli.py
index 9c1c7528..bff9df3d 100644
--- a/tests/test_util_cli.py
+++ b/tests/test_util_cli.py
@@ -32,35 +32,68 @@ def test_tokens_count():
         assert "Token count" in result.output
 
 
-def test_chats_list(mocker):
+def test_chats_list(tmp_path, mocker):
     """Test the chats list command."""
     runner = CliRunner()
 
-    # Test empty list
-    mocker.patch("gptme.util.cli.get_user_conversations", return_value=[])
+    # Create test conversations
+    logs_dir = tmp_path / "logs"
+    logs_dir.mkdir()
+
+    # Mock both the logs directory and the conversation listing
+    mocker.patch("gptme.dirs.get_logs_dir", return_value=str(logs_dir))
+    mocker.patch("gptme.logmanager.get_user_conversations", return_value=[])
+
+    # Test empty list (should work now since we're using our empty logs_dir)
     result = runner.invoke(main, ["chats", "ls"])
     assert result.exit_code == 0
     assert "No conversations found" in result.output
 
-    # Test with conversations
-    class MockConv:
-        def __init__(self, name, messages, modified):
-            self.name = name
-            self.messages = messages
-            self.modified = modified
-
-    mock_convs = [
-        MockConv("test-1", 5, "2024-01-01"),
-        MockConv("test-2", 10, "2024-01-02"),
-    ]
-    mocker.patch("gptme.util.cli.get_user_conversations", return_value=mock_convs)
+    # Create test conversation files with names that won't be filtered
+    conv1_dir = logs_dir / "2024-01-01-chat-one"
+    conv1_dir.mkdir()
+    (conv1_dir / "conversation.jsonl").write_text(
+        '{"role": "user", "content": "hello", "timestamp": "2024-01-01T00:00:00"}\n'
+    )
+
+    conv2_dir = logs_dir / "2024-01-01-chat-two"
+    conv2_dir.mkdir()
+    (conv2_dir / "conversation.jsonl").write_text(
+        '{"role": "user", "content": "hello", "timestamp": "2024-01-01T00:00:00"}\n'
+        '{"role": "assistant", "content": "hi", "timestamp": "2024-01-01T00:00:01"}\n'
+    )
 
+    # Create ConversationMeta objects for our test conversations
+    from gptme.logmanager import ConversationMeta
+    import time
+
+    conv1 = ConversationMeta(
+        name="2024-01-01-chat-one",
+        path=str(conv1_dir / "conversation.jsonl"),
+        created=time.time(),
+        modified=time.time(),
+        messages=1,
+        branches=1,
+    )
+    conv2 = ConversationMeta(
+        name="2024-01-01-chat-two",
+        path=str(conv2_dir / "conversation.jsonl"),
+        created=time.time(),
+        modified=time.time(),
+        messages=2,
+        branches=1,
+    )
+
+    # Update the mock to return our test conversations
+    mocker.patch("gptme.logmanager.get_user_conversations", return_value=[conv1, conv2])
+
+    # Test with conversations
     result = runner.invoke(main, ["chats", "ls"])
     assert result.exit_code == 0
-    assert "test-1" in result.output
-    assert "test-2" in result.output
-    assert "5 messages" in result.output
-    assert "10 messages" in result.output
+    assert "chat-one" in result.output
+    assert "chat-two" in result.output
+    assert "Messages: 1" in result.output  # First chat has 1 message
+    assert "Messages: 2" in result.output  # Second chat has 2 messages
 
 
 @pytest.mark.skip("Waiting for context module PR")