diff --git a/tests/chat/test_chat_app.py b/tests/chat/test_chat_app.py
index d0f38eab..dcfd572a 100644
--- a/tests/chat/test_chat_app.py
+++ b/tests/chat/test_chat_app.py
@@ -21,6 +21,15 @@ def test_chat_command():
         mock_chat.assert_called_once()
 
 
+def test_run_chat_creates_app_and_runs(temp_db_path: Path):
+    """Test run_chat() creates a ChatApp and calls run()."""
+    with patch("haiku.rag.chat.app.ChatApp.run") as mock_run:
+        from haiku.rag.chat import run_chat
+
+        run_chat(db_path=temp_db_path)
+        mock_run.assert_called_once()
+
+
 def _make_mock_client():
     """Create a mock HaikuRAG client."""
     mock_client = AsyncMock()
diff --git a/tests/skills/test_rag.py b/tests/skills/test_rag.py
index deafd57a..1819537c 100644
--- a/tests/skills/test_rag.py
+++ b/tests/skills/test_rag.py
@@ -337,6 +337,48 @@ class TestAskTool:
         # State should store the original question, not the augmented one
         assert state.qa_history[-1].question == query_text
 
+    async def test_ask_embeds_prior_qa_on_demand(self, rag_db, monkeypatch):
+        from haiku.rag.skills.rag import RAGState, create_skill
+        from tests.skills.conftest import VECTOR_DIM
+
+        captured_questions = []
+
+        async def mock_ask(self, question, **kwargs):
+            captured_questions.append(question)
+            return ("Answer about AI.", [])
+
+        monkeypatch.setattr(HaikuRAG, "ask", mock_ask)
+
+        skill = create_skill(db_path=rag_db)
+        ask = _get_tool(skill, "ask")
+
+        # Use the same question text for the prior QA entry and query so
+        # their fake embeddings are identical (cosine similarity = 1.0).
+        prior_question = "Tell me about AI"
+        query_text = prior_question
+
+        # Leave question_embedding=None to exercise the lazy embedding path
+        state = RAGState(
+            qa_history=[
+                QAHistoryEntry(
+                    question=prior_question,
+                    answer="AI is the simulation of human intelligence by machines.",
+                    question_embedding=None,
+                ),
+            ]
+        )
+        ctx = _make_ctx(state)
+        await ask(ctx, question=query_text)
+
+        # The lazy embedding should have populated question_embedding
+        assert state.qa_history[0].question_embedding is not None
+        assert len(state.qa_history[0].question_embedding) == VECTOR_DIM
+
+        # The augmented question should include prior context
+        assert len(captured_questions) == 1
+        assert "Context from prior questions" in captured_questions[0]
+        assert prior_question in captured_questions[0]
+
     async def test_ask_no_prior_qa_context_when_irrelevant(self, rag_db, monkeypatch):
         from haiku.rag.skills.rag import RAGState, create_skill
         from tests.skills.conftest import VECTOR_DIM
diff --git a/tests/test_utils.py b/tests/test_utils.py
index f486d120..fd3ad83d 100644
--- a/tests/test_utils.py
+++ b/tests/test_utils.py
@@ -579,6 +579,14 @@ def test_build_prompt_with_preamble():
 # --- is_up_to_date tests ---
 
 
+def test_cosine_similarity_zero_norm():
+    from haiku.rag.utils import cosine_similarity
+
+    assert cosine_similarity([0, 0, 0], [1, 2, 3]) == 0.0
+    assert cosine_similarity([1, 2, 3], [0, 0, 0]) == 0.0
+    assert cosine_similarity([0, 0], [0, 0]) == 0.0
+
+
 @pytest.mark.asyncio
 async def test_is_up_to_date(monkeypatch):
     from unittest.mock import AsyncMock, MagicMock