fix: ChatContext is not a sequence — use .items for len/list; add test suite

Bug fix:
- GemmaLLM.chat(): len(chat_ctx) → len(chat_ctx.items)
- GemmaLLM._compact_context(): list(chat_ctx) → list(chat_ctx.items)
  ChatContext in livekit-agents 1.7 is not iterable or sized directly;
  it exposes an .items list. The TypeError was silently killing all LLM
  responses (agent heard user but never spoke back).

Test suite (tests/):
- test_web_ui.py: static assets, token endpoint, LiveKit WS proxy
- test_llm_api.py: completion, tool calling, thinking-disabled latency, streaming
- test_mcp_tools.py: weather (Fahrenheit), time, memory, skills — all via
  docker exec + in-container MCP client
- test_agent_integration.py: process alive, worker registered, supervisord,
  container health, LiveKit API
- test_compaction.py: size-triggered compaction, system prompt preservation,
  short-context no-op
- conftest.py: shared fixtures (HTTPS client, token, LLM, docker exec helpers)
- Run with: .venv-tests/bin/python -m pytest tests/ -v

tests/test_llm_api.py

@@ -0,0 +1,106 @@
+"""Tests for the Gemma LLM API — the exact endpoint Hope's agent uses."""
+
+import time
+
+import pytest
+
+from conftest import GEMMA_BASE_URL, GEMMA_MODEL
+
+WEATHER_TOOL = {
+    "type": "function",
+    "function": {
+        "name": "get_weather",
+        "description": "Get the current weather and today's forecast for a location.",
+        "parameters": {
+            "type": "object",
+            "properties": {"location": {"type": "string"}},
+            "required": ["location"],
+        },
+    },
+}
+
+
+def _chat(llm_client, messages, *, tools=None, stream=False, max_tokens=200):
+    payload = {
+        "model": GEMMA_MODEL,
+        "messages": messages,
+        "max_tokens": max_tokens,
+        "temperature": 0.3,
+        "chat_template_kwargs": {"enable_thinking": False},
+    }
+    if tools:
+        payload["tools"] = tools
+    if stream:
+        payload["stream"] = True
+    return llm_client.post("/chat/completions", json=payload)
+
+
+@pytest.mark.slow
+def test_simple_completion(llm_client):
+    resp = _chat(
+        llm_client,
+        [{"role": "user", "content": "What is 2+2? Answer with just the number."}],
+        max_tokens=50,
+    )
+    assert resp.status_code == 200
+    data = resp.json()
+    content = (data["choices"][0]["message"].get("content") or "").strip()
+    assert "4" in content
+
+
+@pytest.mark.slow
+def test_tool_call_weather(llm_client):
+    resp = _chat(
+        llm_client,
+        [{"role": "user", "content": "What is the weather like in Raleigh right now?"}],
+        tools=[WEATHER_TOOL],
+        max_tokens=100,
+    )
+    assert resp.status_code == 200
+    choice = resp.json()["choices"][0]
+    assert choice["finish_reason"] == "tool_calls"
+    tool_calls = choice["message"]["tool_calls"]
+    assert any(tc["function"]["name"] == "get_weather" for tc in tool_calls)
+
+
+@pytest.mark.slow
+def test_thinking_disabled(llm_client):
+    """A normally-slow prompt must answer fast when enable_thinking=false."""
+    messages = [
+        {"role": "user", "content": "Count from 1 to 20, then tell me the sum."},
+    ]
+    start = time.monotonic()
+    resp = _chat(llm_client, messages, max_tokens=300)
+    elapsed = time.monotonic() - start
+    assert resp.status_code == 200
+    content = (resp.json()["choices"][0]["message"].get("content") or "").strip()
+    assert len(content) > 10
+    # With thinking enabled this prompt takes well over 3s on the xNAS.
+    assert elapsed < 3, f"response took {elapsed:.1f}s — thinking may be enabled"
+
+
+@pytest.mark.slow
+def test_streaming(llm_client):
+    resp = _chat(
+        llm_client,
+        [{"role": "user", "content": "Say the words: the quick brown fox jumps over the lazy dog."}],
+        stream=True,
+        max_tokens=100,
+    )
+    assert resp.status_code == 200
+
+    chunks = []
+    for line in resp.iter_lines():
+        if not line or not line.startswith("data: "):
+            continue
+        payload = line[6:].strip()
+        if payload == "[DONE]":
+            break
+        chunk = __import__("json").loads(payload)
+        delta = (chunk.get("choices") or [{}])[0].get("delta", {})
+        if delta.get("content"):
+            chunks.append(delta["content"])
+
+    assert len(chunks) > 1, f"expected multiple streamed chunks, got {len(chunks)}"
+    assembled = "".join(chunks)
+    assert "quick brown fox" in assembled.lower()