test: never call a live LLM from tests, whatever the local .env says

With OPENROUTER_API_KEY in a developer's .env the chat agent switched on in dialog tests (real API calls,
spent money, flaky results). An autouse fixture turns the agent off; agent tests pass explicit providers.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

tests/conftest.py

@@ -0,0 +1,17 @@
+"""Test-wide isolation: a developer's local `.env` (e.g. OPENROUTER_API_KEY) must never make tests call a real
+LLM provider — that costs money and makes results depend on the network. Agent tests pass explicit
+providers/settings; everything else runs with the chat agent off."""
+
+import pytest
+
+from app.config import get_settings
+
+
+@pytest.fixture(autouse=True)
+def _no_live_llm(monkeypatch: pytest.MonkeyPatch):
+    monkeypatch.setenv("CHAT_AGENT", "off")
+    monkeypatch.setenv("OPENROUTER_API_KEY", "")
+    monkeypatch.setenv("LLM_PROVIDER", "mock")
+    get_settings.cache_clear()
+    yield
+    get_settings.cache_clear()