Add mock-based E2E tests and gate live tests to main/nightly

Introduce a new mock-based E2E test suite that uses pytest_httpserver to
simulate LLM provider responses, eliminating the need for real API keys
on PR builds. The mock tests cover model alias routing, protocol
transformation (OpenAI↔Anthropic), Responses API passthrough/translation,
streaming, tool calls, thinking mode, and multi-turn state management.

CI changes:
- Add mock-e2e-tests job (zero secrets, runs on every PR)
- Gate all live E2E jobs to main pushes + nightly schedule
- Scope secrets to only the keys each job actually needs
- Add daily cron schedule for full live test coverage

Also relaxes exact-match assertions in live e2e tests to structural
checks (non-null, non-empty) since LLM output is non-deterministic.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

tests/archgw/config_mock_llm.yaml

@@ -0,0 +1,42 @@
+version: v0.3.0
+
+listeners:
+  - type: model
+    name: model_listener
+    port: 12000
+
+model_providers:
+  # OpenAI Models - all point to mock server
+  - model: openai/gpt-5-mini-2025-08-07
+    access_key: $OPENAI_API_KEY
+    default: true
+    base_url: http://host.docker.internal:51001
+
+  - model: openai/gpt-4o-mini
+    access_key: $OPENAI_API_KEY
+    base_url: http://host.docker.internal:51001
+
+  - model: openai/o3
+    access_key: $OPENAI_API_KEY
+    base_url: http://host.docker.internal:51001
+
+  - model: openai/gpt-4o
+    access_key: $OPENAI_API_KEY
+    base_url: http://host.docker.internal:51001
+
+  # Anthropic Models - point to mock server
+  - model: anthropic/claude-sonnet-4-20250514
+    access_key: $ANTHROPIC_API_KEY
+    base_url: http://host.docker.internal:51001
+
+# Model aliases
+model_aliases:
+  arch.summarize.v1:
+    target: gpt-5-mini-2025-08-07
+
+  arch.v1:
+    target: o3
+
+# State storage for v1/responses API multi-turn tests
+state_storage:
+  type: memory