Add dummy LLM provider, update Concelier sources and JobEngine endpoints

- AdvisoryAI: DummyLlmProvider for offline/testing scenarios,
  wire in LlmProviderFactory
- Concelier: source definitions, registry, and management endpoint updates
- JobEngine: approval and release endpoint updates
- etc/llm-providers/dummy.yaml config

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

etc/llm-providers/dummy.yaml

@@ -0,0 +1,17 @@
+# Dummy echo-reverse provider for testing.
+# Requires no external API — reverses user input as the "answer".
+# Priority 1 = highest priority, so it's selected as the default provider.
+
+enabled: true
+priority: 1
+
+model:
+  name: "dummy-echo-reverse"
+
+inference:
+  temperature: 0
+  maxTokens: 4096
+
+request:
+  timeout: "00:00:05"
+  maxRetries: 0