feat: inject briefing article content for deep article Q&A

Add _build_briefing_article_context() helper to llm.py that reads
rss_item_ids from briefing message metadata and injects article content
into the system prompt. Pass conv_id through build_context() and
generation_task.py.

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

tests/test_news_api.py

@@ -56,6 +56,28 @@ def test_list_news_item_serialisation():
     assert result["published_at"] == pub.isoformat()
 
 
+
+def test_build_briefing_article_context_metadata_extraction():
+    """Verify the rss_item_ids extraction logic from message metadata."""
+    import json
+
+    # Simulates the logic in _build_briefing_article_context
+    def extract_ids(msg_metadata):
+        meta = msg_metadata or {}
+        if isinstance(meta, str):
+            try:
+                meta = json.loads(meta)
+            except Exception:
+                return []
+        return meta.get("rss_item_ids") or []
+
+    assert extract_ids({}) == []
+    assert extract_ids(None) == []
+    assert extract_ids({"rss_item_ids": [1, 2, 3]}) == [1, 2, 3]
+    assert extract_ids(json.dumps({"rss_item_ids": [4, 5]})) == [4, 5]
+    assert extract_ids("not json") == []
+
+
 def test_list_news_null_content_serialisation():
     """Null content should produce empty snippet without error."""
     item = {