Update debug and system prompt

Tune system prompt to avoid asking questions at the end of the
output.  Update debug thinking to include the LLM model used
for that chat.

ObsidianRAGPipe.py

@@ -96,6 +96,7 @@ class Pipe:
         # Start thinking block immediately
         if think:
             yield "<think>\n"
+            yield f"**LLM Model: ** {self.valves.llm_model}\n\n"
             yield f"**Query:** {query}\n\n"
 
         # ─────────────────────────────────────────────
@@ -328,7 +329,7 @@ Rewrite the question to be standalone (respond with ONLY the rewritten question,
             content = payload.get("content", "").strip()
             source = payload.get("source", "")
 
-            part = f"### Note {i}: {file_name}\n"
+            part = f"### Note Name {i}: {file_name}\n"
             if source:
                 part += f"Original source: {source}\n"
             part += f"\n{content}"
@@ -361,6 +362,7 @@ Rewrite the question to be standalone (respond with ONLY the rewritten question,
                         Your knowledge is strictly limited to the provided <notes>.
                         - IF the answer is in the notes: Synthesize it warmly and cite it.
                         - IF the answer is NOT in the notes: You must admit it. Say: "I checked your notes, but I couldn't find info on that."
+                        - Be honest with the user.  The user does not want blind support.  You are a friendly research assistant not an overly supportive friend.
                         - **EXCEPTION:** ONLY if the user explicitly types the trigger phrase "System: Add Context" are you allowed to use outside knowledge.
                         
                         ### INSTRUCTIONS
@@ -368,6 +370,7 @@ Rewrite the question to be standalone (respond with ONLY the rewritten question,
                         2. **Synthesize:** You may combine facts from different notes to build a complete answer.
                         3. **Cite Everything:** Every single statement of fact must end with a citation in this format: `[Note Name]`.
                         4. **Tone:** Be conversational but professional. Avoid robotic phrases like "According to the provided text." Instead, say "Your note on [Topic] mentions..."
+                        5. **Additional:** Avoid asking follow up questions at the end of your output. 
                         
                         ### EXAMPLES (Follow this pattern)