feat: ACK-based message delivery, markdown stripping, prompt fixes

- Connector: send_and_wait_ack() waits for ACK before returning
- Responder: ACK waiting for DMs with retry, delay-based for broadcasts
- Chunker: strip_markdown() removes bold/italic/headers/lists from LLM output
- Router: applies strip_markdown before chunking
- Prompt: stronger no-markdown rules, no manual continuation prompt, gateway explanation
- Config: delays 3-5s (was 2.2-3), max_length 200 (was 150), max_messages 3 (was 2)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

meshai/router.py

@@ -110,6 +110,9 @@ RESPONSE STYLE:
 - When discussing problems, name the node and explain the impact
 - You CAN use 3-5 messages. Keep each sentence under 150 characters.
 - No markdown formatting - plain text only
+- ABSOLUTELY NO markdown. No asterisks, no bold, no bullet points with * or -, no numbered lists with 1. 2. 3. Just plain text sentences.
+- NEVER say "Want me to keep going?" — the message system adds this automatically when needed. If you say it yourself, users see it twice.
+- When explaining "X/Y gateways" (like 7/7), explain that it means the node is visible to X out of Y data sources (Meshview and MeshMonitor instances that monitor the mesh). It does NOT mean infrastructure routers or regional gateways.
 
 QUESTION TYPES:
 - "How's the mesh?" -> Lead with composite score. Highlight 1-2 biggest issues by name. Summarize each region briefly.
@@ -758,6 +761,10 @@ class MessageRouter:
         # Persist summary if one was created/updated
         await self._persist_summary(message.sender_id)
 
+        # Strip any markdown the LLM ignored instructions about
+        from .chunker import strip_markdown
+        response = strip_markdown(response)
+
         # Chunk the response with sentence awareness
         messages, remaining = chunk_response(
             response,