fix(critical): Senpai using Helion's memory — 3 root causes fixed

1. YAML structure bug: Senpai was in `policies:` instead of `agents:`
   in router-config.yml. Router couldn't find Senpai config → no routing
   rule → fallback to local model.

2. tool_manager agent_id not passed: memory_search and graph_query
   tools were called without agent_id → defaulted to "helion" →
   ALL agents' tool calls searched Helion's Qdrant collections.
   Fixed: agent_id now flows from main.py → execute_tool → _memory_search.

3. Config not mounted: router-config.yml was baked into Docker image,
   host changes had no effect. Added volume mount in docker-compose.

Also added:
- Sofiia agent config + routing rule (was completely missing)
- Senpai routing rule: cloud_deepseek (was falling to local qwen3:8b)
- Anti-echo instruction for memory brief injection

Deployed and verified on NODE1: Senpai now searches senpai_* collections.

Co-authored-by: Cursor <cursoragent@cursor.com>

services/router/tool_manager.py

@@ -351,9 +351,9 @@ class ToolManager:
         try:
             # Priority 1: Memory/Knowledge tools
             if tool_name == "memory_search":
-                return await self._memory_search(arguments)
+                return await self._memory_search(arguments, agent_id=agent_id)
             elif tool_name == "graph_query":
-                return await self._graph_query(arguments)
+                return await self._graph_query(arguments, agent_id=agent_id)
             # Priority 2: Web tools
             elif tool_name == "web_search":
                 return await self._web_search(arguments)
@@ -382,7 +382,7 @@ class ToolManager:
             logger.error(f"Tool execution failed: {e}")
             return ToolResult(success=False, result=None, error=str(e))
     
-    async def _memory_search(self, args: Dict) -> ToolResult:
+    async def _memory_search(self, args: Dict, agent_id: str = None) -> ToolResult:
         """Search in Qdrant vector memory using Router's memory_retrieval - PRIORITY 1"""
         query = args.get("query")
         
@@ -393,6 +393,7 @@ class ToolManager:
             if memory_retrieval and memory_retrieval.qdrant_client:
                 results = await memory_retrieval.search_memories(
                     query=query,
+                    agent_id=agent_id or helion,
                     limit=5
                 )
                 
@@ -543,7 +544,7 @@ class ToolManager:
         except Exception as e:
             return ToolResult(success=False, result=None, error=str(e))
     
-    async def _graph_query(self, args: Dict) -> ToolResult:
+    async def _graph_query(self, args: Dict, agent_id: str = None) -> ToolResult:
         """Query knowledge graph"""
         query = args.get("query")
         entity_type = args.get("entity_type")