diff --git a/agent.py b/agent.py
new file mode 100644
index 0000000..4cda3cf
--- /dev/null
+++ b/agent.py
@@ -0,0 +1,51 @@
+#!/usr/bin/env python3
+"""RAG Agent module - creates an agent with RAG tools and system prompt."""
+
+from langchain_community.chat_models import ChatOllama
+from langgraph.prebuilt import create_agent
+
+from rag_tools import search_knowledge_base, add_to_knowledge_base
+
+
+SYSTEM_PROMPT = """You are a helpful AI assistant with access to a knowledge base via RAG (Retrieval-Augmended Generation).
+
+Rules for using the knowledge base:
+
+1. ALWAYS search the knowledge base first before answering any question.
+   Use the `search_knowledge_base` tool with a relevant query.
+
+. If the search returns relevant results, use them to provide accurate, factual answers.
+3. If the search returns no results, inform the user that the information is not in the knowledge base,
+   and offer to add it using the `add_to_knowledge_base` tool.
+4. When a user provides new information or asks you to remember something,
+   use the `add_to_knowledge_base` tool to store it.
+5. Always cite the source (title) of documents from the knowledge base in your answers.
+
+Be concise, accurate, and helpful."""
+
+
+def create_rag_agent(ollama_base_url="http://localhost:11434", model="llama3"):
+    """Create and return a RAG agent with knowledge base tools.
+
+    Args:
+        ollama_base_url: Ollama server URL.
+        model: Ollama model name.
+
+    Returns:
+        Configured agent executor.
+    """
+    llm = ChatOllama(
+        model=model,
+        base_url=ollama_base_url,
+        temperature=0.0,
+    )
+
+    tools = [search_knowledge_base, add_to_knowledge_base]
+
+    agent = create_agent(
+        llm=llm,
+        tools=tools,
+        system_prompt=SYSTEM_PROMPT,
+    )
+
+    return agent