diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
new file mode 100644
index 0000000..cde1094
--- /dev/null
+++ b/.github/workflows/ci.yml
@@ -0,0 +1,34 @@
+name: CI
+
+on:
+  push:
+    branches: [main]
+  pull_request:
+
+jobs:
+  lint:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: astral-sh/setup-uv@v5
+        with:
+          enable-cache: true
+
+      - run: uv python install 3.12
+      - run: uv sync --frozen
+      - run: uv run ruff check .
+      - run: uv run mypy .
+
+  test:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: astral-sh/setup-uv@v5
+        with:
+          enable-cache: true
+
+      - run: uv python install 3.12
+      - run: uv sync --frozen
+      - run: uv run pytest tests/ -v
diff --git a/agent/character_agent.py b/agent/character_agent.py
index cdf853b..133c858 100644
--- a/agent/character_agent.py
+++ b/agent/character_agent.py
@@ -14,7 +14,7 @@
     retrieve_news,
     route_lore_retrieval,
 )
-from agent.state import State
+from agent.state import State, StateUpdate
 from tools.helldivers.training_manual_types import (
     convert_current_event_list,
     convert_major_order_list,
@@ -129,7 +129,7 @@ def build_graph(self) -> CompiledStateGraph:
             num_predict=4096,
         ).bind_tools(list(planet_tools.values()))
 
-        def retrieve_lore(state: State) -> State:
+        def retrieve_lore(state: State) -> StateUpdate:
             """Step 1: Retrieve lore using vector search and store in state"""
             messages = state["messages"]
             user_text = None
@@ -153,7 +153,7 @@ def retrieve_lore(state: State) -> State:
                 logger.warning("Lore retrieval failed: %s", e)
                 return {"retrieved_lore_docs": []}
 
-        def retrieve_style(state: State) -> State:
+        def retrieve_style(state: State) -> StateUpdate:
             """Step 2: Retrieve style messages"""
             messages = state["messages"]
             user_text = ""
@@ -189,7 +189,7 @@ def retrieve_style(state: State) -> State:
 
             return {"retrieved_style_docs": style_docs}
 
-        def retrieve_context(state: State) -> State:
+        def retrieve_context(state: State) -> StateUpdate:
             """Step 5: Create context retriever that picks additional documents"""
             messages = state["messages"]
             retrieved_lore_docs = state.get("retrieved_lore_docs", [])
@@ -258,7 +258,7 @@ def retrieve_context(state: State) -> State:
                 logger.warning("Context retrieval failed: %s", e)
                 return {"tool_messages": []}
 
-        def chatbot(state: State) -> State:
+        def chatbot(state: State) -> StateUpdate:
             """Main chatbot node that uses all retrieved documents"""
             messages = state["messages"]
             retrieved_lore_docs = state.get("retrieved_lore_docs", [])
diff --git a/agent/helldivers/nodes.py b/agent/helldivers/nodes.py
index cb8ffb6..c76f106 100644
--- a/agent/helldivers/nodes.py
+++ b/agent/helldivers/nodes.py
@@ -1,5 +1,5 @@
 from langchain.tools import Tool
-from agent.state import State
+from agent.state import State, StateUpdate
 from tools.helldivers.training_manual_api import (
     get_campaigns,
     get_current_status,
@@ -22,13 +22,13 @@ def __init__(self, tools: dict[str, Tool]):
         self.tools = {tool.name: tool for tool in tools.values()}
         self.tools_to_state_key = {tool.name: key for key, tool in self.tools.items()}
 
-    def __call__(self, state: State) -> State:
+    def __call__(self, state: State) -> StateUpdate:
         if tool_messages := state.get("tool_messages"):
             tool_message = tool_messages[-1]
         else:
             raise ValueError("No tool messages in state")
 
-        result: State = {"tool_messages": []}
+        result: StateUpdate = {"tool_messages": []}
         for tool_call in tool_message.tool_calls:  # type: ignore
             tool_name = tool_call["name"]
             if tool_name not in self.tools:
@@ -50,25 +50,25 @@ def __call__(self, state: State) -> State:
         return result
 
 
-def retrieve_campaigns(_: State) -> State:
+def retrieve_campaigns(_: State) -> StateUpdate:
     """Step 3: Retrieve campaigns"""
     campaigns = get_campaigns()
     return {"active_campaigns": campaigns}  # type: ignore
 
 
-def retrieve_major_orders(_: State) -> State:
+def retrieve_major_orders(_: State) -> StateUpdate:
     """Step 4: Retrieve major orders"""
     major_orders = get_major_orders()
     return {"active_major_orders": major_orders}  # type: ignore
 
 
-def retrieve_current_status(_: State) -> State:
+def retrieve_current_status(_: State) -> StateUpdate:
     """Step 5: Retrieve current status"""
     current_status = get_current_status()
     return {"current_status": current_status}  # type: ignore
 
 
-def retrieve_news(state: State) -> State:
+def retrieve_news(state: State) -> StateUpdate:
     """Step 6: Retrieve news"""
     news = get_past_week_news(state)
     return {"past_week_news": news}  # type: ignore
diff --git a/agent/state.py b/agent/state.py
index de6e02b..e2820c0 100644
--- a/agent/state.py
+++ b/agent/state.py
@@ -25,3 +25,17 @@ class State(TypedDict):
     past_week_news: list[News]  # News from the past week
     current_status: CurrentStatus  # Current status, including time and current events
     tool_messages: Annotated[list[BaseMessage], add_messages]
+
+
+class StateUpdate(TypedDict, total=False):
+    messages: Annotated[list[BaseMessage], add_messages]
+    loop_count: int
+    retrieved_lore_docs: list[Document]
+    retrieved_style_docs: list[Document]
+    retrieved_context_docs: list[Document]
+    retrieved_planet_lore: list[Document]
+    active_campaigns: list[CampaignPlanet]
+    active_major_orders: list[MajorOrder]
+    past_week_news: list[News]
+    current_status: CurrentStatus
+    tool_messages: Annotated[list[BaseMessage], add_messages]