feat(agent): add LangGraph scaffold for assignment 1

- src/agent/state.py: minimal five-field AgentState
- src/agent/graph.py: START -> normalize_input -> reply -> END,
  compiled with InMemorySaver; reply emits <state>1002</state> prefix
- examples/langgraph_smoke.py: offline smoke, turns 1/2/1 verified
- test/agent/test_scaffold.py: turn accumulation, thread isolation,
  fixed-template assertions (3 passed)
- requirements.txt: pin langgraph==1.2.9

Known limit: InMemorySaver state disappears on process exit.

src/agent/state.py

@@ -0,0 +1,22 @@
+"""图状态定义（作业 1）。
+
+最小状态只包含脚手架所需的五个字段；作业 3 会扩充领域状态
+（阶段、事故字段、人员字段、澄清计数、本轮 patch 等）。
+所有字段必须可序列化，不得保存 HTTP client、数据库连接等资源。
+"""
+
+from typing import TypedDict
+
+
+class AgentState(TypedDict):
+    """一次对话轮次在图中流动的状态。
+
+    ``turn_count`` 由同一 ``thread_id`` 的 checkpoint 承载，
+    节点每轮只负责 +1，不负责跨会话记忆。
+    """
+
+    session_id: str
+    turn_count: int
+    input_text: str
+    response_text: str
+    stage_code: str