feat(engine): 实现行动点系统

- 扩展 AgentState 添加 action_points, max_action_points, last_action_tick 字段
- 新增 ActionFeedback 模型用于返回行动执行结果
- 创建 action_points.py 模块实现行动点消耗与恢复逻辑
- 行动消耗表: vote=1, trigger_skill=2, influence=2, comment/support/chaos=0
- 每 tick 恢复 1 点行动点（不超过 max）
- 行动点不足时拒绝执行并返回失败反馈
- 新增 7 个测试用例，全部 37 个测试通过

🤖 Generated with [Claude Code](https://claude.com/claude-code)

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

engine-python/app/engine.py

@@ -2,7 +2,7 @@ from __future__ import annotations
 from typing import List, Optional, Tuple
 from .models import (
     WorldState, AgentState, Event, Action, Emotion, Weather,
-    GlobalEventResult, GlobalEventInfo, FactionEventResult
+    GlobalEventResult, GlobalEventInfo, FactionEventResult, ActionFeedback
 )
 from .global_events import GLOBAL_EVENT_POOL, GlobalEvent
 from .opinions import generate_opinions
@@ -12,6 +12,9 @@ from .factions import (
 )
 from .voting import process_votes, apply_votes_to_factions
 from .faction_skills import check_and_unlock_skills, apply_skill_effects
+from .action_points import (
+    regenerate_action_points, process_event_with_ap, get_action_cost
+)
 
 MAX_EVENTS = 20
 MAX_MEMORY_PER_AGENT = 3
@@ -40,11 +43,16 @@ def update_world_effects(state: WorldState) -> None:
     state.world_effects = active_effects
 
 
-def process_events(state: WorldState, events: List[Event]) -> WorldState:
-    """处理事件并更新世界状态"""
+def process_events(
+    state: WorldState, events: List[Event]
+) -> Tuple[WorldState, List[ActionFeedback]]:
+    """处理事件并更新世界状态，返回行动反馈列表"""
     # tick 递增
     state.tick += 1
 
+    # ★ 行动点恢复
+    regenerate_action_points(state)
+
     # ★ 在 tick 开始时：将上一轮投票累加到阵营能量
     apply_votes_to_factions(state)
 
@@ -58,7 +66,17 @@ def process_events(state: WorldState, events: List[Event]) -> WorldState:
     # ★ 处理投票事件（记录到 votes，下一 tick 生效）
     process_votes(state, events)
 
+    # 收集行动反馈
+    action_feedbacks: List[ActionFeedback] = []
+
     for event in events:
+        # ★ 行动点检查
+        allowed, feedback = process_event_with_ap(state, event)
+        action_feedbacks.append(feedback)
+
+        if not allowed:
+            continue  # 行动点不足，跳过此事件
+
         text = event.text
 
         # 累计能量
@@ -113,7 +131,7 @@ def process_events(state: WorldState, events: List[Event]) -> WorldState:
     # 应用已解锁技能效果
     apply_skill_effects(state)
 
-    return state
+    return state, action_feedbacks
 
 
 def check_and_trigger_global_event(