```

feat(engine): 添加技能查看工具并优化异步任务管理

- 添加SkillViewTool到引擎加载器中，增强技能管理功能
- 在AgentLoop中引入_active_direct_task来跟踪活跃任务
- 实现直接任务执行时的同步处理逻辑
- 更新工具实例化方式以支持依赖注入

feat(config): 增加智能体运行时参数配置支持

- 扩展AgentDefaultsConfig添加max_tokens和temperature字段
- 实现配置解析函数_first_config_value处理多个配置源
- 支持通过Web API动态更新智能体运行时参数
- 添加前端页面配置表单和验证逻辑

refactor(provider): 统一最大令牌数参数类型为可选整型

- 将所有LLM提供者的max_tokens参数改为int | None类型
- 为AnthropicProvider实现模型特定的最大令牌数默认值
- 调整参数传递逻辑，优先级：调用参数 > 配置文件 > 模型默认值
- 移除硬编码的默认值，改用条件判断

feat(process): 增强事件投影功能

- 添加工具调用开始/结束事件的映射逻辑
- 实现技能激活事件的识别和展示
- 添加辅助函数处理工具调用名称和参数提取
- 优化运行记录关联逻辑，提升事件匹配准确性

fix(web): 更新网络请求客户端信任环境设置

- 将WebFetchTool和WebSearchTool的trust_env参数设为True
- 确保HTTP客户端能够正确使用系统代理配置
- 修复可能的网络连接问题

test: 添加配置加载和事件投影相关测试

- 新增智能体默认参数配置测试用例
- 实现API配置持久化和重载测试
- 添加技能卡片和工具事件的投影测试
```

app-instance/backend/tests/unit/test_config_loader.py

@@ -1,10 +1,12 @@
 import json
 
+from fastapi.testclient import TestClient
+
 from beaver.engine import AgentLoop, EngineLoader
 from beaver.engine.providers import make_provider_bundle
 from beaver.engine.providers.litellm import LiteLLMProvider
 from beaver.foundation.config import load_config
-from beaver.interfaces.web.app import _reload_agent_config
+from beaver.interfaces.web.app import create_app, _reload_agent_config
 from beaver.services.agent_service import AgentService
 
 
@@ -161,6 +163,88 @@ def test_reload_agent_config_updates_booted_loop_config(tmp_path) -> None:
     service.close()
 
 
+def test_agent_defaults_include_runtime_controls(tmp_path) -> None:
+    config_path = tmp_path / "config.json"
+    config_path.write_text(
+        json.dumps(
+            {
+                "agents": {
+                    "defaults": {
+                        "maxTokens": 12345,
+                        "temperature": 0.4,
+                        "maxToolIterations": 9,
+                    }
+                }
+            }
+        ),
+        encoding="utf-8",
+    )
+
+    config = load_config(config_path=config_path)
+    service = AgentService(config_path=config_path)
+
+    assert config.agents_defaults.max_tokens == 12345
+    assert config.agents_defaults.temperature == 0.4
+    assert config.agents_defaults.max_tool_iterations == 9
+    assert service.profile.max_tokens == 12345
+    assert service.profile.temperature == 0.4
+    assert service.profile.max_tool_iterations == 9
+    service.close()
+
+
+def test_agent_config_api_persists_and_reloads_defaults(tmp_path) -> None:
+    config_path = tmp_path / "config.json"
+    config_path.write_text(json.dumps({"agents": {"defaults": {}}}), encoding="utf-8")
+    service = AgentService(config_path=config_path)
+    app = create_app(service=service, manage_service_lifecycle=False)
+
+    with TestClient(app) as client:
+        response = client.post(
+            "/api/agent-config",
+            json={"max_tokens": 8192, "temperature": 0.6, "max_tool_iterations": 12},
+        )
+        status = client.get("/api/status")
+
+    saved = json.loads(config_path.read_text(encoding="utf-8"))
+    defaults = saved["agents"]["defaults"]
+
+    assert response.status_code == 200
+    assert response.json() == {"ok": True}
+    assert defaults["maxTokens"] == 8192
+    assert defaults["temperature"] == 0.6
+    assert defaults["maxToolIterations"] == 12
+    assert service.profile.max_tokens == 8192
+    assert service.profile.temperature == 0.6
+    assert service.profile.max_tool_iterations == 12
+    assert status.json()["max_tokens"] == 8192
+    assert status.json()["temperature"] == 0.6
+    assert status.json()["max_tool_iterations"] == 12
+    service.close()
+
+
+def test_agent_config_api_accepts_zero_temperature_and_iterations(tmp_path) -> None:
+    config_path = tmp_path / "config.json"
+    service = AgentService(config_path=config_path)
+    app = create_app(service=service, manage_service_lifecycle=False)
+
+    with TestClient(app) as client:
+        response = client.post(
+            "/api/agent-config",
+            json={"max_tokens": None, "temperature": 0, "max_tool_iterations": 0},
+        )
+
+    config = load_config(config_path=config_path)
+
+    assert response.status_code == 200
+    assert config.agents_defaults.max_tokens is None
+    assert config.agents_defaults.temperature == 0
+    assert config.agents_defaults.max_tool_iterations == 0
+    assert service.profile.max_tokens is None
+    assert service.profile.temperature == 0
+    assert service.profile.max_tool_iterations == 0
+    service.close()
+
+
 def test_openai_compatible_qwen_config_keeps_openai_provider() -> None:
     bundle = make_provider_bundle(
         model="qwen-plus",