### Full Code Development Example (Python)

Source: https://maafw.com/docs/1.1-QuickStarted/index

This Python pseudocode illustrates a full code development approach within MaaFramework. It shows how to execute predefined JSON tasks, perform code-level operations like clicking, capture screenshots, register custom actions, and execute mixed task chains. This method offers deep customization but bypasses some visual tooling.

```python
# Python 伪代码示例
def main():
    # 执行 JSON 预定义任务
    result = tasker.post_task("点击开始按钮").wait().get()
    
    if result.completed:
        # 执行代码级操作
        tasker.controller.post_click(100, 100)
    else:
        # 获取当前屏幕截图
        image = tasker.controller.cached_image
        # 注册自定义动作
        tasker.resource.register_custom_action("MyAction", MyAction())
        # 执行混合任务链
        tasker.post_task("点击确认图标").wait()
```

--------------------------------

### Low-Code JSON Automation

Source: https://maafw.com/docs/1.1-QuickStarted/index

This JSON configuration defines an automated workflow with two steps: clicking a 'Start' button identified by OCR and then clicking a 'Confirm' icon identified by template matching. It specifies the recognition method, expected text or template, and the action to perform, along with the next step in the sequence.

```jsonc
{
    "点击开始按钮": {
        "recognition": "OCR",          // 文字识别引擎
        "expected": "开始",            // 目标文本
        "action": "Click",             // 执行点击操作
        "next": ["点击确认图标"]        // 后续任务链
    },
    "点击确认图标": {
        "recognition": "TemplateMatch",// 图像模板匹配
        "template": "确认.png",        // 匹配素材路径
        "action": "Click"
    }
}
```

--------------------------------

### Hybrid JSON with Custom Logic (Python)

Source: https://maafw.com/docs/1.1-QuickStarted/index

This JSON configuration extends a workflow by introducing custom recognition and action modules. The Python code demonstrates how to register these custom modules ('MyReco' and 'MyAct') with the AgentServer, allowing for complex logic within the framework. The custom recognizer returns coordinates, and the custom action performs a click and dynamically adjusts the task flow.

```jsonc
{
    "点击确认图标": {
        "next": ["自定义处理模块"]
    },
    "自定义处理模块": {
        "recognition": "Custom",
        "custom_recognition": "MyReco",  // 自定义识别器ID
        "action": "Custom",
        "custom_action": "MyAct"         // 自定义动作ID
    }
}
```

```python
# Python 伪代码示例
from maa.agent.agent_server import AgentServer

# 注册自定义识别器
@AgentServer.custom_recognition("MyReco")
class CustomReco:
    def analyze(ctx):
        return (10,10,100,100)  # 返回您自己处理的识别结果

# 注册自定义动作 
@AgentServer.custom_action("MyAct")
class CustomAction:
    def run(ctx):
        ctx.controller.post_click(100, 10).wait()  # 执行点击
        ctx.override_next(["TaskA", "TaskB"])      # 动态调整任务流

# 启动Agent服务
AgentServer.start_up(sock_id)
```

=== COMPLETE CONTENT === This response contains all available snippets from this library. No additional content exists. Do not make further requests.