01 · Agentic Architecture
27%多 Agent 编排、路由、交接和 orchestrator/worker 设计。
Interactive demos · see the exam logic in action
Not rote memorization — the five official exam domains laid out as a follow-along learning path: interactive demos walk you through each domain's decision process, then modular training locks it in.
The five exam domains form the course backbone; each module is ordered by official weight, paired with interactive demos and scenario training.
多 Agent 编排、路由、交接和 orchestrator/worker 设计。
项目指令、工具配置、技能与开发工作流。
提示词结构、输出约束、评估与迭代。
工具设计、MCP 集成、部署、通知与采样。
上下文组织、冷热记忆、可靠性和错误恢复。
Free demo · play anonymously
No sign-up needed — step through the decision process of an Agentic loop right in the interactive demo below; to see all the demos, head to the demo catalog.
发请求 · 读 stop_reason · 执行工具 · 回灌结果——循环的生命周期。单步跑一遍：stop_reason 如何驱动工具调用与终止。
This demo is unofficial, original practice content written from public exam objectives — not real exam questions.
messages = [{"role": "user", "content": prompt}] while True: resp = client.messages.create( model="claude-sonnet-4-6", tools=tools, messages=messages, ) messages.append({"role": "assistant", "content": resp.content}) if resp.stop_reason == "end_turn": break tool_results = [] for block in resp.content: if block.type == "tool_use": output = run_tool(block.name, block.input) tool_results.append({ "type": "tool_result", "tool_use_id": block.id, "content": output, }) messages.append({"role": "user", "content": tool_results})统计本周新增的注册用户，并按来源渠道分组。
请求尚未返回，下一步将看到 stop_reason。
本步未发生工具调用。
对话刚开始，messages 数组里只有一条 user 消息。点「下一步」把它发给 Claude，看它决定先做什么——而不是我们替它决定。
开始 · 2 tools
Jump in from any domain and move at your own pace; the interactive demos and practice inside the classroom are always there to revisit.
Enter the classroom备考时想系统查阅官方一手资源与 104 页课程文档目录？前往 CCAF 备考资料库，按五大考域定位薄弱环节、逐一查漏补缺。