| Lab 01 | Basic Claude API | 03-claude-api/labs（首次调用/请求结构）+ examples/api_walkthrough.py | 🔑 |
| Lab 02 | Structured output | 05-prompting-structured-output/labs + examples/validate_retry_harness.py | 无 |
| Lab 03 | Tool use | 04-tool-use/labs + examples/tool_loop.py | 无 |
| Lab 04 | Multiple tools | 04-tool-use/labs（多工具选择）+ 06-mcp/examples/01_server_primitives_demo.py | 无 |
| Lab 05 | Error handling | 03-claude-api/examples/retry_backoff.py（指数退避实跑）+ 06-mcp/examples/02_tool_error_handling.py | 无 |
| Lab 06 | Simple RAG | 07-rag-context-engineering/labs（Lab A） | 无 |
| Lab 07 | Hybrid retrieval | 07-rag/examples/01_hybrid_vs_single_retrieval.py（BM25+向量+RRF） | 无 |
| Lab 08 | MCP server | 06-mcp/labs（Lab A–D）+ examples/01_server_primitives_demo.py | 无 |
| Lab 09 | Claude Code project setup | 09-claude-code/labs Lab 1–2（agent loop 解剖+配置静态校验） | 无 |
| Lab 10 | Claude Code + MCP | 09-claude-code/labs Lab 6（mcp-example.json 配置）+ 06-mcp README §集成模式 | 🔑 |
| Lab 11 | Agent loop | 08-agents/labs + examples/01_agent_loop_simulator.py（停止条件） | 无 |
| Lab 12 | Human-in-the-loop approval | 04-tool-use/labs（审批门）+ 13-capstone/examples/capstone_skeleton.py（审批门场景 S2） | 无 |
| Lab 13 | Evaluation harness | 12-evaluation-observability/labs Lab A——examples/01_golden_dataset_scorer.py 本身就是完整 harness（金标→trial→grader→门禁） | 无 |
| Lab 14 | Prompt injection defense | 11-security/labs（场景 B/C 对照）+ examples/indirect_injection_demo.py、least_privilege_gateway.py | 无 |