- 官方 agent 原语主轴：OpenAI
New tools for building agents/Agents、AnthropicBuilding effective agents/Create custom subagents、GoogleADK、MicrosoftFoundry Agent Service - 经典 agent 理论主轴：
Reactive Reasoning and Planning、Intelligent Agents: Theory and Practice、BDI Agents: From Theory to Practice、Reinforcement Learning: A Survey、Between MDPs and Semi-MDPs - 多步闭环与环境交互主轴：
ReAct、Reflexion、Voyager、SWE-agent、AnthropicHow to implement tool use、AnthropicClaude Agent SDK、OpenAI CookbookHandling Function Calls with Reasoning Models/Build a coding agent with GPT 5.1 - 状态、记忆与上下文管理主轴：OpenAI
Conversation state、AnthropicEffective context engineering for AI agents、LangChainMemory overview、Generative Agents、CoALA、MemGPT、MemoryBank - 观测、评测与可运维性主轴：OpenAI
Tracing/Trace grading、AnthropicDemystifying evals for AI agents、LangChainMulti-turn Evals/LangSmith Alerts、MicrosoftLLM Observability、GoogleTrace an agent、AgentOps - 工程失败模式与系统化设计主轴：OpenAI
A practical guide to building agents/Operator System Card、GoogleChoose a design pattern for your agentic AI system/Choose your agentic AI architecture components/Memory Bank overview/How Google secures AI Agents、AWSEffectively building AI agents on AWS Serverless、AutoGen
附录 G：分章资料清单