[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$fW4pMilJFZwMX9LT4nnXaVBM75eKdy074vKpCG_Z2EBc":3},[4,14,20,27],{"slug":5,"title":6,"description":7,"pub_date":8,"tags":9,"draft":13},"llm-prompt-engineering","Prompt Engineering 实战：让 LLM 真正听话的技巧","System prompt 怎么写、Few-shot 怎么设计、Chain-of-Thought 原理，以及常见失败模式和调试方法。","2026-05-03",[10,11,12],"ai","llm","工程实践",false,{"slug":15,"title":16,"description":17,"pub_date":8,"tags":18,"draft":13},"rag-system-design","RAG 系统设计：从 naive 到 production-ready","Retrieval-Augmented Generation 不只是「向量数据库 + LLM」，分块策略、召回质量、重排序、缓存才是工程核心。",[10,19,11,12],"rag",{"slug":21,"title":22,"description":23,"pub_date":24,"tags":25,"draft":13},"openclaw-vs-hermes-agent","OpenClaw vs Hermes Agent：两个本地优先 Agent 的设计差异","OpenClaw（Novita AI）和 Hermes Agent（Nous Research）都是本地运行的个人 AI Agent，但在记忆系统、技能学习、运行环境和模型生态上走了不同的路。深入对比两种架构的核心差异。","2026-05-01",[10,26,11],"agent",{"slug":28,"title":29,"description":30,"pub_date":31,"tags":32,"draft":13},"ai-agent-what-is","什么是 AI Agent？从 LLM 到自主执行","LLM 本身是无状态问答机，Agent 是什么让它’动’起来的？本文深入解析 Agent 的四个核心能力、ReAct 框架、工具调用原理，以及主流框架横向对比。","2026-04-30",[10,26,11]]