Agent on 时影https://huggingaha.github.io/topics/agent/Recent content in Agent on 时影Hugo -- gohugo.iozh-cnhuggingaha@gmail.com (时影)huggingaha@gmail.com (时影)© 2026 时影Sat, 12 Sep 2026 03:00:00 +0800Great Evals 十讲：Madhu Guru 的 Eval 建设方法论https://huggingaha.github.io/posts/how-to-build-great-evals/Sat, 12 Sep 2026 03:00:00 +0800huggingaha@gmail.com (时影)https://huggingaha.github.io/posts/how-to-build-great-evals/每个开发者都应该了解的 30 个核心智能体工程概念https://huggingaha.github.io/posts/30-core-agentic-engineering-concepts/Wed, 01 Jul 2026 00:00:00 +0000huggingaha@gmail.com (时影)https://huggingaha.github.io/posts/30-core-agentic-engineering-concepts/Beads：Git 原生问题追踪解析https://huggingaha.github.io/posts/beads/Thu, 05 Feb 2026 00:00:00 +0000huggingaha@gmail.com (时影)https://huggingaha.github.io/posts/beads/解析 AI Agent 评估https://huggingaha.github.io/posts/demystifying-evals-for-ai-agents/Mon, 12 Jan 2026 00:00:00 +0000huggingaha@gmail.com (时影)https://huggingaha.github.io/posts/demystifying-evals-for-ai-agents/Superpowers：构建可组合、工程化的 AI 智能体框架https://huggingaha.github.io/posts/superpowers/Thu, 08 Jan 2026 00:00:00 +0000huggingaha@gmail.com (时影)https://huggingaha.github.io/posts/superpowers/Agent 系统中“压缩器-预测器”的信息瓶颈与 Scaling Lawhttps://huggingaha.github.io/posts/information-theory-of-agentic-systems/Wed, 31 Dec 2025 00:00:00 +0000huggingaha@gmail.com (时影)https://huggingaha.github.io/posts/information-theory-of-agentic-systems/LatentMAS：当大模型学会“脑对脑”直接通信https://huggingaha.github.io/posts/latentmas/Sat, 06 Dec 2025 09:10:00 +0800huggingaha@gmail.com (时影)https://huggingaha.github.io/posts/latentmas/LatentMAS 源码分析https://huggingaha.github.io/posts/latentmas-code-analysis/Sat, 06 Dec 2025 09:00:00 +0800huggingaha@gmail.com (时影)https://huggingaha.github.io/posts/latentmas-code-analysis/构建鲁棒的 AI 系统：Context Engineering 2.0中的分层记忆、隔离与协调https://huggingaha.github.io/posts/context-engineering-2-0/Fri, 07 Nov 2025 00:00:00 +0000huggingaha@gmail.com (时影)https://huggingaha.github.io/posts/context-engineering-2-0/Claude Agent Skills：元工具架构分析https://huggingaha.github.io/posts/claude-skills/Sat, 01 Nov 2025 09:10:00 +0800huggingaha@gmail.com (时影)https://huggingaha.github.io/posts/claude-skills/Claude Agent Skills：第一性原理深度解析https://huggingaha.github.io/posts/claude-agent-skills-first-principles/Sat, 01 Nov 2025 09:00:00 +0800huggingaha@gmail.com (时影)https://huggingaha.github.io/posts/claude-agent-skills-first-principles/AI Agent 的有效上下文工程https://huggingaha.github.io/posts/effective-context-engineering-for-ai-agents/Fri, 03 Oct 2025 00:00:00 +0000huggingaha@gmail.com (时影)https://huggingaha.github.io/posts/effective-context-engineering-for-ai-agents/Manus：Agent 的上下文工程经验总结https://huggingaha.github.io/posts/manus-context-engineering-lessons/Sun, 20 Jul 2025 00:00:00 +0000huggingaha@gmail.com (时影)https://huggingaha.github.io/posts/manus-context-engineering-lessons/Context Engineering 综述解读https://huggingaha.github.io/posts/context-engineering-survey/Sat, 19 Jul 2025 00:00:00 +0000huggingaha@gmail.com (时影)https://huggingaha.github.io/posts/context-engineering-survey/构建可靠 AI Agent 的 12 条架构铁律https://huggingaha.github.io/posts/agent-12-factor/Sat, 12 Jul 2025 00:00:00 +0000huggingaha@gmail.com (时影)https://huggingaha.github.io/posts/agent-12-factor/构建多智能体研究系统https://huggingaha.github.io/posts/built-multi-agent-research-system/Sun, 15 Jun 2025 00:00:00 +0000huggingaha@gmail.com (时影)https://huggingaha.github.io/posts/built-multi-agent-research-system/“思考”工具：使 Claude 在复杂的工具使用情况下能够停止并思考https://huggingaha.github.io/posts/claude-think-tool/Sat, 22 Mar 2025 09:00:00 +0800huggingaha@gmail.com (时影)https://huggingaha.github.io/posts/claude-think-tool/构建高效 AI Agenthttps://huggingaha.github.io/posts/building-effective-agents/Sat, 21 Dec 2024 09:00:00 +0800huggingaha@gmail.com (时影)https://huggingaha.github.io/posts/building-effective-agents/