https://index.zshipu.com/ai001/tags/AI-Agents/ Recent content in AI Agents on 知识铺的博客 Hugo -- gohugo.io zh-CN Fri, 04 Sep 2026 00:00:00 +0000 https://index.zshipu.com/ai001/post/20260904/AI%E4%BB%A3%E7%90%86%E4%BB%8E%E7%AE%80%E5%8D%95%E5%93%8D%E5%BA%94%E8%BD%AC%E5%90%91%E8%87%AA%E4%B8%BB%E5%BE%AA%E7%8E%AF%E5%B8%A6%E6%9D%A5%E7%9A%84%E5%AE%89%E5%85%A8%E5%B7%A5%E7%A8%8B%E6%8C%91%E6%88%98/ Fri, 04 Sep 2026 00:00:00 +0000 https://index.zshipu.com/ai001/post/20260904/AI%E4%BB%A3%E7%90%86%E4%BB%8E%E7%AE%80%E5%8D%95%E5%93%8D%E5%BA%94%E8%BD%AC%E5%90%91%E8%87%AA%E4%B8%BB%E5%BE%AA%E7%8E%AF%E5%B8%A6%E6%9D%A5%E7%9A%84%E5%AE%89%E5%85%A8%E5%B7%A5%E7%A8%8B%E6%8C%91%E6%88%98/ AI代理从简单响应转向自主循环带来的安全工程挑战 从User→Model→Response到Goal→Plan→Retrieve→Reason→Propose Action→Verify→Execute→Observe→Continue的转变，让AI代理层面临大量工程与安全要求。这一