
		<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
			<channel>
				<title>Samuel Wang – AI-native Systems Builder</title>
				<link>https://wanglushan.cn/blog</link>
				<description>AI-native Systems Builder，专注把模糊意图、AI 协作、工程流程和运行证据组织成可执行、可验证、可持续迭代的产品系统。</description>
				<language>zh-CN</language>
				<managingEditor>wanglushan3344@gmail.com (samuel)</managingEditor>
				<webMaster>wanglushan3344@gmail.com (samuel)</webMaster>
				<lastBuildDate>Fri, 29 May 2026 00:00:00 GMT</lastBuildDate>
				<atom:link href="https://wanglushan.cn/tags/workflow/feed.xml" rel="self" type="application/rss+xml"/>
				
		<item>
			<guid>https://wanglushan.cn/blog/skill-workflow-agent-productivity-route</guid>
			<title>为什么我认为 skill -&gt; workflow -&gt; agent 才是生产力路线</title>
			<link>https://wanglushan.cn/blog/skill-workflow-agent-productivity-route</link>
			<description>生产力不是从更松散的 agent 开始，而是从可复用 skill、可验证 workflow 和可治理 agent 逐层上升。</description>
			<pubDate>Fri, 29 May 2026 00:00:00 GMT</pubDate>
			<author>wanglushan3344@gmail.com (samuel)</author>
			<category>wiki-core</category><category>ai-agent</category><category>ai-native-delivery</category><category>workflow</category>
		</item>
	
		<item>
			<guid>https://wanglushan.cn/blog/workflow-first-agent-design-principles</guid>
			<title>为什么松散 agent 不好用：workflow-first agent 的设计原则</title>
			<link>https://wanglushan.cn/blog/workflow-first-agent-design-principles</link>
			<description>松散 agent 缺少边界、回放和验收。workflow-first agent 先定义工作系统，再让 agent 执行。</description>
			<pubDate>Sun, 31 May 2026 00:00:00 GMT</pubDate>
			<author>wanglushan3344@gmail.com (samuel)</author>
			<category>wiki-core</category><category>ai-agent</category><category>workflow</category><category>validation</category>
		</item>
	
			</channel>
		</rss>
	