<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Workflow on Time Friend</title><link>https://time-friend.com/en/tags/workflow/</link><description>Recent content in Workflow on Time Friend</description><generator>Hugo</generator><language>en-GB</language><lastBuildDate>Wed, 29 Jul 2026 15:15:07 +0800</lastBuildDate><atom:link href="https://time-friend.com/en/tags/workflow/index.xml" rel="self" type="application/rss+xml"/><item><title>Five Tasks You Shouldn't Give Your Agent</title><link>https://time-friend.com/en/archive/claude-code-tasks-to-avoid/</link><pubDate>Sun, 22 Mar 2026 08:50:00 +0000</pubDate><guid>https://time-friend.com/en/archive/claude-code-tasks-to-avoid/</guid><description>&lt;p&gt;After a few months with agents, it&amp;rsquo;s easy to fall into the trap of throwing everything at it.&lt;/p&gt;
&lt;p&gt;But some tasks really shouldn&amp;rsquo;t be delegated — not because the agent isn&amp;rsquo;t capable, but because the risk/reward ratio is wrong.&lt;/p&gt;
&lt;h2 id="one-anything-touching-production-data"&gt;One: anything touching production data&lt;/h2&gt;
&lt;p&gt;Running SQL against a production database — even &lt;code&gt;SELECT&lt;/code&gt; — I won&amp;rsquo;t trust an agent with. Claude won&amp;rsquo;t intentionally break things, but it might scan a full table because it misunderstood the query.&lt;/p&gt;</description></item><item><title>Agent and Assistant Are Not the Same Thing</title><link>https://time-friend.com/en/archive/claude-code-agent-vs-assistant/</link><pubDate>Sat, 14 Mar 2026 15:40:00 +0000</pubDate><guid>https://time-friend.com/en/archive/claude-code-agent-vs-assistant/</guid><description>&lt;p&gt;When I first started using Claude Code, I spent a lot of time wondering: is this thing an agent or an assistant?&lt;/p&gt;
&lt;p&gt;Most articles try to draw a clean line — assistant does one step at a time, agent pursues a goal autonomously. But after using it daily, the line is blurry. Most tools can switch between both modes.&lt;/p&gt;
&lt;p&gt;The distinction that actually matters isn&amp;rsquo;t what you call it. It&amp;rsquo;s &lt;strong&gt;who initiates, who controls, and where it runs&lt;/strong&gt;.&lt;/p&gt;</description></item><item><title>Keeping Your Agent on Track with Plan → Code</title><link>https://time-friend.com/en/archive/claude-code-plan-code-loop/</link><pubDate>Thu, 08 Jan 2026 10:30:00 +0000</pubDate><guid>https://time-friend.com/en/archive/claude-code-plan-code-loop/</guid><description>&lt;p&gt;When I first started using Claude Code, I had a bad habit: open the terminal, type one sentence telling it what to change, watch it go, and realize later it went in the wrong direction.&lt;/p&gt;
&lt;p&gt;Here&amp;rsquo;s what I learned — agentic coding is a different skill from writing code. When you write code yourself, your hands follow your brain. When you use an agent, your value shifts to defining the goal and reviewing the result. Not typing faster.&lt;/p&gt;</description></item></channel></rss>