<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/">
  <channel>
    <title>Ddl on Mini Fish</title>
    <link>https://blog.minifish.org/tags/ddl/</link>
    <description>Recent content in Ddl on Mini Fish</description>
    <image>
      <title>Mini Fish</title>
      <url>https://blog.minifish.org/android-chrome-512x512.png</url>
      <link>https://blog.minifish.org/android-chrome-512x512.png</link>
    </image>
    <generator>Hugo -- 0.161.1</generator>
    <language>en-US</language>
    <copyright>Mini Fish 2014-present. Licensed under CC-BY-NC</copyright>
    <lastBuildDate>Thu, 28 May 2026 09:00:00 +0800</lastBuildDate>
    <atom:link href="https://blog.minifish.org/tags/ddl/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>ob-skills: Packaging OceanBase Knowledge as Agent Skills</title>
      <link>https://blog.minifish.org/posts/ob-skills-oceanbase-agent-skills/</link>
      <pubDate>Thu, 28 May 2026 09:00:00 +0800</pubDate>
      <guid>https://blog.minifish.org/posts/ob-skills-oceanbase-agent-skills/</guid>
      <description>A project note on ob-skills, a set of OceanBase Cursor agent skills for solution architecture, DDL syntax checking, and embedded validation workflows.</description>
      <content:encoded><![CDATA[<p><code>ob-skills</code> packages OceanBase knowledge as agent skills. It is not an application in the usual sense. It is a set of reusable instructions and workflows for database architecture, DDL review, and tooling guidance.</p>
<p>The project exists because raw prompting is not enough for repeated technical work.</p>
<h2 id="what-is-inside">What is inside</h2>
<p>The repo contains skills such as:</p>
<ul>
<li>OceanBase solution architect guidance</li>
<li>OceanBase DDL syntax checking</li>
<li>seekdb-backed DDL validation</li>
</ul>
<p>Each skill tries to narrow the agent&rsquo;s behavior for a specific kind of work. Instead of asking a general model to “help with OceanBase,” the skill defines what kind of answer is expected, which constraints matter, and which validation path should be used.</p>
<h2 id="why-skills-instead-of-notes">Why skills instead of notes</h2>
<p>Traditional notes help humans. Skills help agents act more consistently.</p>
<p>That difference matters. A good skill should:</p>
<ul>
<li>define the task boundary</li>
<li>mention common failure modes</li>
<li>route to validation where possible</li>
<li>make assumptions explicit</li>
<li>avoid generic advice when concrete checks exist</li>
</ul>
<p>For database work, this is especially useful because the cost of vague answers is high. A DDL suggestion that sounds plausible but violates grammar or deployment constraints is worse than no suggestion.</p>
<h2 id="the-useful-pattern">The useful pattern</h2>
<p>The best skills are not huge knowledge dumps. They are small operational wrappers around repeatable judgment.</p>
<p>For example, a DDL validation skill should not merely explain DDL. It should push the agent toward an actual validation path and make clear when offline validation is possible.</p>
<p>That is the difference between “LLM as autocomplete” and “LLM as a guided operator.”</p>
<h2 id="what-i-learned">What I learned</h2>
<p>Agent quality improves when domain knowledge is packaged close to the workflow. It is not enough to rely on the model remembering a product correctly.</p>
<p>Skills are also easier to maintain than long prompts embedded in many tools. If a rule changes, update the skill. If a validation command improves, update the skill. The agent workflows that use it inherit the better behavior.</p>
<h2 id="open-source-status">Open source status</h2>
<p><code>ob-skills</code> is public because the project is mostly guidance and validation workflow. It is a good fit for sharing: others can inspect the assumptions, reuse the structure, or adapt the skill pattern to their own database work.</p>
]]></content:encoded>
    </item>
  </channel>
</rss>
