<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: Antonio Marasco</title>
    <description>The latest articles on DEV Community by Antonio Marasco (@razorphish).</description>
    <link>https://dev.to/razorphish</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4026496%2F2d43d34c-89da-4aa6-a37c-c023c98723ef.gif</url>
      <title>DEV Community: Antonio Marasco</title>
      <link>https://dev.to/razorphish</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/razorphish"/>
    <language>en</language>
    <item>
      <title>55 Production-Ready Claude Prompts for Engineers</title>
      <dc:creator>Antonio Marasco</dc:creator>
      <pubDate>Wed, 15 Jul 2026 20:55:14 +0000</pubDate>
      <link>https://dev.to/razorphish/55-production-ready-claude-prompts-for-engineers-5g11</link>
      <guid>https://dev.to/razorphish/55-production-ready-claude-prompts-for-engineers-5g11</guid>
      <description>&lt;p&gt;I've been using Claude in production for a while, and I kept noticing the same thing: I was re-deriving the same prompts across every project. Code review, SQL optimization, IAM policies, refactoring, the same patterns, over and over.&lt;/p&gt;

&lt;p&gt;So I organized the 55 I actually reach for. Here's what's in them, plus 7 you can grab free.&lt;/p&gt;

&lt;h2&gt;
  
  
  7 free prompts (no signup)
&lt;/h2&gt;

&lt;p&gt;Before the pitch, here are 7 of them, free, on GitHub:&lt;br&gt;
&lt;strong&gt;&lt;a href="https://gist.github.com/razorphish/5469109bbe6f20a49280c5b5002d1a5a" rel="noopener noreferrer"&gt;https://gist.github.com/razorphish/5469109bbe6f20a49280c5b5002d1a5a&lt;/a&gt;&lt;/strong&gt;&lt;/p&gt;

&lt;p&gt;They cover SQL optimization, async refactoring, API validation, React, Python ETL, IAM policies, and microservice design. Copy, paste your context into the brackets, iterate.&lt;/p&gt;

&lt;h2&gt;
  
  
  What the full set covers
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Code review &amp;amp; architecture&lt;/strong&gt; (15): e.g. &lt;em&gt;"Review this PR for security issues, edge cases, and performance."&lt;/em&gt; Great for async review; catches things a first human pass misses.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Security &amp;amp; compliance&lt;/strong&gt; (10): e.g. &lt;em&gt;"Threat-model this system from an attacker's perspective."&lt;/em&gt; Maps to the OWASP Top 10.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Testing &amp;amp; DevOps&lt;/strong&gt; (10): integration test generation, CI/CD, failure-scenario planning.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;AI &amp;amp; agentic loops&lt;/strong&gt; (5): multi-agent designs, tool binding, function definitions.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Data &amp;amp; analytics&lt;/strong&gt; (5): e.g. &lt;em&gt;"Optimize this SQL query for a 100M-row table."&lt;/em&gt;
&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Other specialized&lt;/strong&gt; (10): the rest I use regularly.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Why package them?
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Most engineers rediscover the same ~50 prompts independently.&lt;/li&gt;
&lt;li&gt;One good prompt saves 2 to 4 hours. The time-to-value is lopsided.&lt;/li&gt;
&lt;li&gt;These are tested in production, not theoretical.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Each prompt in the full set includes the full text, a real example input, a production-ready example output, and notes on adapting it to your project, so you see how it behaves before you run it.&lt;/p&gt;

&lt;h2&gt;
  
  
  A free sample prompt
&lt;/h2&gt;

&lt;blockquote&gt;
&lt;p&gt;"Audit this Python code for security vulnerabilities. List issues by severity (critical, high, medium) and suggest a fix for each."&lt;/p&gt;
&lt;/blockquote&gt;

&lt;p&gt;Paste your code after it; Claude returns issues and fixes ordered by severity.&lt;/p&gt;

&lt;h2&gt;
  
  
  Get them
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;7 free (no signup):&lt;/strong&gt; &lt;a href="https://gist.github.com/razorphish/5469109bbe6f20a49280c5b5002d1a5a" rel="noopener noreferrer"&gt;https://gist.github.com/razorphish/5469109bbe6f20a49280c5b5002d1a5a&lt;/a&gt;
&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Full set of 55&lt;/strong&gt; (each with a worked example + real output): &lt;a href="https://2797536849428.gumroad.com/l/vxetvk" rel="noopener noreferrer"&gt;https://2797536849428.gumroad.com/l/vxetvk&lt;/a&gt; , $49, one-time.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Questions? Ask in the comments, happy to explain any prompt or use case. And I'm curious: which prompts do &lt;em&gt;you&lt;/em&gt; keep re-writing?&lt;/p&gt;

</description>
      <category>claude</category>
      <category>ai</category>
      <category>productivity</category>
      <category>programming</category>
    </item>
    <item>
      <title>The Claude prompt I use to optimize slow SQL queries</title>
      <dc:creator>Antonio Marasco</dc:creator>
      <pubDate>Wed, 15 Jul 2026 18:50:37 +0000</pubDate>
      <link>https://dev.to/razorphish/the-claude-prompt-i-use-to-optimize-slow-sql-queries-38ik</link>
      <guid>https://dev.to/razorphish/the-claude-prompt-i-use-to-optimize-slow-sql-queries-38ik</guid>
      <description>&lt;p&gt;Here's the single Claude prompt I reach for most when a query is dragging.&lt;/p&gt;

&lt;h2&gt;
  
  
  Optimize a slow SQL query
&lt;/h2&gt;



&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight sql"&gt;&lt;code&gt;&lt;span class="k"&gt;Analyze&lt;/span&gt; &lt;span class="k"&gt;and&lt;/span&gt; &lt;span class="n"&gt;optimize&lt;/span&gt; &lt;span class="n"&gt;this&lt;/span&gt; &lt;span class="n"&gt;slow&lt;/span&gt; &lt;span class="k"&gt;SQL&lt;/span&gt; &lt;span class="n"&gt;query&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;
&lt;span class="p"&gt;[&lt;/span&gt;&lt;span class="n"&gt;paste&lt;/span&gt; &lt;span class="n"&gt;query&lt;/span&gt;&lt;span class="p"&gt;]&lt;/span&gt;

&lt;span class="n"&gt;Context&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;
&lt;span class="o"&gt;-&lt;/span&gt; &lt;span class="k"&gt;Table&lt;/span&gt; &lt;span class="k"&gt;size&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="p"&gt;[&lt;/span&gt;&lt;span class="k"&gt;rows&lt;/span&gt;&lt;span class="p"&gt;]&lt;/span&gt;
&lt;span class="o"&gt;-&lt;/span&gt; &lt;span class="n"&gt;Query&lt;/span&gt; &lt;span class="n"&gt;frequency&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="p"&gt;[&lt;/span&gt;&lt;span class="n"&gt;per&lt;/span&gt; &lt;span class="k"&gt;second&lt;/span&gt;&lt;span class="o"&gt;/&lt;/span&gt;&lt;span class="k"&gt;minute&lt;/span&gt;&lt;span class="p"&gt;]&lt;/span&gt;
&lt;span class="o"&gt;-&lt;/span&gt; &lt;span class="k"&gt;Current&lt;/span&gt; &lt;span class="n"&gt;performance&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="p"&gt;[&lt;/span&gt;&lt;span class="n"&gt;ms&lt;/span&gt;&lt;span class="p"&gt;]&lt;/span&gt;  &lt;span class="o"&gt;-&amp;gt;&lt;/span&gt;  &lt;span class="n"&gt;Target&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="p"&gt;[&lt;/span&gt;&lt;span class="n"&gt;ms&lt;/span&gt;&lt;span class="p"&gt;]&lt;/span&gt;

&lt;span class="n"&gt;Suggest&lt;/span&gt; &lt;span class="n"&gt;indexes&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="n"&gt;rewrites&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="k"&gt;and&lt;/span&gt; &lt;span class="k"&gt;explain&lt;/span&gt; &lt;span class="n"&gt;the&lt;/span&gt; &lt;span class="n"&gt;trade&lt;/span&gt;&lt;span class="o"&gt;-&lt;/span&gt;&lt;span class="n"&gt;offs&lt;/span&gt; &lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="k"&gt;read&lt;/span&gt; &lt;span class="n"&gt;speed&lt;/span&gt; &lt;span class="n"&gt;vs&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt; &lt;span class="k"&gt;write&lt;/span&gt; &lt;span class="n"&gt;cost&lt;/span&gt;&lt;span class="p"&gt;).&lt;/span&gt;
&lt;span class="k"&gt;Show&lt;/span&gt; &lt;span class="k"&gt;EXPLAIN&lt;/span&gt; &lt;span class="n"&gt;reasoning&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="n"&gt;what&lt;/span&gt; &lt;span class="n"&gt;scans&lt;/span&gt; &lt;span class="n"&gt;it&lt;/span&gt; &lt;span class="n"&gt;eliminates&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;&lt;strong&gt;Use it when&lt;/strong&gt; a query is slow and you are not sure whether the fix is an index, a rewrite, or a schema change. Asking for the EXPLAIN reasoning and the read-vs-write trade-off is what turns a generic answer into one you can actually ship.&lt;/p&gt;

&lt;p&gt;This is one of 55 prompts I use in production. &lt;strong&gt;7 are free, no signup:&lt;/strong&gt; &lt;a href="https://gist.github.com/razorphish/5469109bbe6f20a49280c5b5002d1a5a" rel="noopener noreferrer"&gt;https://gist.github.com/razorphish/5469109bbe6f20a49280c5b5002d1a5a&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;The full set (each with a worked example input and a real output) is here: &lt;a href="https://2797536849428.gumroad.com/l/vxetvk" rel="noopener noreferrer"&gt;https://2797536849428.gumroad.com/l/vxetvk&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;What prompt do you reach for most? I am curious whether the "worked example + output" format beats a bare prompt list.&lt;/p&gt;

</description>
      <category>claude</category>
      <category>ai</category>
      <category>sql</category>
      <category>database</category>
    </item>
  </channel>
</rss>
