<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0"><channel>
    <title>Sintropyc News</title>
    <link>https://news.sintropyc.com/</link>
    <description>Product updates and notes on AI and application security from Sintropyc.</description>
    <language>en</language>
    <lastBuildDate>Thu, 22 Oct 2026 14:04:01 +0000</lastBuildDate>
    <atom:link xmlns:atom="http://www.w3.org/2005/Atom" href="https://news.sintropyc.com/rss.xml" rel="self" type="application/rss+xml"/>
    <item>
      <title>New PHP security fixes — the no-DevOps guide to updating your site</title>
      <link>https://news.sintropyc.com/php-vulnerability</link>
      <guid isPermaLink="true">https://news.sintropyc.com/php-vulnerability</guid>
      <category>Security</category>
      <pubDate>Thu, 22 Oct 2026 00:00:00 +0000</pubDate>
      <description>PHP&#x27;s September and October 2026 updates patched high-severity bugs in SOAP, the PostgreSQL extension, OpenSSL verification and PHP-FPM. If your site runs PHP and nobody on your team speaks server, here is the shortest path to getting on a safe version today.</description>
    </item>
    <item>
      <title>AI agents can now write, run and publish their own exploit code</title>
      <link>https://news.sintropyc.com/ai-agents-that-write-and-ship-exploits</link>
      <guid isPermaLink="true">https://news.sintropyc.com/ai-agents-that-write-and-ship-exploits</guid>
      <category>AI</category>
      <pubDate>Thu, 15 Oct 2026 00:00:00 +0000</pubDate>
      <description>A new wave of AI models can draft an attack, run it, read the output and iterate without a human in the loop. That raises the bar for what counts as proof, and it is exactly how the Sintropyc agent already works.</description>
    </item>
    <item>
      <title>Continuous now keeps agent memory between runs</title>
      <link>https://news.sintropyc.com/continuous-tier-agent-memory</link>
      <guid isPermaLink="true">https://news.sintropyc.com/continuous-tier-agent-memory</guid>
      <category>Release</category>
      <pubDate>Thu, 08 Oct 2026 00:00:00 +0000</pubDate>
      <description>The Continuous tier now carries what the agent learned about your app from one audit into the next, so each re-test starts from proven context instead of a cold start.</description>
    </item>
</channel></rss>
