<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Swagger on Cdani&#39;s Blog</title>
    <link>https://c-daniele.github.io/en/tags/swagger/</link>
    <description>Recent content in Swagger on Cdani&#39;s Blog</description>
    <generator>Hugo</generator>
    <language>en-US</language>
    <lastBuildDate>Thu, 15 May 2025 00:00:00 +0200</lastBuildDate>
    <atom:link href="https://c-daniele.github.io/en/tags/swagger/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Policy Puppetry Prompt Injection</title>
      <link>https://c-daniele.github.io/en/posts/2025-05-15-policy-puppetry/</link>
      <pubDate>Thu, 15 May 2025 00:00:00 +0200</pubDate>
      <guid>https://c-daniele.github.io/en/posts/2025-05-15-policy-puppetry/</guid>
      <description>&lt;h1 id=&#34;policy-puppetry-prompt-injection&#34;&gt;Policy Puppetry Prompt Injection&lt;/h1&gt;&#xA;&lt;p&gt;A few days ago, I experimented with some Jailbreaking techniques, which I share in the &lt;a href=&#34;https://github.com/c-daniele/policy-puppetry&#34; target=&#34;_blank&#34; rel=&#34;noopener noreffer &#34;&gt;repo&lt;/a&gt;.&lt;br&gt;&#xA;I started from a &lt;a href=&#34;https://hiddenlayer.com/innovation-hub/novel-universal-bypass-for-all-major-llms/&#34; target=&#34;_blank&#34; rel=&#34;noopener noreffer &#34;&gt;HiddenLayer article&lt;/a&gt; published a few weeks ago, where the research team described a rather creative and ingenious &lt;strong&gt;jailbreaking&lt;/strong&gt; technique to bypass safety guardrails and the alignment of frontier models.&lt;br&gt;&#xA;The technique appears to be &lt;strong&gt;universal&lt;/strong&gt; and applicable with a &lt;strong&gt;single prompt&lt;/strong&gt; to multiple models, capable of revealing typically unsafe content or even portions of the native system prompt.&lt;/p&gt;</description>
    </item>
    <item>
      <title>Langchain pt. 3 - How to call Rest API in natural language</title>
      <link>https://c-daniele.github.io/en/posts/2024-04-20-langchain-api/</link>
      <pubDate>Sat, 20 Apr 2024 19:00:00 +0200</pubDate>
      <guid>https://c-daniele.github.io/en/posts/2024-04-20-langchain-api/</guid>
      <description>&lt;h2 id=&#34;intro&#34;&gt;Intro&lt;/h2&gt;&#xA;&lt;p&gt;Last year, Gartner put Generative AI at the peak of inflated expectations in its AI &lt;a href=&#34;https://www.gartner.com/en/articles/what-s-new-in-artificial-intelligence-from-the-2023-gartner-hype-cycle&#34; target=&#34;_blank&#34; rel=&#34;noopener noreffer &#34;&gt;Hype Cycle&lt;/a&gt;.&lt;/p&gt;&#xA;&lt;p&gt;Recently, big tech leaders &lt;a href=&#34;https://www.wired.com/story/amazons-cloud-boss-selipsky-generative-ai-hype/&#34; target=&#34;_blank&#34; rel=&#34;noopener noreffer &#34;&gt;compared the hype around GenAI to the &lt;em&gt;dotcom&lt;/em&gt; bubble&lt;/a&gt;.&#xA;Furthermore, according to some &lt;a href=&#34;https://www.theinformation.com/articles/generative-ai-providers-quietly-tamp-down-expectations?ref=wheresyoured.at&#34; target=&#34;_blank&#34; rel=&#34;noopener noreffer &#34;&gt;rumors&lt;/a&gt;, the main Cloud Providers are even giving instructions to their Sales Teams to slow down the enthusiasm towards customers regarding GenAI initiatives and promoting cost-vs-benefits awareness.&#xA;Has the drop into the &lt;a href=&#34;https://it.wikipedia.org/wiki/Hype_cycle&#34; target=&#34;_blank&#34; rel=&#34;noopener noreffer &#34;&gt;trough of disillusionment&lt;/a&gt; already begun?&lt;/p&gt;</description>
    </item>
  </channel>
</rss>
