<?xml version="1.0" encoding="utf-8"?>
<feed xmlns="http://www.w3.org/2005/Atom">
  <title>Condukt blog</title>
  <subtitle>A framework for building AI agents in Elixir.</subtitle>
  <link href="https://condukt.tuist.dev/blog/feed.xml" rel="self" />
  <link href="https://condukt.tuist.dev/blog/" />
  <id>https://condukt.tuist.dev/blog/</id>
  
  <updated>2026-05-03T00:00:00.000Z</updated>
  <author>
    <name>Tuist</name>
    <uri>https://tuist.dev</uri>
  </author>
  
  <entry>
    <title>Where an agent&#39;s tools should run</title>
    <link href="https://condukt.tuist.dev/blog/sandboxes-for-tool-calls/" />
    <id>https://condukt.tuist.dev/blog/sandboxes-for-tool-calls/</id>
    <updated>2026-05-03T00:00:00.000Z</updated>
    <published>2026-05-03T00:00:00.000Z</published>
    <summary>We just landed sandboxes in Condukt. Same agent, swap where the tool calls actually execute. Here is the thinking behind it.</summary>
    <author><name>The Condukt team</name></author>
  </entry>
  
  <entry>
    <title>Secrets belong in the session</title>
    <link href="https://condukt.tuist.dev/blog/agent-session-secrets/" />
    <id>https://condukt.tuist.dev/blog/agent-session-secrets/</id>
    <updated>2026-05-03T00:00:00.000Z</updated>
    <published>2026-05-03T00:00:00.000Z</published>
    <summary>Agents need to act against real systems. We added session secrets to Condukt so credentials become part of the execution boundary, not part of the conversation.</summary>
    <author><name>The Condukt team</name></author>
  </entry>
  
</feed>
