<?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>Systemd on thenybble.de</title>
    <link>https://thenybble.de/tags/systemd/</link>
    <description>Recent content in Systemd on thenybble.de</description>
    <generator>Hugo</generator>
    <language>en-us</language>
    <lastBuildDate>Mon, 02 Nov 2026 13:19:13 +0100</lastBuildDate>
    <atom:link href="https://thenybble.de/tags/systemd/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Using Yubikey challenge-response with sandboxed polkit-agent-helper</title>
      <link>https://thenybble.de/posts/systemd-yubico/</link>
      <pubDate>Mon, 02 Nov 2026 13:19:13 +0100</pubDate>
      <guid>https://thenybble.de/posts/systemd-yubico/</guid>
      <description>&lt;p&gt;I&amp;rsquo;ve got my computer setup to use yubikey challenge-response authentication for most things
(unlocking my KeepassXC databasee, user logins and disk encryption). I&amp;rsquo;ve had problems multiple
times with permissions for this setup: The Yubikey challenge-response implementation needs to
maintain some state. This state can either be stored in the user&amp;rsquo;s home directory, or some
root-accessible directory.&lt;/p&gt;
&lt;p&gt;Tools like &lt;code&gt;swaylock&lt;/code&gt; generally do not run as root, and so the per-user implementation is generally
more useful (it took me a while to figure this out). For system authentication via PAM, the
authentication is done via the &lt;code&gt;polkit-agent-helper&lt;/code&gt; that runs as root. Because this helper is very
security-sensitive, it is heavily sandboxed as of the time of writing.&lt;/p&gt;</description>
    </item>
  </channel>
</rss>
