<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Fallback on freshestweb</title>
    <link>https://freshestweb.com/en/tags/fallback/</link>
    <description>Recent content in Fallback on freshestweb</description>
    <generator>Hugo</generator>
    <language>en</language>
    <lastBuildDate>Mon, 16 Mar 2026 20:10:07 +0000</lastBuildDate>
    <atom:link href="https://freshestweb.com/en/tags/fallback/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>If your agent suddenly seems stupid: openclaw tools.profile-Fail</title>
      <link>https://freshestweb.com/en/blog/if-your-agent-suddenly-seems-stupid-openclaw-toolsprofile-fail/</link>
      <pubDate>Mon, 16 Mar 2026 20:09:57 +0000</pubDate>
      <guid>https://freshestweb.com/en/blog/if-your-agent-suddenly-seems-stupid-openclaw-toolsprofile-fail/</guid>
      <description>&lt;p&gt;&lt;strong&gt;Last updated:&lt;/strong&gt; March 16, 2026&lt;/p&gt;&#xA;&lt;p&gt;You switch from the top model to a cheaper model (or slip into a fallback) and suddenly your agent seems thwarted: a lot of explanation, little execution. This feels like “model got dumber”, but it is often a different error: &lt;strong&gt;Capability drift due to an incorrect tool profile&lt;/strong&gt;.&lt;/p&gt;&#xA;&lt;p&gt;We had exactly this practical case in the OpenClaw setup: With &lt;code&gt;tools.profile: messaging&lt;/code&gt; only messaging/sessions were available - but no &lt;code&gt;exec&lt;/code&gt;, no file system, no browser automation. Result: The agent plans properly, but can no longer carry out its own steps.&lt;/p&gt;</description>
    </item>
  </channel>
</rss>
