<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:skuf="https://skufosaur.com/ns/radar/1">
  <channel>
    <title>Skufosaur · Radar</title>
    <link>https://skufosaur.com/en/radar/</link>
    <description>Technology Radar, lab notes and project updates published in English.</description>
    <language>en</language>
    <atom:link href="https://skufosaur.com/en/radar/rss.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Before comparing agents, check the benchmark version</title>
      <link>https://skufosaur.com/en/analysis/terminal-bench-version/</link>
      <guid isPermaLink="true">https://skufosaur.com/en/analysis/terminal-bench-version/</guid>
      <description>Editorial importance 80%. Terminal-Bench 2.1 repaired 28 of 89 tasks. Compare agents on the same version.</description>
      <pubDate>Mon, 09 Nov 2026 00:00:00 GMT</pubDate>
      <category>analysis</category>
      <skuf:importance>80</skuf:importance>
    </item>
    <item>
      <title>The tool was allowed. The recipient was not.</title>
      <link>https://skufosaur.com/en/labs/recipient-binding/</link>
      <guid isPermaLink="true">https://skufosaur.com/en/labs/recipient-binding/</guid>
      <description>Editorial importance 85%. A two-case fixture for catching recipient substitution before an agent sends anything.</description>
      <pubDate>Mon, 09 Nov 2026 00:00:00 GMT</pubDate>
      <category>lab</category>
      <skuf:importance>85</skuf:importance>
    </item>
    <item>
      <title>Speed is only part of the story. Reliability becomes a metric.</title>
      <link>https://skufosaur.com/en/radar/2026-11-07/</link>
      <guid isPermaLink="true">https://skufosaur.com/en/radar/2026-11-07/</guid>
      <description>Editorial importance 85%. Terminal-Bench 2.1, KOSMOS kernels, robot failure prediction and sensor security.</description>
      <pubDate>Sat, 07 Nov 2026 00:00:00 GMT</pubDate>
      <category>radar</category>
      <skuf:importance>85</skuf:importance>
    </item>
    <item>
      <title>From analysis to a working template</title>
      <link>https://skufosaur.com/en/news/practical-collection/</link>
      <guid isPermaLink="true">https://skufosaur.com/en/news/practical-collection/</guid>
      <description>Editorial importance 45%. The collection grew to 14 materials. It added handoff notes, targeted reading, tool responses, caching, acceptance checks and stop conditions. The recipes use synthetic examples; they do not report measured savings.</description>
      <pubDate>Sat, 03 Oct 2026 00:00:00 GMT</pubDate>
      <category>update</category>
      <skuf:importance>45</skuf:importance>
    </item>
    <item>
      <title>Five primary sources and what they support</title>
      <link>https://skufosaur.com/en/news/source-library/</link>
      <guid isPermaLink="true">https://skufosaur.com/en/news/source-library/</guid>
      <description>Editorial importance 45%. The collection added OpenAI documentation on caching and evaluation, alongside work on tool responses and Anthropic materials. The date I reviewed a source is separate from its publication date; an unknown date stays unknown.</description>
      <pubDate>Sat, 03 Oct 2026 00:00:00 GMT</pubDate>
      <category>update</category>
      <skuf:importance>45</skuf:importance>
    </item>
    <item>
      <title>The first knowledge collection</title>
      <link>https://skufosaur.com/en/news/knowledge-start/</link>
      <guid isPermaLink="true">https://skufosaur.com/en/news/knowledge-start/</guid>
      <description>Editorial importance 35%. Six materials cover context, tools, retries, parallel work, cost accounting and comparisons. Each entry explains when to use the technique and where its limits are.</description>
      <pubDate>Sat, 03 Oct 2026 00:00:00 GMT</pubDate>
      <category>update</category>
      <skuf:importance>35</skuf:importance>
    </item>
  </channel>
</rss>
