<?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>Extensions on Max Collier</title><link>https://mcwalrus.github.io/categories/extensions/</link><description>Recent content in Extensions on Max Collier</description><generator>Hugo</generator><language>en</language><lastBuildDate>Sat, 12 Sep 2026 17:04:02 +1000</lastBuildDate><atom:link href="https://mcwalrus.github.io/categories/extensions/index.xml" rel="self" type="application/rss+xml"/><item><title>pi.dev self-help on extensions</title><link>https://mcwalrus.github.io/posts/pi-dev-extension-documentation/</link><pubDate>Sat, 12 Sep 2026 17:04:02 +1000</pubDate><guid>https://mcwalrus.github.io/posts/pi-dev-extension-documentation/</guid><description>&lt;p&gt;Notably, when you install a &lt;a href="https://pi.dev/"&gt;pi.dev&lt;/a&gt; extension, the harness never actually reads or tries to present the README.md documentation associated with the extensions installed. Often, the new ecosystem of harness extensions coming forward are not just only extensible, but may also require some aspect of configuration up-front.&lt;/p&gt;
&lt;p&gt;&lt;a href="https://pi.dev/packages"&gt;https://pi.dev/packages&lt;/a&gt;&lt;/p&gt;
&lt;div class="highlight"&gt;&lt;pre tabindex="0" style="color:#f8f8f2;background-color:#272822;-moz-tab-size:4;-o-tab-size:4;tab-size:4;-webkit-text-size-adjust:none;"&gt;&lt;code class="language-bash" data-lang="bash"&gt;&lt;span style="display:flex;"&gt;&lt;span&gt;$ pi --version
&lt;/span&gt;&lt;/span&gt;&lt;span style="display:flex;"&gt;&lt;span&gt;0.85.1
&lt;/span&gt;&lt;/span&gt;&lt;/code&gt;&lt;/pre&gt;&lt;/div&gt;&lt;p&gt;The &lt;a href="https://pi.dev/"&gt;pi.dev&lt;/a&gt; system prompt already advertises skills with their locations and descriptions. However, this rule is not applied for extensions. Therefore pi.dev agent treats extensions like opaque blobs of code, without any help documentation.&lt;/p&gt;</description></item><item><title>ECC research plugin</title><link>https://mcwalrus.github.io/posts/ecc-research-plugin/</link><pubDate>Sat, 12 Sep 2026 14:21:35 +1000</pubDate><guid>https://mcwalrus.github.io/posts/ecc-research-plugin/</guid><description>&lt;p&gt;&lt;a href="https://github.com/affaan-m/ECC"&gt;https://github.com/affaan-m/ECC&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;em&gt;The Basic Premise:&lt;/em&gt;&lt;/p&gt;
&lt;pre tabindex="0"&gt;&lt;code&gt;plan -&amp;gt; test -&amp;gt; implement -&amp;gt; review -&amp;gt; verify -&amp;gt; remember -&amp;gt; improve
&lt;/code&gt;&lt;/pre&gt;&lt;blockquote&gt;
&lt;p&gt;Your agent can write code, but ECC gives it a coordinated engineering system and toolbox: it plans before it builds, verifies changes with tests, reviews its own work from a fresh context, remembers what matters, and turns repeated wins into reusable skills and workflows.&lt;/p&gt;
&lt;/blockquote&gt;
&lt;p&gt;I am not a fan of this harness extension for a &lt;em&gt;variety of reasons:&lt;/em&gt;&lt;/p&gt;</description></item></channel></rss>