<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>iautist</title>
    <link>https://iautist.com/</link>
    <atom:link href="https://iautist.com/feed.xml" rel="self" type="application/rss+xml" />
    <description>New artifacts and methodology updates from iautist. AI that respects user self-direction.</description>
    <language>en-US</language>
    <lastBuildDate>Mon, 11 May 2026 00:00:00 +0000</lastBuildDate>
    <generator>hand-authored</generator>
    <managingEditor>hello@ijarvis.ai (iJarvis LLC)</managingEditor>
    <copyright>Methodology and prose dedicated to the public domain (CC0 1.0). Code MIT-licensed.</copyright>

    <item>
      <title>v0.1 launch: methodology, reference prompt, and open-source eval harness</title>
      <link>https://iautist.com/</link>
      <guid isPermaLink="false">iautist-launch-v0.1-2026-05-11</guid>
      <pubDate>Mon, 11 May 2026 00:00:00 +0000</pubDate>
      <dc:creator>iJarvis LLC</dc:creator>
      <description><![CDATA[
        <p>iautist.com launches with three artifacts:</p>
        <ol>
          <li><strong>v0.1 reference disclosure-safe system prompt</strong>, shipped on the page as a copy-to-clipboard block. License: CC0 1.0.</li>
          <li><strong>Disclosure-Safe Eval Harness</strong>, an open-source Python package for measuring how LLM output changes when users disclose cognitive-profile information. License: MIT. Repository: <a href="https://github.com/ijarviscom/iautist-eval-harness">github.com/ijarviscom/iautist-eval-harness</a>.</li>
          <li><strong>Methodology v0.1</strong>, a six-stage evaluation protocol that generalizes the Wohn et al. CHI 2026 audit pipeline. License: CC0 1.0.</li>
        </ol>
        <p>Initial-run results against frontier production models are forthcoming.</p>
      ]]></description>
    </item>
  </channel>
</rss>
