<?xml version="1.0" encoding="UTF-8" ?>
<rss version="2.0">
    <channel>
      <title>The Behavioral Layer</title>
      <link>https://behaviorlayer.ai</link>
      <description>Last 10 notes on The Behavioral Layer</description>
      <generator>Quartz -- quartz.jzhao.xyz</generator>
      <item>
    <title>OpenAI Model Spec</title>
    <link>https://behaviorlayer.ai/behavior/openai-model-spec</link>
    <guid>https://behaviorlayer.ai/behavior/openai-model-spec</guid>
    <description><![CDATA[ A public, versioned document in which a frontier lab specifies how its models should behave: how they follow instructions, resolve conflicts between competing direct... ]]></description>
    <pubDate>Tue, 18 Aug 2026 00:00:00 GMT</pubDate>
  </item><item>
    <title>Sparrow</title>
    <link>https://behaviorlayer.ai/behavior/sparrow</link>
    <guid>https://behaviorlayer.ai/behavior/sparrow</guid>
    <description><![CDATA[ An early, concrete demonstration that you can take “behave well” and break it into specific rules, then train and judge an agent against each one. ]]></description>
    <pubDate>Tue, 18 Aug 2026 00:00:00 GMT</pubDate>
  </item><item>
    <title>NIST AI RMF</title>
    <link>https://behaviorlayer.ai/briefings/nist-ai-rmf</link>
    <guid>https://behaviorlayer.ai/briefings/nist-ai-rmf</guid>
    <description><![CDATA[ The vendor-neutral framework that gives the behavioral layer an institutional vocabulary. ]]></description>
    <pubDate>Tue, 18 Aug 2026 00:00:00 GMT</pubDate>
  </item><item>
    <title>Building Effective Agents</title>
    <link>https://behaviorlayer.ai/build/building-effective-agents</link>
    <guid>https://behaviorlayer.ai/build/building-effective-agents</guid>
    <description><![CDATA[ A practitioner guide whose most important argument is a restraint: give an agent the least autonomy the task needs, not the most the model can handle. ]]></description>
    <pubDate>Tue, 18 Aug 2026 00:00:00 GMT</pubDate>
  </item><item>
    <title>Llama Guard</title>
    <link>https://behaviorlayer.ai/build/llama-guard</link>
    <guid>https://behaviorlayer.ai/build/llama-guard</guid>
    <description><![CDATA[ A separate model whose only job is to read what goes into and comes out of another model and classify it as safe or unsafe against a defined policy. ]]></description>
    <pubDate>Tue, 18 Aug 2026 00:00:00 GMT</pubDate>
  </item><item>
    <title>Anthropic Reports That Benign Agent Habits Compound Into Systemic Failures</title>
    <link>https://behaviorlayer.ai/signal/anthropic-multiagent-failure-patterns</link>
    <guid>https://behaviorlayer.ai/signal/anthropic-multiagent-failure-patterns</guid>
    <description><![CDATA[ Anthropic’s Frontier Red Team published experiments on 2026-08-13 in which swarms of Claude agents collude on price, converge on identical decisions, fail to surface... ]]></description>
    <pubDate>Tue, 18 Aug 2026 00:00:00 GMT</pubDate>
  </item><item>
    <title>Agents Trading With Each Other Produce Misaligned Messages Without Anyone Eliciting Them</title>
    <link>https://behaviorlayer.ai/signal/multi-agent-commerce-misaligned-communication</link>
    <guid>https://behaviorlayer.ai/signal/multi-agent-commerce-misaligned-communication</guid>
    <description><![CDATA[ A paper submitted to arXiv on 2026-08-14 classifies 2,583 emails that frontier-model agents sent each other while competing in a year-long simulated vending business... ]]></description>
    <pubDate>Tue, 18 Aug 2026 00:00:00 GMT</pubDate>
  </item><item>
    <title>OpenAI Ships a Model Spec Revision That Makes the User&#039;s Mental Model the Assistant&#039;s Job</title>
    <link>https://behaviorlayer.ai/signal/openai-model-spec-2026-08-18</link>
    <guid>https://behaviorlayer.ai/signal/openai-model-spec-2026-08-18</guid>
    <description><![CDATA[ model-spec.openai.com now serves a version dated 2026-08-18, which adds a new guideline making the accuracy of the user’s mental model of the assistant an explicit b... ]]></description>
    <pubDate>Tue, 18 Aug 2026 00:00:00 GMT</pubDate>
  </item><item>
    <title>About</title>
    <link>https://behaviorlayer.ai/about</link>
    <guid>https://behaviorlayer.ai/about</guid>
    <description><![CDATA[ This page records who maintains The Behavioral Layer and how its notes are made. Attribution here is provenance, not promotion. ]]></description>
    <pubDate>Sun, 09 Aug 2026 18:51:46 GMT</pubDate>
  </item><item>
    <title>Four Weeks Down the Stack</title>
    <link>https://behaviorlayer.ai/briefings/four-weeks-down-the-stack</link>
    <guid>https://behaviorlayer.ai/briefings/four-weeks-down-the-stack</guid>
    <description><![CDATA[ This digest covers four weeks of signal, July 13 through August 3, in one sitting. ]]></description>
    <pubDate>Sun, 09 Aug 2026 00:00:00 GMT</pubDate>
  </item>
    </channel>
  </rss>