<?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>Confidence and Disclosure</title>
    <link>https://behaviorlayer.ai/behavior/confidence-and-disclosure</link>
    <guid>https://behaviorlayer.ai/behavior/confidence-and-disclosure</guid>
    <description><![CDATA[ A trustworthy agent is not the one that is always right. ]]></description>
    <pubDate>Fri, 10 Jul 2026 03:12:13 GMT</pubDate>
  </item><item>
    <title>Escalation</title>
    <link>https://behaviorlayer.ai/behavior/escalation</link>
    <guid>https://behaviorlayer.ai/behavior/escalation</guid>
    <description><![CDATA[ Escalation is the contract clause most teams write last and need first: the specified moment an agent stops acting on its own and hands control to a human, with enou... ]]></description>
    <pubDate>Fri, 10 Jul 2026 03:12:13 GMT</pubDate>
  </item><item>
    <title>Guardrails</title>
    <link>https://behaviorlayer.ai/behavior/guardrails</link>
    <guid>https://behaviorlayer.ai/behavior/guardrails</guid>
    <description><![CDATA[ Guardrails are the enforcement half of a behavioral contract. ]]></description>
    <pubDate>Fri, 10 Jul 2026 03:12:13 GMT</pubDate>
  </item><item>
    <title>Trust Scaffolding</title>
    <link>https://behaviorlayer.ai/behavior/trust-scaffolding</link>
    <guid>https://behaviorlayer.ai/behavior/trust-scaffolding</guid>
    <description><![CDATA[ Trust in an agent is not granted. ]]></description>
    <pubDate>Fri, 10 Jul 2026 03:12:13 GMT</pubDate>
  </item><item>
    <title>Specified, Measured, Governed</title>
    <link>https://behaviorlayer.ai/briefings/specified-measured-governed</link>
    <guid>https://behaviorlayer.ai/briefings/specified-measured-governed</guid>
    <description><![CDATA[ Three artifacts shipped in recent months from three organizations that were not coordinating: a protocol rebuilt for governance, an eval that scores a model against ... ]]></description>
    <pubDate>Fri, 10 Jul 2026 03:12:13 GMT</pubDate>
  </item><item>
    <title>Claude&#039;s Constitution</title>
    <link>https://behaviorlayer.ai/behavior/claudes-constitution</link>
    <guid>https://behaviorlayer.ai/behavior/claudes-constitution</guid>
    <description><![CDATA[ The normative document Anthropic trains Claude toward, published in full: “a detailed description of Anthropic’s intentions for Claude’s values and behavior” that “p... ]]></description>
    <pubDate>Fri, 10 Jul 2026 03:02:04 GMT</pubDate>
  </item><item>
    <title>Practices for Governing Agentic AI Systems</title>
    <link>https://behaviorlayer.ai/behavior/practices-for-governing-agentic-ai-systems</link>
    <guid>https://behaviorlayer.ai/behavior/practices-for-governing-agentic-ai-systems</guid>
    <description><![CDATA[ OpenAI’s white paper defining agentic AI systems and proposing seven baseline practices for keeping “agents’ operations safe and accountable,” written before the age... ]]></description>
    <pubDate>Fri, 10 Jul 2026 03:02:04 GMT</pubDate>
  </item><item>
    <title>Agent2Agent (A2A) Protocol</title>
    <link>https://behaviorlayer.ai/build/a2a-protocol</link>
    <guid>https://behaviorlayer.ai/build/a2a-protocol</guid>
    <description><![CDATA[ A2A is “an open standard designed to facilitate communication and interoperability between independent, potentially opaque AI agent systems”: a common language for a... ]]></description>
    <pubDate>Fri, 10 Jul 2026 03:02:04 GMT</pubDate>
  </item><item>
    <title>AutoGen</title>
    <link>https://behaviorlayer.ai/build/autogen</link>
    <guid>https://behaviorlayer.ai/build/autogen</guid>
    <description><![CDATA[ AutoGen is “an open-source framework that allows developers to build LLM applications via multiple agents that can converse with each other to accomplish tasks,” the... ]]></description>
    <pubDate>Fri, 10 Jul 2026 03:02:04 GMT</pubDate>
  </item><item>
    <title>NeMo Guardrails</title>
    <link>https://behaviorlayer.ai/build/nemo-guardrails</link>
    <guid>https://behaviorlayer.ai/build/nemo-guardrails</guid>
    <description><![CDATA[ NeMo Guardrails is “an open-source toolkit for easily adding programmable guardrails to LLM-based conversational systems”: constraints you write as code and enforce ... ]]></description>
    <pubDate>Fri, 10 Jul 2026 03:02:04 GMT</pubDate>
  </item>
    </channel>
  </rss>