<?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>Permission-Gate on 兜忆轩 | Cedric's Compiler Notes</title><link>https://douyixuan.github.io/tags/permission-gate/</link><description>Recent content in Permission-Gate on 兜忆轩 | Cedric's Compiler Notes</description><image><title>兜忆轩 | Cedric's Compiler Notes</title><url>https://douyixuan.github.io/images/papermod-cover.png</url><link>https://douyixuan.github.io/images/papermod-cover.png</link></image><generator>Hugo -- 0.146.0</generator><language>en-us</language><lastBuildDate>Tue, 03 Nov 2026 08:00:00 +0800</lastBuildDate><atom:link href="https://douyixuan.github.io/tags/permission-gate/index.xml" rel="self" type="application/rss+xml"/><item><title>AI Native Digest 2026-11-03</title><link>https://douyixuan.github.io/posts/ai-native/2026-11-03/</link><pubDate>Tue, 03 Nov 2026 08:00:00 +0800</pubDate><guid>https://douyixuan.github.io/posts/ai-native/2026-11-03/</guid><description>两篇 10 月 29 日的原创论文从命令执行前后研究 Coding Agent 的可信控制：Caper 在快照沙箱内预执行命令，根据真实副作用审批；Soei 将 Agent 行为、证据与完成声明分开记录，构建可核验的跨会话执行链。</description></item></channel></rss>