<?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>CI on yishiashia's Blog</title><link>https://yishiashia.github.io/tags/ci/</link><description>Recent content in CI on yishiashia's Blog</description><generator>Hugo -- gohugo.io</generator><language>zh-TW</language><copyright>Copyright © 2023 yishiashia@gmail.com. All rights reserved.</copyright><lastBuildDate>Thu, 10 Sep 2026 22:53:00 +0800</lastBuildDate><atom:link href="https://yishiashia.github.io/tags/ci/index.xml" rel="self" type="application/rss+xml"/><item><title>[AI 軟體工程] 驗收即證據：LLM 審查不能單獨放行，審查通過也不等於可以合併</title><link>https://yishiashia.github.io/posts/reproducible-verification-and-attestation/</link><pubDate>Thu, 10 Sep 2026 22:53:00 +0800</pubDate><guid>https://yishiashia.github.io/posts/reproducible-verification-and-attestation/</guid><description>在 harness engineering 的信任鏈裡，驗收分成兩半：機器產出的「可重現證據」與獨立審查者或審查機制進行的「獨立確認」。本文說明 LLM review 的適當邊界、如何把 Evidence 綁定到確切版本，以及為什麼「審查通過」不等於可以合併或上線。</description></item></channel></rss>