<?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>Prompt Injection on Vinh Thang's AI Tech Notes</title><link>https://vinhthang.dev/tags/prompt-injection/</link><description>Recent content in Prompt Injection on Vinh Thang's AI Tech Notes</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Sat, 05 Sep 2026 23:50:00 +0700</lastBuildDate><atom:link href="https://vinhthang.dev/tags/prompt-injection/index.xml" rel="self" type="application/rss+xml"/><item><title>Building an Autonomous AI Code Reviewer: Lessons in Prompt Injection and Orchestration</title><link>https://vinhthang.dev/posts/autonomous-ai-code-reviewer/</link><pubDate>Sat, 05 Sep 2026 23:50:00 +0700</pubDate><guid>https://vinhthang.dev/posts/autonomous-ai-code-reviewer/</guid><description>&lt;p&gt;What happens when you unleash a highly autonomous, multi-model AI agent to peer-review its own codebase?&lt;/p&gt;
&lt;p&gt;Over the past few weeks, I built the &lt;code&gt;ai-review-plugin&lt;/code&gt;—an Antigravity plugin designed to enforce rigorous, multi-model peer reviews on code changes before they are merged. The goal was simple: use a smaller, faster model (Model A) to orchestrate the Git workflow, and use a massive, frontier-class model (Model B - Codex) to act as a ruthless Principal Engineer reviewing the code.&lt;/p&gt;</description></item></channel></rss>