<?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>Add-Skill on Aonan Guan</title><link>https://oddguan.com/tags/add-skill/</link><description>Recent content in Add-Skill on Aonan Guan</description><image><title>Aonan Guan</title><url>https://oddguan.com/avatar.webp</url><link>https://oddguan.com/avatar.webp</link></image><generator>Hugo -- 0.147.7</generator><language>en-us</language><lastBuildDate>Sun, 08 Mar 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://oddguan.com/tags/add-skill/index.xml" rel="self" type="application/rss+xml"/><item><title>Agent SkillSlip: Path Traversal in Google Gemini CLI, Anthropic Claude Code, and Vercel add-skill</title><link>https://oddguan.com/blog/agent-skillslip-google-gemini-cli-anthropic-claude-code-vercel-add-skill-path-traversal/</link><pubDate>Sun, 08 Mar 2026 00:00:00 +0000</pubDate><guid>https://oddguan.com/blog/agent-skillslip-google-gemini-cli-anthropic-claude-code-vercel-add-skill-path-traversal/</guid><description>Google Gemini CLI, Anthropic Claude Code, and Vercel add-skill share the same path traversal flaw: the name field in skill metadata is passed to path.join() without validation, enabling VS Code hijacking and SSH key injection — invisible to the user.</description></item></channel></rss>