<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>leeky</title><description>Application and AI security. Writeups on vulnerability discovery at scale.</description><link>https://leeky.sh/</link><item><title>Building an LLM Vulnerability Scanner That Doesn&apos;t Drown You in False Positives</title><link>https://leeky.sh/writing/web3-security-scanner/</link><guid isPermaLink="true">https://leeky.sh/writing/web3-security-scanner/</guid><description>Static analysis finds patterns, LLMs find intent, and neither survives alone. So I built a scanner where five pipelines hunt and an adversarial validator kills what they get wrong.</description><pubDate>Mon, 20 Jul 2026 00:00:00 GMT</pubDate></item><item><title>I Vibecoded a Medicare Fraud Detector in a Weekend to See How Hard It Actually Is</title><link>https://leeky.sh/writing/medfraud-medicare-detection/</link><guid isPermaLink="true">https://leeky.sh/writing/medfraud-medicare-detection/</guid><description>2025 and 2026 turned government fraud into a headline argument with numbers ranging from tens of millions to nine billion for the same programs. The data behind it is public, so I built a screening system over 47 million rows to find out what it can actually show.</description><pubDate>Tue, 30 Jun 2026 00:00:00 GMT</pubDate></item></channel></rss>