<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Fulvio Favilla</title><description>Blog posts and security writeups.</description><link>https://fulvio.sh/</link><item><title>HTB Web | Spookifier: not all fonts are equal</title><link>https://fulvio.sh/writeups/htb-spookifier/</link><guid isPermaLink="true">https://fulvio.sh/writeups/htb-spookifier/</guid><description>A Halloween name generator with four spooky fonts. The scariest thing about it was the sanitization.</description><pubDate>Sat, 23 May 2026 00:00:00 GMT</pubDate></item><item><title>HTB Web | Notebook Converter Pro: two CVEs, one flag</title><link>https://fulvio.sh/writeups/htb-notebookconverterpro/</link><guid isPermaLink="true">https://fulvio.sh/writeups/htb-notebookconverterpro/</guid><description>A Jupyter notebook conversion service running a vulnerable nbconvert. Neither CVE alone gets the flag.</description><pubDate>Fri, 15 May 2026 00:00:00 GMT</pubDate></item><item><title>HTB Web | Flag Command: apparently a non-winnable game</title><link>https://fulvio.sh/writeups/htb-flagcommand/</link><guid isPermaLink="true">https://fulvio.sh/writeups/htb-flagcommand/</guid><description>A horror D&amp;D text adventure where every path leads to death. The way out was never inside the game.</description><pubDate>Thu, 07 May 2026 00:00:00 GMT</pubDate></item><item><title>HTB Web | ReactOOPS: a static UI hiding an RCE</title><link>https://fulvio.sh/writeups/htb-reactoops/</link><guid isPermaLink="true">https://fulvio.sh/writeups/htb-reactoops/</guid><description>A static Next.js interface with no inputs and no obvious attack surface. The vulnerability was in the framework itself.</description><pubDate>Thu, 30 Apr 2026 00:00:00 GMT</pubDate></item><item><title>Not all skills are written down: building skill extraction for ResumeRadar</title><link>https://fulvio.sh/blog/llm-vs-keyword-extraction/</link><guid isPermaLink="true">https://fulvio.sh/blog/llm-vs-keyword-extraction/</guid><description>Building ResumeRadar forced me to figure out exactly where LLMs outperform keyword matching for skill extraction, and where they don&apos;t.</description><pubDate>Tue, 28 Apr 2026 00:00:00 GMT</pubDate></item></channel></rss>