<?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>MCP on IamWin's Blog</title><link>https://blog.iamwin.xyz/tags/mcp/</link><description>Recent content in MCP on IamWin's Blog</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Thu, 06 Aug 2026 00:47:30 -0500</lastBuildDate><atom:link href="https://blog.iamwin.xyz/tags/mcp/index.xml" rel="self" type="application/rss+xml"/><item><title>Mythic C2 MCP: Integrating AI with Command and Control Operations</title><link>https://blog.iamwin.xyz/posts/mcpformythicc2/</link><pubDate>Thu, 06 Aug 2026 00:47:30 -0500</pubDate><guid>https://blog.iamwin.xyz/posts/mcpformythicc2/</guid><description>&lt;h2 id="introduction"&gt;Introduction&lt;/h2&gt;
&lt;p&gt;A few days ago I was working through an AI Offensive course where I covered several interesting topics. Some of them I had already explored on my own out of curiosity, but studying a subject more than once reinforces the concepts and improves retention. In this case, I will focus on the integration of a Mythic C2 MCP server with &lt;strong&gt;OpenCode&lt;/strong&gt;, my current LLM of choice.&lt;/p&gt;
&lt;p&gt;I previously used Claude — the course itself uses it — but I repeatedly ran into guardrails that block anything related to C2 infrastructure, implants, malware, and similar topics, which limits the learning experience considerably. OpenCode does not impose the same restrictions, making it the tool I rely on most for assistance with tooling syntax, language-specific questions, payload concepts, and similar offensive research tasks.&lt;/p&gt;</description></item><item><title>BloodHound MCP: Integrating Claude with BloodHound</title><link>https://blog.iamwin.xyz/posts/mcpbloodhoundwithclaude/</link><pubDate>Fri, 03 Jul 2026 12:00:00 -0500</pubDate><guid>https://blog.iamwin.xyz/posts/mcpbloodhoundwithclaude/</guid><description>&lt;h2 id="introduction"&gt;Introduction&lt;/h2&gt;
&lt;p&gt;About 2 weeks ago I found out about the existence of an MCP (Model Context Protocol) server for BloodHound that allows us to integrate it with Claude. Below I share the publications I read on the topic:&lt;/p&gt;
&lt;blockquote&gt;
&lt;p&gt;&lt;strong&gt;SpecterOps article:&lt;/strong&gt;
&lt;a href="https://specterops.io/blog/2026/06/18/bloodhound-mcp-one-year-later-what-i-learned-about-mcps-models-and-context/#h-introduction-the-first-version-worked-but-needed-improvement"&gt;BloodHound MCP: One Year Later — What I Learned About MCPs, Models, and Context&lt;/a&gt;&lt;/p&gt;
&lt;/blockquote&gt;
&lt;blockquote&gt;
&lt;p&gt;&lt;strong&gt;GitHub Repository:&lt;/strong&gt;
&lt;a href="https://github.com/mwnickerson/bloodhound_mcp"&gt;BloodHound MCP - GitHub Repository&lt;/a&gt;&lt;/p&gt;
&lt;/blockquote&gt;
&lt;h3 id="what-is-this"&gt;What is this?&lt;/h3&gt;
&lt;p&gt;The BloodHound MCP allows you to query your Active Directory data in natural language directly from Claude Desktop. Instead of writing Cypher queries manually, you ask Claude things like &amp;ldquo;show me all Domain Admins&amp;rdquo; or &amp;ldquo;find kerberoastable users.&amp;rdquo;&lt;/p&gt;</description></item></channel></rss>