<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>Forem: mcp-fortress</title>
    <description>The latest articles on Forem by mcp-fortress (@mcp-fortress).</description>
    <link>https://forem.com/mcp-fortress</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F3629304%2F73857fbb-632f-41c3-a139-f73d8bc383ff.png</url>
      <title>Forem: mcp-fortress</title>
      <link>https://forem.com/mcp-fortress</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://forem.com/feed/mcp-fortress"/>
    <language>en</language>
    <item>
      <title>I Scanned 13 Popular MCP Servers. Here's What I Found. 🔐</title>
      <dc:creator>mcp-fortress</dc:creator>
      <pubDate>Tue, 25 Nov 2025 20:51:09 +0000</pubDate>
      <link>https://forem.com/mcp-fortress/i-scanned-13-popular-mcp-servers-heres-what-i-found-3i3l</link>
      <guid>https://forem.com/mcp-fortress/i-scanned-13-popular-mcp-servers-heres-what-i-found-3i3l</guid>
      <description>&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Farticles%2Fykn49kfeuyith9acwz7c.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Farticles%2Fykn49kfeuyith9acwz7c.png" alt="mcp security scan report" width="800" height="564"&gt;&lt;/a&gt;&lt;/p&gt;




&lt;p&gt;Model Context Protocol (MCP) servers are becoming essential tools for AI workflows. But with great power comes great security risk.&lt;/p&gt;

&lt;p&gt;I just finished scanning 13 of the most popular MCP servers using &lt;a href="https://github.com/mcp-fortress/mcp-fortress" rel="noopener noreferrer"&gt;mcp-fortress&lt;/a&gt;, an open-source security scanner I built for the MCP ecosystem.&lt;/p&gt;

&lt;p&gt;Here's what every MCP user needs to know.&lt;/p&gt;

&lt;h2&gt;
  
  
  🚨 The First Malicious MCP Server
&lt;/h2&gt;

&lt;p&gt;Before I even started scanning, security researchers at &lt;a href="https://semgrep.dev/blog/2025/so-the-first-malicious-mcp-server-has-been-found-on-npm-what-does-this-mean-for-mcp-security/" rel="noopener noreferrer"&gt;Semgrep&lt;/a&gt; and &lt;a href="https://snyk.io/blog/malicious-mcp-server-on-npm-postmark-mcp-harvests-emails/" rel="noopener noreferrer"&gt;Snyk&lt;/a&gt; discovered &lt;strong&gt;postmark-mcp&lt;/strong&gt; - the first confirmed malicious MCP server on npm.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;What it did:&lt;/strong&gt; Added a hidden BCC to all emails sent through AI agents, silently harvesting every email.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Why it matters:&lt;/strong&gt; This proves MCP servers are already being weaponized. It won't be the last.&lt;/p&gt;

&lt;h2&gt;
  
  
  📊 What I Scanned
&lt;/h2&gt;

&lt;p&gt;I scanned 13 packages including:&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Official Anthropic Servers:&lt;/strong&gt;&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;@modelcontextprotocol/server-filesystem&lt;/li&gt;
&lt;li&gt;@modelcontextprotocol/server-puppeteer&lt;/li&gt;
&lt;li&gt;@modelcontextprotocol/server-github&lt;/li&gt;
&lt;li&gt;@modelcontextprotocol/server-slack&lt;/li&gt;
&lt;li&gt;...and more&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;&lt;strong&gt;Popular Third-Party Servers:&lt;/strong&gt;&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;mcp-framework (80+ projects depend on it)&lt;/li&gt;
&lt;li&gt;mcp-use (complete MCP framework)&lt;/li&gt;
&lt;li&gt;@contentful/mcp-server&lt;/li&gt;
&lt;li&gt;...and more&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  ✅ The Good News
&lt;/h2&gt;

&lt;p&gt;All official @modelcontextprotocol servers are exceptionally secure:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;0 CVEs found&lt;/li&gt;
&lt;li&gt;Minimal dependencies (&amp;lt;15 for most)&lt;/li&gt;
&lt;li&gt;No suspicious patterns&lt;/li&gt;
&lt;li&gt;Regular updates&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;&lt;strong&gt;Verdict:&lt;/strong&gt; Trust and use official servers with confidence.&lt;/p&gt;

&lt;h2&gt;
  
  
  ⚠️ The Concerning Patterns
&lt;/h2&gt;

&lt;h3&gt;
  
  
  Undisclosed Telemetry
&lt;/h3&gt;

&lt;p&gt;&lt;strong&gt;mcp-use&lt;/strong&gt; makes 6 external network requests to:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;PostHog (analytics)&lt;/li&gt;
&lt;li&gt;Langfuse (LLM tracking)&lt;/li&gt;
&lt;li&gt;Scarf (package analytics)&lt;/li&gt;
&lt;li&gt;Vendor cloud&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;&lt;strong&gt;Issue:&lt;/strong&gt; Not disclosed in documentation. Users should know their AI interactions are being tracked.&lt;/p&gt;

&lt;h3&gt;
  
  
  Dependency Risk
&lt;/h3&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Package&lt;/th&gt;
&lt;th&gt;Dependencies&lt;/th&gt;
&lt;th&gt;Risk&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;mcp-use&lt;/td&gt;
&lt;td&gt;47&lt;/td&gt;
&lt;td&gt;Medium&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;@modelcontextprotocol/sdk&lt;/td&gt;
&lt;td&gt;37&lt;/td&gt;
&lt;td&gt;Low&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;mcp-framework&lt;/td&gt;
&lt;td&gt;27&lt;/td&gt;
&lt;td&gt;Low&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;More dependencies = more supply chain attack surface.&lt;/p&gt;

&lt;h2&gt;
  
  
  🛡️ How to Protect Yourself
&lt;/h2&gt;

&lt;h3&gt;
  
  
  1. Scan Before Installing
&lt;/h3&gt;



&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;&lt;span class="c"&gt;# Quick scan&lt;/span&gt;
npx mcp-fortress scan @modelcontextprotocol/server-github

&lt;span class="c"&gt;# Install tool&lt;/span&gt;
npm &lt;span class="nb"&gt;install&lt;/span&gt; &lt;span class="nt"&gt;-g&lt;/span&gt; mcp-fortress
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;h3&gt;
  
  
  2. Check These Red Flags
&lt;/h3&gt;

&lt;p&gt;🚩 Published &amp;lt; 1 week ago&lt;br&gt;
🚩 0 downloads or GitHub stars&lt;br&gt;
🚩 No source code repository&lt;br&gt;
🚩 Typosquatting names&lt;br&gt;
🚩 300+ dependencies&lt;br&gt;
🚩 Undisclosed external requests&lt;/p&gt;
&lt;h3&gt;
  
  
  3. Use Sandbox Environments
&lt;/h3&gt;

&lt;p&gt;Test new MCP servers in isolated VMs/containers before giving production access.&lt;/p&gt;
&lt;h2&gt;
  
  
  🎯 Key Takeaways
&lt;/h2&gt;

&lt;ol&gt;
&lt;li&gt;
&lt;strong&gt;Official servers are safe&lt;/strong&gt; - use them with confidence&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Third-party servers need scrutiny&lt;/strong&gt; - scan before installing&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Malicious servers exist&lt;/strong&gt; - postmark-mcp proves this&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Always scan packages&lt;/strong&gt; - &lt;code&gt;npx mcp-fortress scan &amp;lt;package&amp;gt;&lt;/code&gt;
&lt;/li&gt;
&lt;/ol&gt;
&lt;h2&gt;
  
  
  🌐 What Needs to Happen
&lt;/h2&gt;

&lt;p&gt;The MCP ecosystem needs:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Security guidelines for developers&lt;/li&gt;
&lt;li&gt;Vetting process for popular servers&lt;/li&gt;
&lt;li&gt;Transparency requirements (disclose telemetry)&lt;/li&gt;
&lt;li&gt;Built-in scanning in Claude Desktop, Cursor, etc.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;We're in the "npm 10 years ago" phase. Let's not repeat the same security mistakes.&lt;/p&gt;
&lt;h2&gt;
  
  
  🏰 About MCP Fortress
&lt;/h2&gt;

&lt;p&gt;mcp-fortress is a free, open-source security scanner for MCP servers that works &lt;strong&gt;three ways&lt;/strong&gt;:&lt;/p&gt;
&lt;h3&gt;
  
  
  1️⃣ CLI Tool
&lt;/h3&gt;


&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;npm &lt;span class="nb"&gt;install&lt;/span&gt; &lt;span class="nt"&gt;-g&lt;/span&gt; mcp-fortress
mcp-fortress scan @modelcontextprotocol/server-github
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;

&lt;h3&gt;
  
  
  2️⃣ MCP Server (Scan from Claude Desktop)
&lt;/h3&gt;

&lt;p&gt;Add to your &lt;code&gt;claude_desktop_config.json&lt;/code&gt;:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight json"&gt;&lt;code&gt;&lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="w"&gt;
  &lt;/span&gt;&lt;span class="nl"&gt;"mcpServers"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="w"&gt;
    &lt;/span&gt;&lt;span class="nl"&gt;"mcp-fortress"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="w"&gt;
      &lt;/span&gt;&lt;span class="nl"&gt;"command"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="s2"&gt;"mcp-fortress"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;&lt;span class="w"&gt;
      &lt;/span&gt;&lt;span class="nl"&gt;"args"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;[&lt;/span&gt;&lt;span class="s2"&gt;"serve-mcp"&lt;/span&gt;&lt;span class="p"&gt;]&lt;/span&gt;&lt;span class="w"&gt;
    &lt;/span&gt;&lt;span class="p"&gt;}&lt;/span&gt;&lt;span class="w"&gt;
  &lt;/span&gt;&lt;span class="p"&gt;}&lt;/span&gt;&lt;span class="w"&gt;
&lt;/span&gt;&lt;span class="p"&gt;}&lt;/span&gt;&lt;span class="w"&gt;
&lt;/span&gt;&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Now you can ask Claude: &lt;em&gt;"Scan the puppeteer MCP server for vulnerabilities"&lt;/em&gt;&lt;/p&gt;

&lt;h3&gt;
  
  
  3️⃣ Claude Code Plugin
&lt;/h3&gt;

&lt;p&gt;Install the MCP Fortress plugin in Claude Code to scan MCP servers directly from your coding environment.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Features:&lt;/strong&gt;&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;✅ Scans for CVEs via OSV database&lt;/li&gt;
&lt;li&gt;✅ Detects dangerous code patterns (eval, exec, etc.)&lt;/li&gt;
&lt;li&gt;✅ Finds undisclosed external requests&lt;/li&gt;
&lt;li&gt;✅ Analyzes dependency risks&lt;/li&gt;
&lt;li&gt;✅ Scores security risk&lt;/li&gt;
&lt;li&gt;✅ Detects prompt injection attempts&lt;/li&gt;
&lt;li&gt;✅ Identifies tool poisoning&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;&lt;strong&gt;Use it however you want:&lt;/strong&gt;&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;CLI:&lt;/strong&gt; Quick terminal scans&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;MCP Server:&lt;/strong&gt; Integrate into your AI workflow&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Claude Code Plugin:&lt;/strong&gt; Scan while you code&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;&lt;strong&gt;Links:&lt;/strong&gt;&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;GitHub:&lt;/strong&gt; &lt;a href="https://github.com/mcp-fortress/mcp-fortress" rel="noopener noreferrer"&gt;https://github.com/mcp-fortress/mcp-fortress&lt;/a&gt;
&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;npm:&lt;/strong&gt; &lt;a href="https://www.npmjs.com/package/mcp-fortress" rel="noopener noreferrer"&gt;https://www.npmjs.com/package/mcp-fortress&lt;/a&gt;
&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;MCP Registry:&lt;/strong&gt; &lt;a href="https://registry.modelcontextprotocol.io" rel="noopener noreferrer"&gt;https://registry.modelcontextprotocol.io&lt;/a&gt;
&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Smithery (Remote):&lt;/strong&gt; &lt;a href="https://smithery.ai/server/@mcp-fortress/mcp-fortress-server" rel="noopener noreferrer"&gt;https://smithery.ai/server/@mcp-fortress/mcp-fortress-server&lt;/a&gt;
&lt;/li&gt;
&lt;/ul&gt;




&lt;p&gt;What do you think about MCP security? Share your thoughts in the comments!&lt;/p&gt;

&lt;p&gt;&lt;em&gt;If you found this valuable, star the repo and spread the word. Security is a team sport.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>security</category>
      <category>mcp</category>
      <category>ai</category>
      <category>claudecode</category>
    </item>
  </channel>
</rss>
