David Bombal

#509: AI + Metasploit = Terrifyingly Easy Hacking is here (demo)

Sep 11, 2025
Guest
Kyle Winters
Kyle Winters, security practitioner who builds Metasploit integrations, demonstrates an AI-driven ethical-hacking lab. Short demos show an LLM linked to Metasploit via MCP automating recon, generating a risk report, and running exploits like VSFTPD, EternalBlue and UnrealIRCD. Conversations cover how this lowers barriers for red teams and what defenders should watch for.
Ask episode
AI Snips
Chapters
Transcript
Episode notes
INSIGHT

MCP Bridges LLMs To Real Tools

  • MCP (Model Context Protocol) can bridge LLMs to real tools like Metasploit to perform actions on systems.
  • Kyle Winters shows MCP translates natural language into tool commands, enabling AI-driven interactions with infrastructure.
ANECDOTE

Manual VSFTPD Exploit Demo

  • Kyle runs a VSFTPD backdoor exploit manually in Metasploit to get a root shell on a vulnerable host.
  • He demonstrates basic Metasploit commands and verifies root access before automating with Claude.
ANECDOTE

AI Executes VSFTPD And Writes File

  • Kyle connects Claude (LLM) via an MCP server to Metasploit and asks it to run the same VSFTPD exploit.
  • The AI runs the exploit, obtains a root shell, and creates a /tmp/bumble.txt file containing "Hello World."
Get the Snipd Podcast app to discover more snips from this episode
Get the app