What happens when AI moves beyond answering pentesting questions and starts actively participating in the testing workflow?
In this episode of Hacker in the Loop, Phillip Wylie is joined by his colleague Jacob Krell for a hands-on look at building an agentic hacking harness using Cursor, Metasploit, MCPs, AI models, and custom guardrails. Jacob walks through the process in real time, showing how an AI-powered environment can research techniques, build tools, maintain engagement state, execute testing workflows, troubleshoot problems, and help a penetration tester work significantly faster.
Rather than treating AI as a replacement for the penetration tester, the discussion focuses on using it as an acceleration layer. Jacob demonstrates how a properly designed harness can handle repetitive testing activities while leaving the human operator responsible for direction, judgment, and higher-value investigation.
GitHub for Metasploit Cursor Harness: https://github.com/Suzu-Testing/metasploit-cursor-harness
Connect with Jacob Krell:
LinkedIn: https://www.linkedin.com/in/jacob-krell/
X: https://x.com/hackerfren
=========================
Connect with your host, Phillip Wylie:
LinkedIn: https://linkedin.com/in/phillipwylie
YouTube: https://youtube.com/@PhillipWylie
=========================
Presented by Suzu Labs
=========================
All the ways to connect with @Suzulabs
https://suzulabs.com
https://x.com/suzulabs
https://www.linkedin.com/company/suzu-labs/