People behind the signal.

Agentic cybersecurity, security automation, and trustworthy AI systems.

Adaptive leadership, AI-enabled learning, and future-ready cybersecurity talent.

Starlark is a Python dialect mainly used as a configuration language. Its interoperability with Go makes it a natural choice for Go-based C2 agents. This talk explains the Starlark module system in emp3r0r C2, demonstrates detection evasion compared with traditional BOF modules, and shows how agents can be scripted on the fly.

AI agents connect language models to tools, data and workflows, creating opportunities for attackers to influence their decisions and actions. This session explores how to red team these systems through practical examples of prompt injection, trust boundary failures and tool abuse. We’ll examine how to assess the impact of successful attacks and apply controls that reduce risk while preserving useful capabilities.

How do you turn threat intelligence into an investigation you can run and a detection you can test? This session demonstrates a practical Jupyter notebook workflow for moving from threat reporting and indicators to evidence-driven investigation and detection hypotheses.

Using a worked example, I’ll show how to organize intelligence, examine supporting telemetry, map observed behavior to MITRE ATT&CK, and develop detection logic. We’ll also explore how AI can assist with analysis and documentation while keeping conclusions grounded in source evidence.

The session draws on my experience delivering DEF CON training, developing CTF challenges, and teaching security operations. It is designed for cybersecurity practitioners, researchers, students, and CTF players.