Learn the latest in adversary tradecraft with SpecterOps

Our hands-on Codex crash course for security practitioners will show you how to harness Codex for defense, evolving from a coding assistant to a force multiplier for defensive teams. Built by security practitioners for security practitioners, this session shows what modern detection engineering can look like when defenders embrace Codex.

Frontier models are reshaping what cyber defenders can accomplish, and the security leaders deploying them are learning where these capabilities deliver value and where human expertise remains essential. In this chat, SpecterOps’ Head of Services Robby Winchester sits down with OpenAI’s Cyber Lead Clint Gibler in a conversation moderated by Dave Bittner, Host of the CyberWire Daily, to examine the practical realities of building defense around frontier models.

The conversation moves past the hype to address the questions CISOs are actually facing. How should defenders think about the division of labor between AI and human judgment? Where do frontier models create the most leverage for defenders? What does responsible, scoped deployment of advanced models look like inside real security workflows? And how are partnerships between AI labs and security practitioners shaping the standards for safe access?

Drawing on OpenAI's perspective as a model developer and SpecterOps' adversary-focused tradecraft, this discussion will offer attendees a grounded and actionable view of what frontier models are capable of and how they fit into a defensible security strategy.

Frontier AI agents are moving beyond short, self-contained hacking tasks, but most benchmarks still look like CTFs. In this talk, we'll examine a different approach: using realistic cyber ranges to evaluate how these agents perform in long-horizon offensive scenarios.

Built to mirror modern enterprise environments, the ranges include defensive telemetry, realistic network complexity, and the multi-system traversal that actual attacks require. Philippos Giavridis (UK AI Security Institute) will be joined by Luke Paine (SpecterOps) to showcase the ways UK AI Security Institute leverages these ranges, built in collaboration with SpecterOps, to reveal model bottlenecks, map agent behavior to real-world TTPs, and support more rigorous cyber risk assessments.

This session is expected to fill quickly; book your place today and bring your questions!

How far can an attacker get across AWS accounts with the right escalation primitives? Built to emulate a corporate cloud environment, this hands-on workshop walks you through the answer, covering IAM, EC2, SSM, EKS, CloudFormation, KMS, Lambda, and S3, and culminating in a capstone exercise simulating an attack path across three AWS accounts with conditional access controls. Learn how to use AWSHound to analyze and identify these attack paths, and take it back to use in your own environment. 

How far can an attacker get across AWS accounts with the right escalation primitives? Built to emulate a corporate cloud environment, this hands-on workshop walks you through the answer, covering IAM, EC2, SSM, EKS, CloudFormation, KMS, Lambda, and S3, and culminating in a capstone exercise simulating an attack path across three AWS accounts with conditional access controls. Learn how to use AWSHound to analyze and identify these attack paths, and take it back to use in your own environment.

Want to learn how to graph attack paths for a new technology in BloodHound or already working on your own OpenGraph collector? Join us for a brief intro to OpenGraph, followed by a short hands-on lab and demo, extended Q&A, and DIY development time guided by our team of researchers to equip you with the tools and experience you need to visualize new attack paths on the fly during offensive and defensive operations. Bring your own idea and we’ll help you get it into the graph!

Our hands-on Codex crash course for security practitioners will show you how to harness Codex for defense, evolving from a coding assistant to a force multiplier for defensive teams. Built by security practitioners for security practitioners, this session shows what modern detection engineering can look like when defenders embrace Codex.

How far can an attacker get across AWS accounts with the right escalation primitives? Built to emulate a corporate cloud environment, this hands-on workshop walks you through the answer, covering IAM, EC2, SSM, EKS, CloudFormation, KMS, Lambda, and S3, and culminating in a capstone exercise simulating an attack path across three AWS accounts with conditional access controls. Learn how to use AWSHound to analyze and identify these attack paths, and take it back to use in your own environment.