| Presentation | Author | Description |
|---|---|---|
| Defense-in-Depth Engineering, Blue Team Con 2024 | John Poulin | The 2021 OWASP Top Ten introduced a category “Insecure Design” to focus on risks related to design flaws. In this talk, we will focus on techniques we can use to build defense-in-depth software. What can we do to proactively architect software to be more resilient to attacks? What type of findings may not be discovered via automated static analysis? How can we design our software to be more friendly during incident response scenarios? Throughout this talk, we will focus on identifying often-overlooked architectural anti-patterns and vulnerabilities to be on the lookout for. We will source code to analyze patterns for improvement in both real-world applications as well as intentionally vulnerable applications. Engineers will leave this talk with a solid understanding of defense-in-depth software architecture and design. Security engineers or consultants can expect to leave with an increased understanding of insecure design patterns and vulnerabilities. |
| 60 Minutes to Hack: AWS | John Poulin | Gone in 60 Minutes – A one-hour checklist for pentesting AWS. The detections team is hot on your trail, and your access will be cut in 1-hour – what do you do? This talk will lay out the foundations for red teaming an AWS account in an extremely time-boxed manner. We will identify tools and techniques for exfiltrating data, maintaining persistence, and causing noise. The emphasis of this talk will be laying out things that can actually be accomplished in 1-hour. After all, time limits everything. Attendees will walk away from this talk with a functional and prioritized checklist that they can use during red team operations. |
| Attack of the Prompt | Sean Lyford | Large Language Models (LLMs) have been a hot topic of discussion and engineering since ChatGPT took the internet by storm in 2022. The field has seen a heavy investment monetarily, in engineering efforts, and even electricity consumption. The field has evolved into numerous hosted and local solutions to match business needs. To supplement LLMs, we have seen the rise of techniques that assist in creating new and improved solutions. Retrieval Augmented Generation (RAG) feeds better data into LLM contexts to generate better results, and recent agentic efforts consume LLM outputs to perform actions based on the inputs. In this talk, we will talk about LLMs and what they do. We will go into detail about what an LLM solution may look like before pivoting into attacks such as prompt injection and jailbreaking. We will then introduce RAG and agents, and what a compromised LLM could potentially do. Finally, we will finish with steps to take including better prompt engineering, better AI/ML red teaming, and introduce guardrails. |
| What Hackers See that You Don't: A Real-World Look at Vulnerabilities | Alexandria Poulin | Oftentimes, we see IT teams that work hard at keeping up their systems, whether it's maintaining patches, system updates, and scaling services. IT teams can often be left in charge of security services too, which leaves a lot on their plate and keeps them busy. Due to this, functionality can sometimes obscure vulnerabilities that threat actors may exploit. In my position at Cloud Security Partners, I have seen numerous examples of overlooked risks, such as misconfigured cloud services, insecure applications, and exposed credentials or data. During this talk, we will explore real-world examples that we have faced and discuss how minor oversights can escalate into significant security threats. Rest assured, no complicated red team jargon will be used. You’ll gain insights into how attackers leverage OSINT to gather information and leave with an understanding of the current threats along with tips on prevention. |
| Tool Name | Description |
|---|---|
| Skate | Intentionally vulnerable Ruby-on-Rails application designed to accompany Defense-in-Depth engineering training and workshops. |