Sign up to save your podcasts
Or
Share Interview with Johann Rehberger - Agentic ProbLLMs - Exploiting AI Computer-Use and Coding Agents
Share to email
Share to Facebook
Share to X
Share to email
Share to Facebook
Share to X
Or
Interview with Johann Rehberger - Agentic ProbLLMs - Exploiting AI Computer-Use and Coding Agents
Explore the cutting edge of AI security at BSides Vancouver Island 2025! In this interview, Johann Rehberger—an independent researcher with deep roots in offensive security and AI systems—previews his talk “Agentic ProbLLMs – Exploiting AI Computer-Use and Coding.” Join us as Johann demonstrates real-world prompt injection attacks against autonomous AI agents such as OpenAI Operator, Google Jules, Anthropic Claude Code, Devin, Amazon Q, and more.
In this session, learn:
Why agentic systems are uniquely vulnerable to prompt injection exploits compared to traditional LLM setups.
A surprising case study of Johann’s exploits and how advanced AI responded in the wild.
Direct, practical steps developers and defenders can take to help secure next-gen AI tools today.
Don’t miss your chance to meet Johann and other cybersecurity experts at the BSides Vancouver Island Security Conference, happening October 3, 2025 at the Victoria Conference Centre. Secure your spot at bsidesvi.com by September 19th to snag a limited-edition black hacker T-shirt and connect with the island’s premier community-driven infosec event.
✅ Get tickets: bsidesvi.com
🪙 Early Bird Offer: T-shirt until September 19
🧑💻 Join our online community: #bsidesvi
—
About BSides Vancouver Island:
BSidesVI delivers hands-on learning, insightful keynotes, and networking for cybersecurity pros and enthusiasts in Victoria, BC. All social media links and more info are at bsidesvi.com.
Be ready to learn, hack, and connect—see you at BSides Vancouver Island 2025!
#bsidesvi #AIsecurity #Cybersecurity #Conference
View all episodes
By Explore the cutting edge of AI security at BSides Vancouver Island 2025! In this interview, Johann Rehberger—an independent researcher with deep roots in offensive security and AI systems—previews his talk “Agentic ProbLLMs – Exploiting AI Computer-Use and Coding.” Join us as Johann demonstrates real-world prompt injection attacks against autonomous AI agents such as OpenAI Operator, Google Jules, Anthropic Claude Code, Devin, Amazon Q, and more.
In this session, learn:
Why agentic systems are uniquely vulnerable to prompt injection exploits compared to traditional LLM setups.
A surprising case study of Johann’s exploits and how advanced AI responded in the wild.
Direct, practical steps developers and defenders can take to help secure next-gen AI tools today.
Don’t miss your chance to meet Johann and other cybersecurity experts at the BSides Vancouver Island Security Conference, happening October 3, 2025 at the Victoria Conference Centre. Secure your spot at bsidesvi.com by September 19th to snag a limited-edition black hacker T-shirt and connect with the island’s premier community-driven infosec event.
✅ Get tickets: bsidesvi.com
🪙 Early Bird Offer: T-shirt until September 19
🧑💻 Join our online community: #bsidesvi
—
About BSides Vancouver Island:
BSidesVI delivers hands-on learning, insightful keynotes, and networking for cybersecurity pros and enthusiasts in Victoria, BC. All social media links and more info are at bsidesvi.com.
Be ready to learn, hack, and connect—see you at BSides Vancouver Island 2025!
#bsidesvi #AIsecurity #Cybersecurity #Conference