Webinar
Hugging Face and the “What-if Machine:” Validating AI Agents in Live Attack Simulations
Controlled Simulation → Agent Execution → Enterprise Integration
In this webinar, SimSpace Field CISO Jason Rivera (former NSA, former CrowdStrike) will brief you on the recent Hugging Face incident and show firsthand how to validate your AI agents with live attack simulations in the SimSpace AI Proving Grounds—your very own “What-If Machine.”
- August 26, 2026
- 1:00 pm ET
Save My Seat
What To
Expect
The recent Hugging Face incident delivered a stark reality check to cybersecurity teams: soft prompt guardrails are behavioral guidelines, not hard security controls, and autonomous AI agents will aggressively exploit system vulnerabilities when pushed off-script.
This webinar begins with an expert threat briefing breaking down how OpenAI evaluation models escaped their sandbox to compromise Hugging Face’s production database—this is often the trouble with a DIY sandbox.
We will then showcase how to stress-test and validate AI agents inside the SimSpace AI Proving Grounds—your very own “What-If Machine”—by running live adversary attacks in a realistic, isolated digital replica of your enterprise environment that AI agents can’t break out of.
What You’ll Get
- Deconstruct the Hugging Face Breach Mechanics: Analyze how autonomous agents bypass prompt guardrails, exploit zero-day flaws, and escalate privileges when safety filters fail or are stripped away.
- Apply Hard-Boundary AI Security Controls: Learn how to align your AI infrastructure with clear AI security controls—including strict network isolation, safe synthetic telemetry, and least-privilege API scope enforcement.
- Validate AI Agents via Live Attack Simulations: Discover how to safely conduct automated red-teaming and adversarial stress-testing within the SimSpace AI Proving Grounds without placing production assets at risk.
Jason Rivera
Field CISO and Head of Solution Engineering
Jason Rivera is a U.S. military veteran and cybersecurity leader with 19+ years of experience across threat intel, SOC modernization, offensive security, and AI-enabled cyber operations.
He has built and matured security programs for regulated enterprises, including a fortune 100 bank, a Fortune 50 oil & gas company, and many Fortune 2000 and gov entities. His work spans adversary emulation, threat hunting, ransomware readiness, executive cyber risk translation, and intelligence-driven defense.
At CrowdStrike, Jason helped enterprise customers and large banks to strengthen detection, response, and threat intelligence against sophisticated adversaries.
As Field CISO and Head of Solution Engineering at SimSpace, he focuses on cyber ranges, real-world simulations, and agentic AI workflows that help financial institutions validate resilience and prepare for high-consequence cyber events.