Train under pressure. Triage AI-generated vulnerabilities, analyze exploit chains, and validate patches — all against the clock. These scenarios simulate the speed and complexity of Mythos-class AI attacks.
Based on real-world attack patterns documented by Anthropic's Mythos red team assessments and the Glasswing consortium.
An AI scanner has identified 8 vulnerabilities in a production web application. You have 5 minutes to classify each by severity (Critical/High/Medium/Low), determine which are true positives vs false positives, and prioritize the remediation order. Speed and accuracy both count.
A Glasswing-powered scanner has flagged 10 findings across your AWS cloud infrastructure. Triage each finding, identify false positives, and create a prioritized remediation plan. You have 7 minutes.
A Mythos-class AI has discovered a 4-stage exploit chain that escapes a browser sandbox. Analyze each stage, identify the vulnerability classes, map the MITRE ATT&CK techniques, and determine the correct detection points. You have 8 minutes.
An AI agent has autonomously compromised an open-source package to create a supply chain attack. Trace the 5-stage attack, identify indicators of compromise, and map the kill chain. You have 10 minutes.
A developer has submitted 6 code patches to fix SQL injection vulnerabilities found by an AI scanner. Review each patch and determine if it correctly fixes the vulnerability, introduces new issues, or is incomplete. You have 5 minutes.
A Mythos-class AI discovered a zero-day in the Linux kernel's memory management subsystem. Three competing patches have been proposed. Analyze each for correctness, performance impact, and completeness. You have 10 minutes.
Based on actual vulnerability patterns discovered by Mythos-class AI systems and documented by the Glasswing consortium.
Every simulation is timed. Learn to triage, analyze, and respond at the speed that modern AI threats demand.
Your performance is scored and tracked. Compete on leaderboards and earn recognition for your defensive skills.