AI Threat SimulationsTimed Scenarios

AI Threat Simulation Labs

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.

Filter:
Vulnerability Triage

AI Vulnerability Triage: Web Application Scan

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.

Beginner5m100 pts
Vulnerability Triage

AI Vulnerability Triage: Cloud Infrastructure Audit

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.

Intermediate7m120 pts
Exploit Chain Analysis

Exploit Chain Analysis: Browser Sandbox Escape

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.

Intermediate8m100 pts
Exploit Chain Analysis

Exploit Chain Analysis: AI-Powered Supply Chain Attack

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.

Professional10m120 pts
Patch Validation

Patch Validation: SQL Injection Remediation

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.

Beginner5m100 pts
Patch Validation

Patch Validation: Zero-Day Kernel Vulnerability

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.

Professional10m120 pts

Real-World Scenarios

Based on actual vulnerability patterns discovered by Mythos-class AI systems and documented by the Glasswing consortium.

Machine-Speed Pressure

Every simulation is timed. Learn to triage, analyze, and respond at the speed that modern AI threats demand.

Scored & Tracked

Your performance is scored and tracked. Compete on leaderboards and earn recognition for your defensive skills.