CSEM

Vulnerable AI Challenge

Test your skills against AI security vulnerabilities. Each challenge has multiple steps to complete.

Level 1: Prompt Injection
1
2
3
Step 1: Basic prompt injection
Step 2: Advanced system prompt bypass
Step 3: Chain multiple injections
Start Challenge
Level 2: Data Leakage
1
2
3
Step 1: Identify error patterns
Step 2: Extract sensitive data
Step 3: Bypass security filters
Start Challenge
Level 3: API Configuration
1
2
3
Step 1: Discover API endpoints
Step 2: Extract configuration
Step 3: Exploit misconfigurations
Start Challenge
Level 4: Model Poisoning
1
2
3
Step 1: Identify trigger phrases
Step 2: Manipulate responses
Step 3: Chain manipulation attacks
Start Challenge
Level 5: Indirect Injection
1
2
3
Step 1: Find injection points
Step 2: Craft payload chain
Step 3: Execute complex attack
Start Challenge
Level 6: Output Handling
1
2
3
Step 1: Analyze output sanitization
Step 2: Bypass security controls
Step 3: Execute full compromise
Start Challenge
Challenge Guide