KageX

About KageX

KageX builds open-source tools for AI security testing. We believe the best defense starts with offense -- understanding how AI systems break is the first step to making them resilient.

Our Mission

Make AI security testable, teachable, and measurable. Every company shipping an AI product needs to know how its guardrails perform under adversarial pressure. KageX gives them the tools to find out.

What We Build

FreakLab

Hands-on vulnerability labs covering the OWASP LLM Top 10. Real AI models, real attack vectors, real exploits. Learn by breaking.

SentinelGoat

AI-focused Capture The Flag with progressive difficulty. Free to play. Validates that your team can think like an AI attacker.

AgentBreaker

Autonomous red-teaming harness. 14 attack strategies, LLM-guided campaigns, judge-based scoring. Open source. CLI-first.

Our Approach

Open source first. AgentBreaker and FreakLab are open source. We believe security tools should be accessible to everyone. The community makes them better.

Offense informs defense. You can't defend what you don't understand. Our tools teach security teams how AI systems actually break, not how they theoretically might.

Standards-aligned. Everything maps to OWASP LLM Top 10 and MITRE ATLAS. Enterprise teams get compliance-ready reporting. Researchers get a shared taxonomy.

Contact

For enterprise inquiries, partnerships, or speaking engagements:

hello@kagex.ai