About KageX
The Offense-FirstAI Security Company
KageX builds the training, assessment, and automation layer for AI security teams. The best defense starts with offense, but the real gap is repeatability: most teams still don't have a clean way to learn, validate, and operationalize AI security.
// mission
Make every team that ships AI capable of defending it.
Trainable. Testable. Measurable. KageX gives them the workflow to find out whether their people, products, and processes hold up under adversarial pressure.
// 03 reasons
Why now.
Reason 01
AI ships faster than security can adapt.
Teams are deploying copilots, agents, and internal LLM workflows without the same maturity they expect from traditional application security.
Reason 02
The market is fragmented.
Most offerings solve only one slice: training content, point-in-time testing, or red-team services. KageX is building across the full lifecycle.
Reason 03
The category needs workflow, not demos.
The product stack is designed to move from awareness to validated skill to repeatable security execution.
// 03 products
What we build.
Three products. One funnel. Free CTF on top, paid hosted labs in the middle, open-source automation underneath.
Hosted labs
FreakLabs
Hosted cloud labs for hands-on AI vulnerability training. Real models, real weak apps, and a managed environment teams can use without setup friction.
Free CTF
SentinelGoat
Free AI security CTF with progressive challenge design. A fast top-of-funnel product for onboarding, community growth, and skill validation.
Open source
AgentBreaker
Probe-first red-teaming engine that profiles the system surface and generates targeted campaigns. Built for security engineers who need repeatable results.
// 03 principles
Our approach.
01 · Product
Product-led where it matters.
SentinelGoat is free, FreakLabs is hosted and paid, and AgentBreaker is open source. Together they form an acquisition, training, and operational pipeline instead of isolated products.
02 · Practice
Offense informs defense.
You can't defend what you don't understand. The tools teach security teams how AI systems actually break, not how they theoretically might.
03 · Standards
Standards-aligned.
Everything maps to OWASP LLM Top 10 and MITRE ATLAS. Enterprise teams get compliance-ready reporting. Researchers get a shared taxonomy.
// closing
Built on practice.
KageX exists because there's a gap between shipping AI and knowing how it fails. Every product — labs, challenges, tooling — comes out of years of breaking and defending AI systems in production. Offensive security, AI/ML engineering, and developer tooling, fused into one stack.