# KageX

> KageX is an offense-first AI security company. We build hands-on training and
> open-source tooling for testing how AI systems fail — prompt injection, RAG
> and agent attacks, and LLM red teaming.

## Products

- [FreakLabs](https://kagex.ai/products/freaklab): Hosted, hands-on AI security labs — 20 attack labs and a 9-level curriculum ending in a verifiable completion certificate. Paid, with a free tier.
- [SentinelGoat](https://kagex.ai/products/sentinelgoat): Free AI hacking challenges focused on prompt injection and testing AI behaviour.
- [AgentBreaker](https://kagex.ai/products/agentbreaker): Free, open-source AI security testing tool.
- [Mirage](https://kagex.ai/products/mirage): Free, open-source browser extension that red-teams your AI — an autonomous agent running recon, test and judge loops with deterministic scoring.

## Blog

- [Black October: five incidents in one week](https://kagex.ai/blog/black-october): Black October, 17 to 23 October 2026: weak eMudhra keys, PHP2Shell, a hijacked LiteLLM release, a critical GitLab fix, and a Node.js security release.
- [Agent skills are code written in English](https://kagex.ai/blog/agent-skills-security): Agent skills are plain-English instructions your AI agent obeys. On one registry, 5 of the 7 most-downloaded were malware. How to review one safely.
- [Grok's zero-click chat leak: the injection guardrails cannot read](https://kagex.ai/blog/cryptographic-context-injection-grok): A web page can steal a Grok user's chat history with no click, by shipping instructions as ciphertext. Reported to xAI in June, still unpatched.
- [AI “mind viruses”: the attack that spreads between AI agents](https://kagex.ai/blog/ai-agent-memory-poisoning): Anthropic and EPFL built self-spreading “mind viruses” for AI agents. They survived 20 hops in the lab, and one paragraph of prompt text stops them.
- [AI agents escaping the sandbox: OpenAI, Claude and Kimi K3](https://kagex.ai/blog/ai-sandbox-escapes-openai-anthropic-kimi-k3): OpenAI, Anthropic and Moonshot disclosed AI sandbox escapes in three weeks. Same root cause, real impact, and the plain controls that would have stopped them.
- [Introducing Mirage: an AI red-team agent for your browser](https://kagex.ai/blog/mirage-red-team-agent): Mirage is a free, open-source browser extension that runs an autonomous AI red-team agent against any LLM app — recon, attack, and a deterministically confirmed, OWASP-mapped report.
- [What is AI red teaming? A practical guide for cybersecurity professionals](https://kagex.ai/blog/what-is-ai-red-teaming): Learn what AI red teaming is, what AI red teamers test, how it differs from penetration testing, and how to start building practical AI security skills in 2026.

## Company

- [About](https://kagex.ai/about): Who we are and why we take an offense-first approach.
- [Pricing](https://kagex.ai/pricing): Plans for individuals, teams and enterprise.
- [Careers](https://kagex.ai/careers): Open roles.
- [Contact](https://kagex.ai/contact): Enterprise and security inquiries.

## Notes

- Full text of every article in one file: https://kagex.ai/llms-full.txt
- Sitemap: https://kagex.ai/sitemap.xml
- RSS: https://kagex.ai/rss.xml
- All tooling is intended for systems you own or are explicitly authorized to test.