LLMVault

Deliberately vulnerable, CTF-style training platform for the OWASP LLM Top 10.

Webfree

LLMVault is an open-source, deliberately vulnerable training platform for practicing the OWASP Top 10 for LLM Applications, in the spirit of WebGoat or KubeGoat but for AI systems. Play Mode offers 25 scored labs across three unlocking tiers covering jailbreaking, data poisoning, and agent exploitation, while Live Mode points the same techniques at a real local model with no scripted flag. It is aimed at security teams and researchers building hands-on LLM security skills.

Website

Full match profile

Behind the summary, Matchbox keeps a richer profile of LLMVault - the signals our matcher actually reads to decide when to surface it. It stays private; claim the listing to see and control it.

Problem & pain-point mapping
Who we surface it to (audience fit)
What it's a strong alternative to
Trust & credibility signals

Claim LLMVault to unlock →

Try Matchbox with your own problem

Describe what is not working - we’ll show you whether LLMVault (or something else) actually fits.

Try a problem