RESEARCH LAB

ERL Lab

Open-source tools and training modules built by our research team. We build what the security community needs.

LIVEv1.2.0

Schrute App

Advanced Prompt Injection Training Module. Test your social engineering skills against a hardened AI persona. Built to help security professionals understand and defend against prompt injection attacks.

Next.jsGoogle GeminiTypeScript
Launch App →

Features

  • ✓ Multi-level difficulty progression
  • ✓ Real-time AI persona responses
  • ✓ Prompt injection technique library
  • ✓ Leaderboard and completion tracking
COMING SOON

Agentic Pwn

Real-time adversarial agent battles. Pit AI agents against each other in a sandboxed environment to test agent security, tool-use vulnerabilities, and goal hijacking defenses.

PythonLangGraphWebSockets

Features

  • ✓ Multi-agent adversarial arena
  • ✓ Tool-use abuse detection
  • ✓ Goal hijacking scenarios
  • ✓ Agent security scoring
COMING SOON

Deep Fuzzer

Automated vulnerability scanner for LLM endpoints. Detect data leaks, hallucination exploitation, injection vectors, and guardrail bypasses across your AI applications.

PythonFastAPIDocker

Features

  • ✓ Automated prompt fuzzing
  • ✓ Output analysis and leak detection
  • ✓ Guardrail bypass testing
  • ✓ CI/CD integration