Red-team tool for jailbreaking LLMs — swap custom prompts per model to test safety boundaries
Curated collection of system prompts and agent specifications for 25+ frontier AI models
Lightweight OS-level sandboxing for AI agents, MCP servers, and arbitrary processes
Advanced jailbreak prompts and testing toolkit for GPT Codex models with safety-focused evaluation framework
Independent auditing of AI agents - catch your agent's mistakes and blind spots in under 120 seconds
A fast programming language designed for the post-AGI economy that blocks AI mistakes via mathematical proof
The AI red-teaming toolkit for LLM security assessment and vulnerability scanning
Enterprise security system for AI agents - observability, benchmarking, and threat detection for Cursor, Claude Code, and more
Sandbox any AI agent in seconds - zero setup, zero latency sandbox for Claude Code, Copilot, and more