OPFOR

Open-source red-team platform that runs multi-turn adversarial attacks against AI agents, LLM applications and MCP servers, mapping tests to the OWASP LLM, Agentic AI and MCP Top 10 plus API Security and scoring responses with LLM judges. It ships as a CLI, browser extension, MCP server, SDK and IDE skills, logging every prompt, request, response and verdict for reproducibility.