🔐 Pheoeniks

🛡️ Unified AI Security and redteaming tool for Multi-Modal & Agentic Systems

🛡️ AI Security Platform for LLM & Agent ProtectionPhoeniks is an MCP-integrated security platform that detects and blocks prompt injection attacks across major LLM providers (OpenAI, Gemini, Mistral, Phi, Claude). Built on Gradio, it provides real-time threat analysis for both text and vision inputs through a simple, plug-and-play interface.The platform uses advanced LLMs as detection engines to identify jailbreaks, injection attempts, and adversarial prompts with high accuracy. It includes OWASP-aligned attack templates for red-teaming, customizable threat scenarios, and live behavioral monitoring. Fast heuristic pre-filtering ensures low-latency screening suitable for production environments. Deploy Phoeniks as middleware via MCP to protect your AI stack, gain visibility into attack patterns, and enforce security policies across agentic workflows.