mcp-neural-shield
Last released
A neural network-powered security middleware and reverse proxy for the Model Context Protocol (MCP). Detects tool poisoning, shadowing, and prompt injection attacks using deep learning classifiers trained on SentenceTransformer embeddings.
