Securing an MCP-Based LLM
Developed a secure MCP/FastMCP AI assistant with pre-call validation and post-call response filtering. Tested for prompt injection, secret leakage, suspicious API responses, XSS indicators and SQL injection indicators.
Workflow Implementation