Shrike Security

Shrike Security is an MCP (Model Context Protocol) server designed specifically to provide runtime security tools for AI agents and assistants. It exposes 12 real-time scanning tools, enabling detection of prompt injection, jailbreaks, SQL injection, PII exposure, agent-to-agent vulnerabilities, and a wide range of data leakage and manipulation threats. Shrike integrates seamlessly with AI clients and tools that support MCP, such as Claude, Cursor, and Windsurf, as well as via REST API, SDKs (TypeScript, Python, Go), and browser extensions. Its multi-layer detection pipeline combines fast pattern matching with optional LLM-powered semantic analysis, and supports compliance with major frameworks like GDPR, HIPAA, and PCI-DSS, making it a fit for AI developers, security teams, and enterprises building or orchestrating LLM-powered agents.