MCP Security Skill

This skill enforces security best practices for MCP servers and multi-agent pipelines.

5-Layer Defense Architecture

1. Input Validation - Sanitize all user inputs
2. Prompt Injection Prevention - Detect and block injection attempts
3. SQL/NoSQL Validation - Prevent query injection
4. User Context Propagation - Maintain identity through pipeline
5. Authorization (RBAC/ABAC) - Enforce access controls

Prompt Injection Prevention