Shield-MCP

A safety wrapper for AI agent workflows that filters risky tool output, screens for prompt injection patterns, and adds a quieter layer of runtime control.

This product page is intentionally minimal for now. Full launch details are on the way.