Runtime security gateway and FastMCP server that protects MCP clients from tool poisoning, prompt injection, and unauthorized tool schema changes through policy enforcement, fail-closed scanning, and human approval gates.
Enforces deterministic security policies on Model Context Protocol traffic between agents and remote MCP servers, including request validation, signed human approval, response-side credential blocking, prompt-injection flagging, and privacy-minimized auditing.
Provides a governance gateway for LLM and agent traffic via MCP, blocking prompt injection, redacting PII, enforcing per-tenant token budgets, and caching repeat questions semantically. Enables AI agents to route model calls through the same guardrails instead of calling providers directly.