Last released Jun 12, 2026
A permission layer for AI agent tool calls — wrap any MCP server and enforce allow/deny/ask policy with full audit.
Supported by
