PM

MCP-сервер

promptspeak-mcp-server

Автор: chrbailey · Безопасность

Описание из исходного каталога

Pre-execution governance for AI agents. Intercepts and validates every agent tool call through an 8-stage pipeline before execution — risk classification, behavioral drift detection, hold queue for dangerous operations, and complete audit trail. 45 tools, 658 tests

Категория
Безопасность