Observed arrival · 2026-08-29
Gatekeeper puts a human between AI agents and production SQL
An MCP server and Beekeeper Studio plugin that lets AI agents propose database queries while a human reviews and approves each statement.
Field notes
Gatekeeper separates proposal from execution: an agent submits SQL over MCP and receives a ticket, while Beekeeper Studio presents the statement for human approval. The displayed examples cover a customer SELECT flagged for possible PII, an account UPDATE requiring write mode, and a DELETE requiring destructive mode. Results return through get_query_result, and the page says the agent never receives the credentials held by Beekeeper.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
One card from the complete issue