{"id":"postgresql-server-mcp","name":"postgresql-server-mcp","af_score":39.8,"security_score":48.5,"reliability_score":25.0,"what_it_does":"Provides an MCP (Model Context Protocol) server interface for working with PostgreSQL—allowing an AI agent to connect to a Postgres instance and perform database-related operations via MCP tools.","best_when":"You want a standardized way for an LLM agent to interact with PostgreSQL from within an MCP-capable agent/runtime, ideally with restricted credentials and read-only or tightly controlled actions.","avoid_when":"You cannot enforce least-privilege access, query restrictions (e.g., read-only), or proper handling of credentials/network security; or when you need REST/GraphQL/SDK interfaces instead of MCP.","last_evaluated":"2026-04-04T21:48:03.617671+00:00","has_mcp":true,"has_api":false,"auth_methods":["Postgres connection authentication (e.g., user/password, possibly via connection string)"],"has_free_tier":false,"known_gotchas":["Agents may generate unsafe SQL unless the MCP tool enforces query allowlists/read-only constraints.","DB credentials must be scoped tightly; otherwise tool access becomes equivalent to direct DB access.","Network/VPC connectivity and Postgres authentication/pg_hba.conf misconfiguration are common failure modes; absent guidance can slow integration."],"error_quality":0.0}