Databases

Postgres

SQL inspection and query workflows for Postgres-backed products.

Install TestedMCP Rank ReviewedMedium confidencemedium riskclaudecursorcodex

Unranked score draft

Install + docs80
Maintenance75
Auth handling67
Client compatibility84
Real usefulness88
Safety signals66

Evidence history

Review snapshots show why the current score, confidence, and risk labels changed over time.

This review depth is not eligible for MCP Rank leaderboards until a Deep Review is complete.

seed review
DateChangeScoreConfidenceRisk
May 13, 2026
Initial MCP Rank review captured from the curated seed dataset.

The install model is easy to reproduce for teams that already manage Postgres connection strings.

seed snapshot · reviewed

78mediummedium

Evidence notes

The install model is easy to reproduce for teams that already manage Postgres connection strings.
The strongest use case is read-only schema and query analysis, where the agent can explain tables without receiving application admin privileges.
The archived reference path caps confidence at medium even though package and source evidence are directly linkable.
Operational safety is determined by the database role: a readonly reporting replica is a different risk class from a production owner connection.
Review quality depends on inspecting whether the deployed config exposes mutation tools, transaction controls, or unrestricted SQL execution.

Best-fit use cases

Best for engineering teams that need schema exploration, migration planning, or query explanation against a safe replica.

Useful for support analytics when the database role is constrained to approved views.

Not suitable as a default MCP server for production databases without role isolation and query logging.

Risk analysis

The main risk is credential scope, not the MCP protocol surface by itself.

SQL generation can leak or transform sensitive rows even when the agent only appears to be answering a question.

A safe deployment should combine network allowlists, readonly roles, row-level controls where available, and human review for expensive queries.

Use-case examples

Explain table relationships.
Draft a read-only analytics query.
Find likely PII columns.