Evaluation record · mcp-server-postgres

MCP PostgreSQL Server

v0.6.2

Anthropic (Archived)

MCPdatabasesqlmcpmodel-context-protocol

Adequate

About This MCP

ARCHIVED WITH UNPATCHED VULNERABILITY: Former Anthropic reference MCP server for PostgreSQL, archived 2025-05-29. A SQL injection flaw disclosed by Trend Micro (June 2025) remains unpatched because the repo is archived; the npm package (v0.6.2) is still published and downloaded. NOT RECOMMENDED for any use; a patched community fork (@zeddotdev/postgres-context-server, fixed in v0.1.4) and maintained alternatives (e.g., Microsoft's MCP server for Azure Database for PostgreSQL) exist.

Last Evaluated: July 9, 2026

Official Website

Trust Vector Analysis

Dimension Breakdown

🚀Performance & Reliability

query execution accuracy

Query execution testing

Evidence

PostgreSQL Driver Integration — Uses standard PostgreSQL drivers with high query accuracy

highVerified: 2026-07-09

schema introspection

Schema discovery testing

Evidence

MCP Postgres Implementation — Comprehensive schema inspection including tables, columns, indexes, and constraints

highVerified: 2026-07-09

connection stability

Connection reliability testing

Evidence

PostgreSQL Reliability — Built on PostgreSQL's robust connection pooling and stability

highVerified: 2026-07-09

large result handling

Large dataset performance testing

Evidence

MCP Server Implementation — Handles large result sets with pagination and limits

mediumVerified: 2026-07-09

error handling

Error scenario testing

Evidence

PostgreSQL Error Reporting — Detailed error messages from PostgreSQL with proper propagation

highVerified: 2026-07-09

🛡️Security

sql injection protection

Vulnerability disclosure review and exploitability analysis

Evidence

Trend Micro Research — Classic SQL injection vulnerability disclosed in the archived PostgreSQL MCP server; will never be patched in @modelcontextprotocol/server-postgres because the repository is archived

Datadog Security Labs — Case study confirming SQL injection in the PostgreSQL MCP server allows bypassing the read-only transaction restriction; ~21k weekly npm downloads while vulnerable

npm: @modelcontextprotocol/server-postgres — v0.6.2 remains the latest published version as of July 2026; SQL injection still unpatched, no CVE identifier assigned. Patched fork @zeddotdev/postgres-context-server fixed the flaw in v0.1.4

highVerified: 2026-07-09

access control

Permission boundary testing

Evidence

PostgreSQL Permissions — Inherits database user permissions but AI can execute any query within those permissions

Datadog Security Labs — SQL injection allows escaping the server's intended read-only transaction sandbox, defeating its primary access-control mechanism

highVerified: 2026-07-09

data modification risk

Write operation risk assessment

Evidence

Security Analysis — AI can execute INSERT, UPDATE, DELETE operations if database user has permissions

Trend Micro Research — Unpatched SQL injection enables write operations even when the server is configured for read-only access

highVerified: 2026-07-09

credential security

Credential storage review

Evidence

Connection String Management — Credentials stored in configuration; requires secure credential management

mediumVerified: 2026-07-09

audit logging

Logging capabilities assessment

Evidence

PostgreSQL Logging — Depends on PostgreSQL audit configuration; MCP server provides basic query logging

mediumVerified: 2026-07-09

🔒Privacy & Compliance

data exposure risk

Data flow and exposure analysis

Evidence

Data Flow Analysis — Query results including sensitive data sent to LLM provider

highVerified: 2026-07-09

pii protection

Privacy controls assessment

Evidence

MCP Security Guidelines — No built-in PII detection, redaction, or anonymization

highVerified: 2026-07-09

query result filtering

Data filtering capabilities review

Evidence

MCP Server Configuration — Can limit result rows but no column-level filtering for sensitive data

mediumVerified: 2026-07-09

compliance readiness

Compliance framework review

Evidence

Compliance Assessment — GDPR/HIPAA compliance depends on LLM provider and data handling configuration

mediumVerified: 2026-07-09

schema exposure

Metadata exposure analysis

Evidence

Schema Introspection — Database schema metadata shared with LLM for query generation

highVerified: 2026-07-09

👁️Trust & Transparency

documentation quality

Documentation completeness review

Evidence

MCP Postgres Docs — Clear documentation with setup instructions and security considerations

highVerified: 2026-07-09

query visibility

Query traceability assessment

Evidence

MCP Protocol Logging — All SQL queries logged and visible in MCP message stream

highVerified: 2026-07-09

open source code

Source code review

Evidence

GitHub Repository — Fully open source with MIT license

highVerified: 2026-07-09

security documentation

Security documentation review

Evidence

Security Guidelines — Provides security best practices but could be more comprehensive for database access

MCP servers-archived repository README — README states 'NO SECURITY GUARANTEES ARE PROVIDED FOR THESE ARCHIVED SERVERS'; known SQL injection remains undocumented and unpatched in the package

highVerified: 2026-07-09

⚙️Operational Excellence

ease of setup

Setup complexity assessment

Evidence

MCP Quickstart — Straightforward setup requiring connection string configuration

highVerified: 2026-07-09

performance

Performance benchmarking

Evidence

PostgreSQL Performance — Query performance depends on database configuration and optimization

highVerified: 2026-07-09

connection pooling

Connection management review

Evidence

Implementation Details — Supports connection pooling for efficient resource utilization

mediumVerified: 2026-07-09

error diagnostics

Error messaging assessment

Evidence

PostgreSQL Error Messages — Detailed error messages help diagnose issues

mediumVerified: 2026-07-09

community support

Community activity analysis

Evidence

MCP servers-archived repository — Server archived 2025-05-29; repository read-only, no issues or PRs accepted. Patched community fork available as @zeddotdev/postgres-context-server

highVerified: 2026-07-09

Strengths

+Comprehensive PostgreSQL feature support including schema introspection
+Natural language to SQL query generation capabilities
+Excellent for data analysis and business intelligence workflows
+Full query visibility and logging for audit purposes
+Open source code remains publicly auditable in the archived repository
+Supports connection pooling and performance optimization

Limitations

!High risk of exposing sensitive database content to LLM providers
!No built-in PII detection, redaction, or data anonymization
!AI can execute destructive operations (DELETE, DROP) if permissions allow
!Limited granular access control beyond database user permissions
!Query results with sensitive data sent to external APIs
!Compliance challenges for regulated industries (HIPAA, PCI-DSS, GDPR)
!ARCHIVED 2025-05-29 with an UNPATCHED SQL injection vulnerability (Trend Micro, June 2025) confirmed still unpatched in npm v0.6.2 as of July 2026; use the patched fork @zeddotdev/postgres-context-server (v0.1.4+) or a maintained alternative instead

Metadata

license: MIT

supported platforms

0: All platforms with PostgreSQL client support

programming languages

0: TypeScript

1: Python

mcp version: 1.0

github repo: https://github.com/modelcontextprotocol/servers-archived

github stars: 58700

database version: PostgreSQL 10+

connection method: Connection string with credentials

first release: 2024-11

maintained by: None (Archived 2025-05-29)

status: Archived - unpatched SQL injection vulnerability (still unpatched as of 2026-07); patched community fork: @zeddotdev/postgres-context-server (v0.1.4+); maintained alternative: Microsoft MCP server for Azure Database for PostgreSQL

package name: @modelcontextprotocol/server-postgres

transport types

0: stdio

installation methods

0: npm

Use Case Ratings

code generation

Useful for generating database migration scripts and query code

customer support

Can help support teams query customer data, but privacy concerns exist

content creation

Limited applicability; mainly for content stored in databases

data analysis

Excellent for AI-powered data analysis, reporting, and insights generation

research assistant

Good for analyzing research datasets stored in PostgreSQL

legal compliance

High risk due to potential exposure of sensitive legal data to LLM providers

healthcare

Not recommended for PHI due to data exposure risks; HIPAA compliance challenges

financial analysis

Risky for sensitive financial data; requires strict access controls and data filtering

education

Good for analyzing student data and learning analytics

creative writing

Limited utility unless creative content is database-driven