Is TiDB safe to install?

PingCAP · Data · stdio (local process)

Interact with TiDB databases using natural language over MCP.

Cmoderate exposureSource

C

Grade

56/100

Blast radius

7

Tools

2

Capabilities

What TiDB declares it can do

A server's blast radius is the most powerful thing each of its tools can do. This grades the tools it declares, read-only, not whether it is secure.

Execute0
Delete1
Write1
Read5

highTool declares a high-privilege capability×2
mediumTool input schema asks for a credential×2

Tools

Every tool TiDB exposes

7 tools

Control

Govern TiDB with AxioRank

7 tools

Installed as-is, all 7 of TiDB's tools are callable by your agent without restriction, including any that write, delete, or execute.

Flip the switch to see the policy AxioRank would apply.

Advisories

Informational content flags

Keyword matches in tool descriptions and schemas. These are shown for transparency and are not part of the grade.

highCode-execution capability×5
highCredential / secret access capability×2

Install

Add TiDB to your client

Drop this into your MCP client config (Claude Desktop, Cursor, and others).

mcp.json

{
  "mcpServers": {
    "tidb": {
      "command": "uvx",
      "args": [
        "--from",
        "pytidb[mcp]",
        "tidb-mcp-server"
      ]
    }
  }
}

Embed

Show the grade in your README

Links back to this page and updates when the grade changes.

[![AxioRank MCP Security Index grade for TiDB](https://axiorank.com/api/badge/mcp/tidb.svg)](https://axiorank.com/mcp-index/tidb)

More Data servers

ACouchbase APinecone Developer AHugging Face AQdrant AAmazon Redshift AMySQL (benborla)

How it works

About this grade

The grade reflects the blast radius of what TiDB declares it can do, read-only, not whether it is secure or trustworthy, and not a judgment of the vendor. Maintain this server? Claim this listing or request a re-scan. See the methodology.

Email me this scorecard

Get the TiDB grade in your inbox.

Let your agents use TiDB safely

Route this server through AxioRank to allowlist its tools, hold risky calls for approval, and keep a signed audit trail of every action.

Start free