MCP servers · least privilege

MCP servers with no destructive tools

3,151 MCP servers declare at least one tool and none of them is classed destructive: nothing deletes, overwrites or runs a command through a declared tool. 1,622 of them only read; the rest may create or edit.

Ask an AI about this pageGChatGPTCClaudePPerplexityGeGeminiXGrok
More
Cite

Cite this page

ModelsAtlas. “MCP servers with no destructive tools”. Retrieved 30 Oct 2026. https://modelsatlas.com/mcp/no-destructive/page/58

Verdicts · 3,151 servers

✓SAFE1,634!CAUTION522✕BLOCK995

Transports

A server can list more than one transport.

3,151 servers · 1,369–1,392 · ranked by trust score, then stars
Community✓SAFE B

mirascope/mcp-community

Easily run, deploy, and connect to MCP servers

★ 22GitHub stars1 write
Cypress Page Object & Test Generator✓SAFE B

jprealini/cypress-mcp

★ 22GitHub stars1 readstdio
Nearby Search✓SAFE B

kukapay/nearby-search-mcp

An MCP server for nearby place searches with IP-based location detection.

★ 22GitHub stars1 read
Server Proxy✓SAFE B

leizongmin/mcp-server-proxy

Converts MCP protocol's SSE transport layer to a standard HTTP request/response.

★ 22GitHub stars1 read
Trace Search✓SAFE B

songpengfei1111-dot/tracesearch_mcp

★ 22GitHub stars4 readstdio
Code Merge✓SAFE B

yy1588133/code-merge-mcp

★ 21GitHub stars1 writestdio
balatro-agent✓SAFE B

Arcadi4/balatro-agent

Let AI play Balatro

★ 20GitHub stars2 writestdio
cred-1✓SAFE B

aloth/cred-1

CRED-1: Open, reproducible domain credibility dataset for automated pre-bunking of online misinformation. TypeScript library, CLI & MCP server. Check any news source's reliability from terminal, code, or AI agents. CC BY 4.0, DOI-citable.

★ 17GitHub stars5 readstdio
@putervision/agent-reasoning-mcp✓SAFE B

putervision/agent-reasoning-mcp

🧭 Strategic BDI cognitive reasoning engine for AI agents. Goal decomposition DAGs, multi-factor utility scoring, risk evaluation, and adaptive replanning over MCP.

★ 15GitHub stars9 readstdio
Sdk✓SAFE B

settlemint/sdk

Integrate SettleMint into your application with ease.

★ 15GitHub stars19 writestdio
Taiga Ui Mcp✓SAFE B

taiga-family/taiga-ui-mcp

Taiga UI MCP server providing documentation search and scaffolding tools.

★ 15GitHub stars4 readstdio
call-a-human-mcp✓SAFE B

nishantmodak/call-a-human-mcp

MCP server that gives any AI phone a friend options : ask a human or request approval via Slack or macOS dialogs before taking irreversible actions.

★ 15GitHub stars2 readsse
mcp-server-polarion✓SAFE B

devemberx/mcp-server-polarion

MCP server for Polarion ALM — read and write documents, work items, test runs, traceability links, and comments from any MCP-compatible AI assistant

★ 15GitHub stars2 readstdio
mcpql✓SAFE B

hendrickcastro/MCPQL

MCPQL - MCP Server for SQL Server database operations and analysis

★ 15GitHub stars2 writestdio
medical-terminologies-mcp✓SAFE B

SidneyBissoli/medical-terminologies-mcp

Unified MCP server for six medical terminologies (ICD-11, LOINC, RxNorm, MeSH, ATC, CID-10) with authoritative WHO ICD-10→ICD-11 mapping and provenance on every response. Hosted endpoint: no install, no credentials.

★ 15GitHub stars33 readstdio
pubcrawl✓SAFE B

nickjlamb/pubcrawl

MCP server giving AI assistants verifiable access to PubMed, Europe PMC, US/UK drug labelling, and ClinicalTrials.gov. Built by PharmaTools.AI.

★ 15GitHub stars15 readstdio
1point3acres-toolkit✓SAFE B

vivian-labs/1point3acres-toolkit

一亩三分地自动签到 + 每日答题:在你自己电脑的 Chrome 里、用你自己的登录完成,不经过第三方;也是给 Claude Code / Codex 用的 MCP 服务。1point3acres daily check-in & quiz automation, local-only, plus an MCP server for AI assistants.

★ 14GitHub stars1 writestdio
storybook-mcp-server✓SAFE B

stefanoamorelli/storybook-mcp-server

MCP server for Storybook - provides AI assistants access to components, stories, properties and screenshots. Built with TypeScript and Model Context Protocol SDK.

★ 14GitHub stars6 readstdio
whoop-mcp-server!CAUTION B

yuridivonis/whoop-mcp-server

Your WHOOP recovery, sleep, strain and workouts in Claude, ChatGPT or other MCP apps. Self-hosted.

★ 14GitHub stars5 readstdio
htmldrop✓SAFE B

vin-spiegel/htmldrop

Publish HTML for agents — one API call, MCP server included

★ 13GitHub stars1 writestdio
mssql-mcp✓SAFE B

BYMCS/mssql-mcp

MCP server for Microsoft SQL Server — query, schema, and stored-procedure access for Claude Desktop, Cursor, Windsurf, and VS Code.

★ 13GitHub stars4 writestdio
openhammer-api✓SAFE B

EshanPrakash/openhammer-api

Comprehensive Warhammer 40K API with 2,748 units across 35 factions and 2 game editions (10th and 11th). Built with FastAPI, featuring 29 endpoints. Also available as a live MCP server.

★ 13GitHub stars9 readstreamable-http
Nest Openapi✓SAFE B

ts-oas/nest-openapi

A modern, modular set of utilities for building OpenAPI-driven NestJS apps.

★ 12GitHub stars1 writestdio
@putervision/behavior-mcp✓SAFE B

putervision/behavior-mcp

⚡ High-frequency (~60Hz) in-browser behavior tree execution engine for autonomous AI agents. Reactive priority interrupts, safety guardrails & deterministic action replay over MCP.

★ 12GitHub stars1 writestdio

Other risk hubs

By job

Questions

What counts as destructive?
A tool that deletes, overwrites or runs commands. We class every tool the source declares as read, write or destructive; a server is listed here when it has tools and none is destructive.
How is this different from read-only?
Read-only servers have only read tools (1,622 servers). This list also includes servers whose tools write: 1,529 of the 3,151 do.
Does no destructive tools mean safe?
No. 1,634 are SAFE, 522 CAUTION and 995 BLOCK on this list. The verdict comes from the audit of the source, not from the tool list alone.

Keep going in your AI

Opens with this page as the source. We add nothing else to the prompt.