MCP servers · least privilege

MCP servers with no destructive tools

3,145 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. 1,620 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 29 Oct 2026. https://modelsatlas.com/mcp/no-destructive/page/27

Verdicts · 3,145 servers

✓SAFE1,632!CAUTION521✕BLOCK992

Transports

A server can list more than one transport.

3,145 servers

625–648 of 3,145 · ranked by trust score, then stars

Shinzo Ts

shinzo-labs/shinzo-ts

✓SAFE B

TypeScript SDK for MCP server observability, built on OpenTelemetry. Gain insight into agent usage patterns, contextualize tool calls, and analyze server performance across platforms. Integrate with any OpenTelemetry ingest service including the Shinzo platform.

★ 682 writestdio

agentmail-mcp

agentmail-to/agentmail-mcp

✓SAFE B

Email inboxes for AI agents: send, receive, reply, search, and manage threaded email over MCP.

★ 684 readstdio

Deepseek R1

66julienmartin/mcp-server-deepseek_r1

✓SAFE B

A Model Context Protocol (MCP) server implementation connecting Claude Desktop with DeepSeek's language models (R1/V3)

★ 681 readstdio

Lexware Office

jannikwempe/mcp-lexware-office

✓SAFE B

MCP server to interact with Lexware Office

★ 681 writestdio

Xcode Index

block/xcode-index-mcp

✓SAFE B

Give AI agents access to Xcode's index

★ 684 read

Agenite

subeshb1/agenite

✓SAFE B

🤖 Build powerful AI agents with TypeScript. Agenite makes it easy to create, compose, and control AI agents with first-class support for tools, streaming, and multi-agent architectures. Switch seamlessly between providers like OpenAI, Anthropic, AWS Bedrock, and Ollama.

★ 672 write

Claude Code Emacs

yuya373/claude-code-emacs

✓SAFE B

This package provides seamless integration with Claude Code, allowing you to run AI-powered coding sessions directly in your Emacs environment.

★ 6714 readstdio

Guide

qpd-v/mcp-guide

✓SAFE B

A beginner-friendly guide server that helps users understand MCP concepts, provides interactive examples, and demonstrates best practices for building MCP integrations. Features tools for exploring MCP capabilities, resources for learning core concepts, and prompts for guided tutorials.

★ 674 writestdio

MC Modding

ogmatrix/mcmodding-mcp

✓SAFE B

mcmodding-mcp is a Model Context Protocol (MCP) server that gives AI assistants like Claude direct access to Minecraft modding documentation. Instead of relying on potentially outdated training data, your AI assistant can search real documentation, find code examples, and explain concepts accurately

★ 671 writestdio

Public.com

publicdotcom/publicdotcom-mcp-server

✓SAFE B

Connect AI assistants to your Public.com brokerage account. Trade stocks, options, and crypto — get quotes, manage orders, and view your portfolio.

★ 6713 writestdio

Real Estate

agentic-ops/real-estate-mcp

✓SAFE B

A Model Context Protocol server for real estate workflows — search, valuation, mortgage math, and investment analysis.

★ 672 writestreamable-http

Unity API

codeturion/unity-api-mcp

✓SAFE B

Instant, accurate Unity API lookups instead of expensive source file reads, saving your agent tokens, context, and hallucinations

★ 675 readstdio

AgentMail

agentmail-to/agentmail-mcp

✓SAFE B

Email inboxes for AI agents: send, receive, reply, search, and manage threaded email over MCP.

★ 663 readstdio

Alby Bitcoin Payments

getalby/mcp

✓SAFE B

Connect a bitcoin lightning wallet to your LLM using Nostr Wallet Connect and Model Context Protocol

★ 661 writesse

Astro Docs

withastro/docs-mcp

✓SAFE B

Astro Docs MCP Server

★ 661 readstreamable-http

IOTA Agent

scottcjn/iota-agent-mcp

✓SAFE B

MCP server for IOTA blockchain — 20 tools for AI agent integration (wallet, Move CLI, on-chain queries via Model Context Protocol)

★ 666 writestdio

Skill Vision Control

jane-xiaoer/skill-vision-control

✓SAFE B

Safe MCP Skill version manager - detect updates, A/B testing, smart merge

★ 663 writestdio

VibeShift

groundng/vibeshift

✓SAFE B

[MCP Server] The Security Agent for AI assisted coding

★ 661 write

VRChat

sawa-zen/vrchat-mcp

✓SAFE B

This project is a Model Context Protocol (MCP) server for interacting with the VRChat API.

★ 661 writestdio

AI Federation Network

ruvnet/federated-mcp

✓SAFE B

This implementation follows the official MCP specification, including proper message framing, transport layer implementation, and complete protocol lifecycle management. It provides a foundation for building federated MCP systems that can scale across multiple servers while maintaining security and

★ 656 readstdio

Join Cloud

kushneryk/join.cloud

✓SAFE B

Join.cloud lets AI agents work together in real-time rooms. Agents join a room, exchange messages, commit files to shared storage, and optionally review each other's work — all through standard protocols (MCP and A2A).

★ 653 writestdio

Legifrance

pylegifrance/mcp-server-legifrance

✓SAFE B

Legifrance MCP Server

★ 653 readstdio

MyChatArchive

1ch1n/mychatarchive

✓SAFE B

Local-first AI memory archive. Import ChatGPT, Claude, and Grok exports, generate semantic embeddings, and search via MCP server. Zero cloud, zero cost.

★ 656 readsse

ReactBits

ceorkm/reactbits-mcp-server

✓SAFE B

MCP server providing access to 135+ animated React components from ReactBits.dev (9.2/10 test score)

★ 655 readstdio

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 (1,620 servers). This list also includes servers whose tools write: 1,525 of the 3,145 do.
Does no destructive tools mean safe?
No. 1,632 are SAFE, 521 CAUTION and 992 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.