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/73

Verdicts · 3,151 servers

✓SAFE1,634!CAUTION522✕BLOCK995

Transports

A server can list more than one transport.

3,151 servers · 1,729–1,752 · ranked by trust score, then stars
Prefect!CAUTION B

prefecthq/prefect-mcp-server

Prefect MCP server

★ 54GitHub stars1 readstdio
MCPcat!CAUTION B

mcpcat/mcpcat-python-sdk

AgentCat is an analytics platform for MCP server owners 🐱.

★ 53GitHub stars1 writestreamable-http
Modular✓SAFE B

d-kimuson/modular-mcp

A Model Context Protocol (MCP) proxy server that enables efficient management of large tool collections across multiple MCP servers by grouping them and loading tool schemas on-demand.

★ 51GitHub stars1 readstdio
Chrome Control✓SAFE B

lxe/chrome-mcp

Allows LLM agents to control a local chrome instance without taking screenshots

★ 49GitHub stars1 writesse
Flight Ticket!CAUTION B

xiaonieli7/flightticketmcp

Flight Ticket MCP Server 实现了供航空机票相关查询操作的工具和资源。它作为AI助手与航空服务系统之间的桥梁,专注于航班实时动态查询功能。

★ 48GitHub stars10 readsse
OSINT Tools✓SAFE B

himanshusanecha/mcp-osint-server

★ 48GitHub stars7 readstdio
Homelab Manager✓SAFE B

bjeans/homelab-mcp

Model Context Protocol (MCP) servers for managing homelab infrastructure through Claude Desktop. Monitor Docker/Podman containers, Ollama AI models, Pi-hole DNS, Unifi networks, and Ansible inventory. Includes security checks, templates, and automated pre-push validation. Production-ready for homela

★ 43GitHub stars1 writesse
PJE!CAUTION B

chapirousia/pje-mcp-server

★ 42GitHub stars10 readstdio
RootCause✓SAFE B

yindia/rootcause

RootCause is a local-first MCP server that turns natural-language requests into evidence-backed incident analysis, Kubernetes diagnostics, and safer operations.

★ 42GitHub stars2 write
AI Mate!CAUTION B

symfony/ai-mate

AI development assistant CLI for Symfony projects

★ 41GitHub stars1 read
faf-cli✓SAFE B

Wolfe-Jam/faf-cli

The context every AI coding agent reads — authors AGENTS.md, CLAUDE.md, GEMINI.md & .cursorrules from your repo's real stack. IANA-registered .faf format.

★ 40GitHub stars1 writestreamable-http
Sunnyside Figma Context!CAUTION B

tercumantanumut/sunnysidefigma-context-mcp

A comprehensive Model Context Protocol (MCP) server that bridges Figma designs with AI development workflows. It provides 30 specialized tools for extracting pixel-perfect code, assets, and component structures directly from Figma designs.

★ 40GitHub stars1 writesse
Agent Todos✓SAFE B

digitarald/vscode-agent-todos

Gives VS Code agent mode planning superpowers with dynamic todo lists

★ 39GitHub stars1 writestreamable-http
Better Notion✓SAFE B

n24q02m/better-notion-mcp

Markdown-first Notion for AI agents -- pages, databases, blocks, and comments in one call.

★ 36GitHub stars1 writestdio
CFM Tips✓SAFE B

aws-samples/sample-cfm-tips-mcp

A MCP Server that's built on top of AWS Cloud Financial Management (CFM) Technical Implementation Playbooks (TIPs) - our proven collection of cost optimization best practices

★ 34GitHub stars14 writestdio
Advocu!CAUTION B

carlosazaustre/advocu-mcp-server

★ 31GitHub stars11 writestdio
rollbar-mcp-server!CAUTION B

rollbar/rollbar-mcp-server

Model Context Protocol server for Rollbar

★ 28GitHub stars2 writestdio
Interoperable Agent System!CAUTION B

junfanz1/mcp-multiserver-interoperable-agent2agent-langgraph-ai-system

This project demonstrates a decoupled real-time agent architecture that connects LangGraph agents to remote tools served by custom MCP (Modular Command Protocol) servers. The architecture enables a flexible and scalable multi-agent system where each tool can be hosted independently (via SSE or STDIO

★ 28GitHub stars1 writesse
Whoop✓SAFE B

ctvidic/whoop-mcp-server

★ 26GitHub stars16 readstdio
Github Mapper✓SAFE B

dazeb/mcp-github-mapper

MCP GitHub Mapper is a MCP tool that will map any repository remotely and import the map directly into your code editor.

★ 25GitHub stars1 writestdio
Metatron✓SAFE B

kerbelp/metatron

Git-native context layer for AI coding agents. Your team's real engineering decisions — patterns, pitfalls, conventions — live as reviewed markdown files in your repo; agents consult them before writing code and record what they learn. Files-first, no server required; MCP as an optional serving laye

★ 25GitHub stars1 read
Tool Builder!CAUTION B

hanweg/mcp-tool-builder

MCP server that creates its own tools as needed

★ 25GitHub stars1 writestdio
V0 Platform✓SAFE B

longnguyen0612/v0-platform-mcp

★ 25GitHub stars1 writestdio
JigsawStack✓SAFE B

jigsawstack/jigsawstack-mcp-server

Model Context Protocol Server that allows AI models to interact with JigsawStack models!

★ 23GitHub stars13 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 (1,622 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.