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/59

Verdicts · 3,151 servers

✓SAFE1,634!CAUTION522✕BLOCK995

Transports

A server can list more than one transport.

3,151 servers · 1,393–1,416 · ranked by trust score, then stars
ibge-br-mcp✓SAFE B

SidneyBissoli/ibge-br-mcp

MCP Server for IBGE APIs - Brazilian geographic, demographic and statistical data

★ 12GitHub stars21 readstdio
xiangsu-mcp✓SAFE B

sleepyquq/xiangsu-mcp

像塑 MCP | AI Agent scene editing, scripting and playtesting for Douyin Xiangsu. Effect House compatibility unverified.

★ 12GitHub stars1 readstdio
campus-cli!CAUTION B

alejooroncoy/campus-cli

CLI no oficial para UPC Aula Virtual (Blackboard Learn) — terminal + MCP para Claude

★ 11GitHub stars4 writestdio
oauth-callback✓SAFE B

kriasoft/oauth-callback

OAuth 2.0 login for CLI tools and desktop apps: opens the browser and captures the authorization code on a loopback redirect (RFC 8252). Node.js, Deno, Bun. One-line MCP client auth.

★ 11GitHub stars1 readstreamable-http
uiflowchartcreator✓SAFE B

umshere/uiflowchartcreator

MCP server for creating UI flowcharts

★ 11GitHub stars1 readstdio
mcp-listen✓SAFE B

decibri/mcp-listen

Give your AI agents the ability to listen. Microphone capture and speech-to-text tools for MCP-compatible agents.

★ 10GitHub stars3 readstdio
OpenFinData✓SAFE B

robertoecf/OpenFinData

Dados Financeiros Abertos — infraestrutura open source para dados financeiros públicos do Brasil via API REST, MCP, CLI e Python.

★ 10GitHub stars10 read
bcb-br-mcp✓SAFE B

SidneyBissoli/bcb-br-mcp

MCP server for Brazil Central Bank (BCB) data — Selic, IPCA, exchange rates, GDP for AI agents

★ 9GitHub stars8 readstdio
genpark-hub✓SAFE B

alphaparkinc/genpark-hub

GenPark & OpenClaw Premium Growth Hub, Showcase and Live Skill Generator.

★ 9GitHub stars13 read
MeiGen-AI-Design-MCP✓SAFE B

HUANGcvs/MeiGen-AI-Design-MCP

🎨 Turn Claude Code and OpenClaw into a powerful design assistant with local ComfyUI, API integration, and a large pro prompt library.

★ 9GitHub stars3 writestdio
pumpfun-mcp✓SAFE B

nirholas/pumpfun-mcp

Free, read-only pump.fun + Solana MCP server — token discovery, on-chain bonding-curve & holder analysis, creator fee-claim tracking, SNS resolution, KOL signals, and swap quotes. No API keys.

★ 9GitHub stars2 writestdio
solana-sniper-mcp✓SAFE B

nirholas/solana-sniper-mcp

Lightweight, embeddable pump.fun sniper engine for 3D AI agents — multi-agent, multi-user, with pluggable wallet/store/RPC adapters. Library + CLI + MCP server + x402 paid API.

★ 9GitHub stars7 readstdio
Doppler MCP Server✓SAFE B

DopplerHQ/mcp-server

★ 8GitHub stars2 read
governed-workflow✓SAFE B

IgnatRozhkoTR/governed-workflow

Web admin panel and MCP server for Claude Code — phase gates, scope locks, and human approval for agentic coding sessions.

★ 8GitHub stars3 writestdio
avatar-agent-mcp✓SAFE B

nirholas/avatar-agent-mcp

3D AI Agent Avatar — MCP server that spawns a textured GLB avatar, inspects/validates/optimizes any 3D model, gives the agent a Solana wallet + a voice, and ships full pump.fun powers (atomic Jito-bundled launches + creator-fee collection). Powered by three.ws.

★ 7GitHub stars1 readstdio
jev-mcp✓SAFE B

rashedInt32/jev-mcp

MCP server exposing TypeSafe Jev as typed, calibrated judgment tools: classify, score, check, batched ask. Ships as a Claude Code plugin.

★ 7GitHub stars7 readstdio
mcp-server-pixeltable-developer!CAUTION B

pixeltable/mcp-server-pixeltable-developer

A local Model Context Protocol server for building, inspecting, and operating Pixeltable applications.

★ 7GitHub stars3 writestdio
mneme✓SAFE B

AgentGameLab/mneme

Persistent long-term memory for Claude Code agents — SQLite + FTS5 + sqlite-vec + RRF hybrid recall over MCP, with Chinese tokenization. Local-first, graceful degradation.

★ 7GitHub stars2 readstdio
steam-mcp✓SAFE B

Sarg338/steam-mcp

Read-only Steam MCP server for AI agents: 41 tools + prompts + resources across library, friends, achievements, store, reviews, sales, discovery, recommendations, and co-op planning. Official Steam Web API, bring-your-own-key, open source. On PyPI + the MCP Registry.

★ 7GitHub stars2 writestdio
threews-avatar-mcp✓SAFE B

nirholas/threews-avatar-mcp

MCP server for three.ws 3D avatars by three.ws — render a live, rotatable on-chain avatar inline (interactive MCP App), get a paste-anywhere embed iframe, or fetch avatar metadata. Free, no API key.

★ 7GitHub stars2 readstdio
token-naming-mcp✓SAFE B

nirholas/token-naming-mcp

Resolve Solana .sol names and check *.threews.sol agent-identity availability from any AI agent. sns_resolve maps a .sol name to its owner wallet, sns_reverse maps a wallet to its primary name, and threews_availability checks whether a three.ws agent handle is free to claim. No key, no signer, no pa

★ 7GitHub stars1 readstdio
x402-bridge!CAUTION B

nirholas/x402-bridge

MCP stdio bridge that turns any x402-paid HTTP endpoint into a Claude-callable tool. Auto-pays via the @x402 client (EVM exact, SVM exact, EVM batch-settlement) and pre-loads tools from the Coinbase x402 Bazaar.

★ 7GitHub stars3 readstdio
Florentine.ai MCP Server - Talk to your MongoDB & MySQL data✓SAFE B

florentine-ai/mcp

MCP server for Florentine.ai - Natural language to MongoDB aggregations

★ 6GitHub stars2 readstdio
gyazocli✓SAFE B

yuiseki/gyazocli

Fast, script-friendly CLI for Gyazo

★ 6GitHub stars9 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.