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/64

Verdicts · 3,151 servers

✓SAFE1,634!CAUTION522✕BLOCK995

Transports

A server can list more than one transport.

3,151 servers · 1,513–1,536 · ranked by trust score, then stars
Opencode-chrome✓SAFE B

G10hdz/Opencode-chrome

A Chrome extension that gives OpenCode browser control, similar to Claude or Codex in Chrome.

★ 1GitHub stars3 writestdio
options-analysis-suite✓SAFE B

options-analysis-suite/options-analysis-suite-mcp

AI integration for Options Analysis Suite - 46 data tools across Claude, ChatGPT, and Perplexity for personalized options trade analysis

★ 1GitHub stars1 writestdio
orion-celestia-intel!CAUTION B

orioninfra/orion-agent-kit

Celestia network intelligence for AI agents: MCP server over Orion's daily analysis. Read-only.

★ 1GitHub stars11 readstdio
orrery✓SAFE B

chrisfulcher/orrery

Free, open source, self-hosted business development intelligence for U.S. federal government contracting.

★ 1GitHub stars4 write
PaperCortex✓SAFE B

chetanmorey1/PaperCortex

Organize and search Paperless-ngx documents with local AI, auto-classification, receipt extraction, and accounting export.

★ 1GitHub stars5 readsse
postgresql-mcp-js✓SAFE B

s2mgtvodois2/postgresql-mcp-js

🔌 Simplify PostgreSQL interactions with a secure MCP server for LLMs, enabling efficient access to schemas, tables, and query execution.

★ 1GitHub stars1 writestdio
propline-mcp✓SAFE B

proplineapi/propline-mcp

Live odds, cross-book +EV and graded player-prop results across 37 books. Hosted endpoint included.

★ 1GitHub stars30 readstdio
riotskillissue✓SAFE B

Demoen/riotskillissue

The production-ready, auto-updating, and fully typed Python wrapper for the Riot Games API.

★ 1GitHub stars1 writestdio
scopewalker-mcp✓SAFE B

timohaa/scopewalker-mcp

Codebase analysis MCP server for AI agents: complexity metrics, size thresholds, prop drilling, and code smell detection built on tree-sitter and tokei.

★ 1GitHub stars9 readstdio
sheetrender-mcp✓SAFE B

sheetrender/sheetrender-mcp

MCP server for SheetRender: render PDFs from HTML templates and spreadsheet data via the SheetRender API

★ 1GitHub stars4 writestdio
songcheck✓SAFE B

afghanfansmedia-ai/songcheck-mcp

Is this song AI or human? Detect AI-generated music (Suno, Udio, and more) from any AI agent. MCP server by Khaled Media.

★ 1GitHub stars4 read
sovantica-engrava✓SAFE B

sovantica/engrava-mcp

Model Context Protocol server for Engrava — expose an AI agent memory database to any MCP client (Claude Desktop, Cursor, VS Code...) over stdio.

★ 1GitHub stars1 readstdio
substack-mcp-cli✓SAFE B

thenavidm/substack-mcp-cli

Substack MCP server and CLI for Claude Code, Codex and AI agents. 65 tools for drafts, posts, Notes, subscribers, analytics, tags, comments, and researching other writers.

★ 1GitHub stars4 read
tdries-tableau-graphql✓SAFE B

tdries/tableau-graphQL-mcp

Read-only MCP server for Tableau lineage and impact analysis over the Metadata API (GraphQL). Works on Tableau Server and Cloud.

★ 1GitHub stars1 writestdio
trace.moe-mcp✓SAFE B

soruly/trace.moe-mcp

Model Context Protocol (MCP) server for trace.moe anime scene search

★ 1GitHub stars5 readstdio
velric-agent!CAUTION B

ikooky/velric-agent

Agent 自主渗透测试工具: SPA/MPA 接口枚举、运行时流量捕获、权限树还原、参数矩阵合并。

★ 1GitHub stars2 writestdio
web✓SAFE B

GuionAI/web

A web research toolkit with multi-provider search and clean Markdown extraction from static and JavaScript-rendered pages, plus public code and library docs search—available via CLI and MCP.

★ 1GitHub stars14 readstdio
web-search-mcp✓SAFE B

awslew/web-search-mcp

编码 agent 查资料缺搜索 API key?免 key 搜多个引擎、优先可靠来源并抽取可读正文。 Search the web and read pages without an API key.

★ 1GitHub stars3 readstdio
yuanzhu✓SAFE B

RavenWangChina/yuanzhu

Meta-Forge Workshop: turn recurring work into governed AI workflow assets. MCP server + Claude Code plugin + dsh plugin — staged writes, evals-gated forging, inference engine. 元铸工坊 SOSOFAST:把最佳实践铸成可治理的 AI 工作流资产

★ 1GitHub stars1 writestdio
Mcp Ocr Fallback✓SAFE B

hoangdn3/mcp-ocr-fallback

mcp-ocr-fallback

★ 0GitHub stars1 writestdio
Fibel✓SAFE B

k2b-dev/fibel

Docs first - everything else later

★ 0GitHub stars3 readstreamable-http
@orboto/mcp✓SAFE B

orboto/orboto-mcp

★ 0GitHub stars2 readstdio
MCP TypeScript SDK !✓SAFE B

redocly-contrib/mcp-typescript-sdk

The official TypeScript SDK for Model Context Protocol servers and clients

★ 0GitHub stars2 readsse
adaptive-cards-mcp✓SAFE B

hinominomi/adaptive-cards-mcp

Build Adaptive Cards with MCP tools for fast, structured UI creation in TypeScript

★ 0GitHub stars7 writesse

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.