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/119

Verdicts · 3,145 servers

✓SAFE1,632!CAUTION521✕BLOCK992

Transports

A server can list more than one transport.

3,145 servers

2,833–2,856 of 3,145 · ranked by trust score, then stars

Anymd

SylphxAI/anymd

✕BLOCK F

Any file → clean Markdown for AI agents: PDF, Word, PowerPoint, Excel, EPUB, HTML and web pages, images (OCR), audio and video (metadata, subtitles, transcripts). A fast Rust MCP server and CLI that runs on your machine. No API key.

★ 1,0351 readstdio

PDF Reader

sylphxltd/pdf-reader-mcp

✕BLOCK F

Any file → clean Markdown for AI agents: PDF, Word, PowerPoint, Excel, EPUB, HTML and web pages, images (OCR), audio and video (metadata, subtitles, transcripts). A fast Rust MCP server and CLI that runs on your machine. No API key.

★ 1,0201 readstdio

PDF Reader

sylphxai/pdf-reader-mcp

✕BLOCK F

Any file → clean Markdown for AI agents: PDF, Word, PowerPoint, Excel, EPUB, HTML and web pages, images (OCR), audio and video (metadata, subtitles, transcripts). A fast Rust MCP server and CLI that runs on your machine. No API key.

★ 9361 readstdio

Dingo

dataeval/dingo

✕BLOCK F

Dingo: A Comprehensive AI Data, Model and Application Quality Evaluation Tool

★ 7572 writesse

McpStore

ip2a/mcpstore

✕BLOCK F

Elegantly manage your MCP services

★ 4321 readstreamable-http

pgEdge Postgres

pgedge/pgedge-postgres-mcp

✕BLOCK F

pgEdge MCP Server. A PostgreSQL MCP server with a Natural Language Agent CLI and Web UI.

★ 2331 read

Swagger Gateway

zaizaizhao/mcp-swagger-server

✕BLOCK F

MCP Swagger Server 将任何符合 OpenAPI/Swagger 规范的 REST API 转换为 Model Context Protocol (MCP) 格式,让 AI 助手能够理解和调用您的 API。

★ 7715 readsse

Vault

zycaskevin/vault-agent-memory

✕BLOCK F

Local-first memory governance for AI agents: shared, reviewable, auditable memory via SQLite and MCP.

★ 501 write

Vault for LLM

zycaskevin/vault-for-llm

✕BLOCK F

Local-first memory governance for AI agents: shared, reviewable, auditable memory via SQLite and MCP.

★ 501 write

openrod

OpenRod/OpenRod

✕BLOCK F

One-click sandboxes for AI agents, with egress policy, MCPs, skills and audit. Built on NVIDIA OpenShell.

★ 173 readstreamable-http

willow-mcp

willow-memory/willow-mcp

!CAUTION F

Agent-neutral MCP server with persistent memory (SOIL + Postgres KB) and a sandboxed task queue. Manifest-based ACL; works with any stdio MCP client.

★ 35 writestdio

codemcp

mewisme/codemcp

✕BLOCK F

CodeMCP is a workspace-aware execution, context, and agent orchestration server for AI clients.

★ 23 readstreamable-http

agentic-engineering-harness

JamesMorales04/agentic-engineering-harness

✕BLOCK F
★ 04 write

kleo-mcp

tonnooooo/kleo-mcp

✕BLOCK F

Kleo: remote MCP server that turns a brief into a narrated 4K 60 fps film, realistic or animation, from Claude, ChatGPT, Cursor or any MCP client

★ 02 writestreamable-http

Tabularis

tabularisdb/tabularis

✕BLOCK F

Open-source desktop SQL workspace with 3 built-in database drivers and 16 shipped plugins, including SQL Server, DuckDB, ClickHouse and Redis. Built-in MCP server for Claude, Cursor and Devin, SQL notebooks and visual EXPLAIN.

★ 5,0116 read

TradeMemory

mnemox-ai/tradememory-protocol

✕BLOCK F

Decision audit trail + persistent memory for AI trading agents. Outcome-weighted recall, tamper-evident SHA-256 chain with RFC 3161 anchoring, 20 MCP tools.

★ 1,4203 writestdio

sigmap

manojmallick/sigmap

!CAUTION F

~97% token reduction for AI coding sessions — zero deps, 38 languages, MCP server

★ 6462 write

Openllmetry Js

traceloop/openllmetry-js

✕BLOCK F

Sister project to OpenLLMetry, but in Typescript. Open-source observability for your LLM application, based on OpenTelemetry

★ 4121 write

Payload CMS

matmax-worldwide/payloadcmsmcp

✕BLOCK F

Payload CMS MCP Server

★ 1239 readsse

codelocal

codelocal-cloud/codelocal

✕BLOCK F

The open-source Project Brain for AI coding. Connect ChatGPT, Codex, Claude and MCP-compatible agents to your projects, local tools and persistent knowledge. Fork it, build with it, and help shape the open AI development stack. 🧠⚡

★ 872 readstdio

Last.fm MCP Server

rianvdm/lastfm-mcp

✕BLOCK F

An MCP server that provides seamless access to a user's Last.fm listening data and music information via AI assistants like Claude.

★ 4921 readstreamable-http

Tushare

zhewenzhang/tushare_mcp

✕BLOCK F
★ 4866 readstreamable-http

api2ai

OpenSourceAGI/dev-tools-starter-agent

✕BLOCK F

💻⚒️ NPM dev tools and core parts for Full Stack Starter App & Docs 🚀DEMO

★ 389 read

Codingbuddy

jeremydev87/codingbuddy

✕BLOCK F

Codingbuddy orchestrates 29 specialized AI agents to deliver code quality comparable to a team of human experts through a PLAN → ACT → EVAL workflow.

★ 318 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 (1,620 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.