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/101

Verdicts · 3,145 servers

✓SAFE1,632!CAUTION521✕BLOCK992

Transports

A server can list more than one transport.

3,145 servers

2,401–2,424 of 3,145 · ranked by trust score, then stars

memex

STiFLeR7/memex

✕BLOCK D

Persistent memory for AI coding agents via MCP: a bitemporal knowledge graph of your codebase, served to Claude Code, Cursor, Gemini CLI, and any MCP client. Tree-sitter and Gemini Flash into Neo4j via Graphiti. 14 MCP tools, hierarchical clusters, three-regime confidence decay.

★ 123 writestdio

lain

spuentesp/lain

✕BLOCK D

High-performance MCP server for AI coding agents: persistent knowledge graph, blast radius analysis, semantic search. Built in Rust.

★ 94 writestdio

mcp-server

Coinversaa/mcp-server

✕BLOCK D

Hyperliquid trading intelligence for AI agents — 106 available MCP tools: builder analytics (revenue, retention, audience overlap, user journey), cohort analytics, position lifecycles, liquidation heatmaps, trader profiling, HIP-4 outcome contracts, and live market data. Hosted connector at mcp.coin

★ 91 writestdio

warden-ai

rynald0cst0ltziam/Warden-AI

✕BLOCK D

The structurally-verified context layer for AI coding agents. Cut 50-90% of your agent's token spend.

★ 94 writestdio

aura

trotor/aura

✕BLOCK D

Suomalaisen avoimen datan discovery- ja ymmärryspalvelu — MCP-server tekoälyille, avoin web-palvelu ihmisille

★ 81 writestdio

agent-discover

keshrath/agent-discover

✕BLOCK D

MCP server registry and marketplace — discover, install, activate, and manage MCP tools on demand

★ 79 writestdio

aikstockdata

na77tech-creator/aikstockdata

✕BLOCK D

Korean stock data for AI: KOSPI/KOSDAQ/KONEX closes, DART disclosures with receipt time, post-filing price paths. JSON + remote MCP server, no API key.

★ 78 readstreamable-http

Epoch

KyaniteLabs/Epoch

✕BLOCK D

Time estimation MCP server for AI agents: PERT, COCOMO II, Monte Carlo, sprint forecasting, token-to-time mapping, cost estimation, and schedule risk tools.

★ 73 readstdio

arr-mcp

Knuckles-Team/arr-mcp

✕BLOCK D

Arr related MCP and A2A / AG-UI Agent

★ 51 readsse

blender-mcp

Vertiiii/blender-mcp

✕BLOCK D

🛠️ Connect Blender to Claude AI for seamless 3D modeling and scene manipulation using the Model Context Protocol for enhanced creative workflows.

★ 53 write

decisions-judge-mcp

clouatre-labs/decisions-judge-mcp

✕BLOCK D

Typed decisions for AI agents as an MCP tool: yes/no probability (noul), choice, and score in one fast request. Backed by the TypeSafe System One model.

★ 54 readstdio

EPICS-MCP

epicDirk/EPICS-MCP

✕BLOCK D

Ask an AI assistant about your EPICS control system: live PV values, why a PV is disconnected, and where it appears across ChannelFinder, the Archiver and the alarm tree. Read-only by default.

★ 53 readstreamable-http

sandrone

kuuvahki-labs/sandrone

✕BLOCK D

Self-hosted proxy configuration workbench for aggregating subscriptions, probing nodes, and generating Mihomo, sing-box, and Shadowrocket configs.

★ 51 writestreamable-http

OpenAPI to Tools

buildwithlayer/openapi-to-tools

✕BLOCK D
★ 43 read

assemblash

VidGuiCode/assemblash

✕BLOCK D

Local-first visual document engine with a web editor, CLI, API, and MCP server - built for humans and AI agents.

★ 44 writestreamable-http

KGraph

ajksunkang-aios/KGraph

✕BLOCK D

The Compiler-Aware Operating System Kernel CodeGraph Engine for Token-Efficient Structural Codebase Retrieval

★ 413 read

LibreGene

dl-li/LibreGene

✕BLOCK D

An open-source, agent-ready plasmid editor.

★ 42 readstreamable-http

libreoffice-mcp

sandraschi/libreoffice-mcp

✕BLOCK D

LibreOffice headless automation via FastMCP — Writer/Calc/Impress convert, ODT merge, PDF combine, folder watch. For Cursor / Claude Desktop.

★ 43 readstdio

pkgtruth

hxckya/pkgtruth

✕BLOCK D

Ground truth about npm packages for AI coding agents. Catches hallucinated and slopsquatted dependencies before they reach your lockfile.

★ 41 writestdio

Angular Starter

4sh/starter-angular

✕BLOCK D

An Angular component library that natively integrates your design tokens to create a seamless bridge between design and code.

★ 37 readstdio

Neta

sudhanshug16/neta

✕BLOCK D

A leader agent that delegates to tiered worker agents behind the coding-agent CLI you already use — Claude Code, Codex, or OpenCode.

★ 33 write

agent-harness-generator

Daffiburrlike452/agent-harness-generator

✕BLOCK D

Create custom AI agent harnesses for any repository using this CLI generator.

★ 38 writestreamable-http

audio-transcriber

Knuckles-Team/audio-transcriber

✕BLOCK D

Transcribe audio into text. Agentic AI supported through MCP Server.

★ 31 readsse

cal

Dvorinka/cal

✕BLOCK D

Self-hosted daily toolkit — calendar, tasks, notes, links, files, kanban boards, time tracking, GitHub inbox. One Docker container.

★ 31 read

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.