MCP servers that ship as code — a repository or a package you install and run locally (usually over stdio). They have no hosted endpoint, so we cannot probe or grade them: instead each one shows its repository, install command, license, stars and when it was last updated.
Collected from GitHub repositories and npm, PyPI, Cargo, Go, Maven, NuGet and RubyGems packages tagged as MCP servers, the official MCP Registry, Docker's MCP catalog, ModelScope, Baidu MCP World, Cline and other public directories; one entry per repository. A server that also offers a hosted endpoint is listed under Remote instead. These entries are not counted in our reachability, verification or readiness statistics.
-
Official One Identity Safeguard MCP server — AI-powered privileged access management with 1,000+ API endpoints
npm Apache-2.0 -
hipaa-compliance-mcpgithub.com/crawde/hipaa-compliance-mcpupdated this year
MCP server for HIPAA compliance — browse safeguard requirements, assess compliance readiness, generate policy templates, evidence checklists, and gap analysis for healthcare organizations
npm MIT -
Pure Data Provider architecture serving authentic CIS Controls Framework data via MCP and HTTP API. Empowers LLMs with authoritative safeguards data for analysis. Supports Microsoft Copilot custom connectors and cloud deployments.
npm CC-BY-4.0 -
glba-compliance-mcpupdated this year
MCP server for Gramm-Leach-Bliley Act (GLBA) compliance — financial privacy safeguards, risk assessment, information security program requirements, vendor management, and incident response for financial institutions
npm MIT -
MCP server for MongoDB with safeguards against accidental data deletion
-
@websms-nz/websms-mcphttps://gitlab.com/icepicknz/websms-mcpupdated this year
WebSMS MCP server for Claude - Send SMS messages with safeguards
npm MIT -
Persistent SSH sessions for DeepSeek Harness: credential profiles, sudo and root safeguards, fail2ban-aware retry limits, and SFTP.
MIT -
MCP server for iMessage, Apple Notes, and Contacts on macOS. Read-only by default, with prompt-injection safeguards.
MIT -
mcp-mongo-configgithub.com/hexonal/mcp-mongo-configupdated this year
Secure MongoDB MCP server for AI assistants with async operations and read-only-by-default safeguards.
-
🔌 Exposes HIPAA healthcare compliance over MCP — Administrative, Physical, Technical safeguards assessment, PHI handling, BAA templates, breach notification readiness.
pypi MIT -
Local MCP messaging server for macOS Messages and Google Voice with confirmation, idempotency and recipient safeguards.
-
An MCP server that safeguards AI-driven SQLite database writes with impact previews, risk-based approvals, policy enforcement, tamper-evident audit logs, and undo support.
pypi MIT -
mcp-server-fetch-tomupdated this year
A Model Context Protocol server providing tools to fetch and convert web content for usage by LLMs, with prompt injection safeguards
pypi MIT -
🚀 Professional Model Context Protocol (MCP) server for LinkedIn with FastMCP, Playwright, and Human-in-the-Loop safeguards.
MIT ★ 6 -
cdq-mcpgithub.com/mearnsb/cdq-mcpupdated this year
Collibra Data Quality (Classic) MCP Server - HTTP Streamable server for running SQL, adding rules, and creating DQ jobs. 100% REST API. Leverages built-in application level access controls, masking, and SQL safeguards.
★ 2 -
MCP server that lets Claude Code and Claude Desktop drive the jev-ultrafast browser agent in your real Chrome, with a domain-bound safeguard
MIT -
MCP server for Outlook mail and Microsoft Teams via Microsoft Graph, with your own sign-in and preview-before-send safeguards
npm MIT -
Cloudflare to GitHub Backupno update data
Automate the backup of your Cloudflare projects to a GitHub repository. Ensure your projects are securely stored and easily retrievable with this seamless integration. Start safeguarding your Cloudflare data today with minimal setup required.
smithery -
Open-source sovereign workspace for persistent AI coworkers with isolated sandboxed execution, living spaces, and human-in-the-loop safeguards.
MIT ★ 1 -
MCP server for Google Ads - campaign management, reporting, and optimization with safeguards for review before changes go live
npm MIT ★ 1 -
Unofficial Firefox extension + MCP server that lets Claude Code control Firefox, with a sidebar chat panel and prompt-injection safeguards.
MIT -
A browser that AI agents drive through MCP tools and a person can watch and control, with step safeguards, prompt-injection defence, a live viewer and an Electron app
-
windows-mcp-servergithub.com/AhmedLaminou/windows-mcp-serverupdated this year
Windows MCP server with 96 local tools for task management, diagnostics, cleanup, hardware/GPU, security, updates, services, networking, Registry, PowerShell, WSL/Hyper-V, printers, and dev workflows. Includes read-only reports, dry-run safeguards, and safer Recycle Bin cleanup
pypi MIT ★ 4 -
Stdio MCP Server wrapping custom Python runtime (LocalPythonExecutor) from Hugging Faces' `smolagents` framework. The runtime combines the ease of setup (compared to docker, VM, cloud runtimes) while providing safeguards and limiting operations/imports that are allowed inside the runtime.
MIT ★ 48 -
MCP server that lets AI agents drive your real Safari on macOS — per-host read-only policy, prompt-injection safeguards, real and safaridriver backends.
MIT -
ai-employee-hackathon-2026github.com/MrAfoo/ai-employee-hackathon-2026updated this year
Autonomous AI Employee: monitors Gmail, WhatsApp & finances using Groq LLM. Bronze→Silver→Gold→Platinum tiers with Obsidian vault, HITL safeguards, MCP servers & Oracle VM cloud support. 100% free
MIT ★ 1 -
Drive a real Telegram account from an MCP client: chats, messages, media, secret chats, communities, folders, admin rights, invite links and Mini Apps, with multi-account routing and an approval safeguard.
GPL-3.0 ★ 32 -
Claude Code MCP is an implementation of Claude Code as a Model Context Protocol (MCP) server. It exposes Claude Code–style capabilities (file operations, shell execution with safeguards, prompts, and resources) through the standardized MCP interface so that any MCP‑compatible client can integrate with it.
-
Public beta: A Codex plugin for project-based AI-team collaboration with shared meeting rooms, task ownership, recovery, QA safeguards, usage reports, and Unity validation support.
MIT ★ 2 -
MwalimuLens is an evidence-grounded educational AI agent that analyses learning records across terms, identifies potential competency patterns and prepares source-linked insights for teacher review, with human oversight and safeguards against premature learner judgements.
Apache-2.0 -
MCP server for UN human-rights voting records: how States voted at the Commission on Human Rights, the Human Rights Council and the General Assembly Third Committee. Offline, with the dataset's safeguards built in.
NOASSERTION -
GoogleWorkspaceAuditgithub.com/sean-m-sweeney/GoogleWorkspaceAuditupdated this year
Automated compliance auditing for Google Workspace using Claude's MCP. Performs 19 security checks covering access control, authentication, and system protection. Maps findings to CMMC, NIST 800-171, NIST CSF, ISO 27001, HIPAA, and FTC Safeguards frameworks. Built for MSPs and Workspace Admins to streamline security assessments.
★ 5 -
Safeguarded AI research agent: Claude Code runs headless with one tool, a custom MCP fetch gate that masks personal information, while code checks every answer and a person reviews the results. Built for a local business in the Vic Park AI Business Boost program.
MIT -
Tetheredno update data
Control Mac power modes, prevent sleep, keep your Mac awake, and manage background apps during AI coding, builds, rendering, and local model inference. Tethered lets your AI assistant discover and apply saved profiles for Low Power, Automatic, or High Power mode where supported; display brightness; fan settings; and keep-awake behavior. Profile settings can deprioritize or suspend selected background apps, including policies triggered by memory pressure. Inspect memory pressure, RAM and swap usage, session status, and task safeguard warnings. Finite sessions expire automatically; unlimited sessions require an explicit end. Power settings describe requested controls, not independently measured hardware state. Requires an Apple Silicon Mac running macOS 15 or later, Tethered 1.2.0+ installed and running under the same macOS account, Pro access, AI task sessions enabled, and permissions required by the selected profile. Runs locally over stdio. Intel Macs and cloud-only clients are unsupported. No connection credentials or saved profiles are included.
smithery