Stop Searching. Start Trusting.

The curated directory of MCP servers, vetted for security, efficiency, and quality.

Browse DirectoryWhy Us?

Tired of the MCP "Marketplace" Chaos?

We built MCPScout.ai to solve the ecosystems biggest pain points.

No Insecure Dumps

We manually analyze every server for basic security flaws.

Easy Setup

Our gotcha notes warn you about complex setups.

Avoid "Token Hogs"

We estimate token costs for cost-effective agents.

Products, Not Demos

We filter out "Hello World" demos.

CATEGORIES:
SORT:

Vetted Servers(8554)

0
0
Medium Cost
shields icon

lgtmcp

by shields

Sec9

An AI-powered Model Context Protocol (MCP) server that reviews code changes and can automatically commit them or provide detailed feedback.

Setup Requirements

  • ⚠️Requires a Google API Key from Google AI Studio (paid usage applies).
  • ⚠️The `lgtmcp` binary is installed to `~/bin` by default and requires this directory to be in your system's PATH.
  • ⚠️Must be run within a Git repository with staged or unstaged changes.
Verified SafeView Analysis
The server demonstrates robust security practices, particularly in file access and path validation. The `handleFileRetrieval` function in `review.go` includes critical checks for path traversal vulnerabilities (`..`), resolves absolute paths, ensures access is within the repository boundaries, respects `.gitignore` rules using `git check-ignore`, and explicitly handles symlinks to prevent them from pointing outside the repository. Gitleaks is integrated for built-in secret detection on changed code, and the tool uses `exec.CommandContext` for git operations with timeouts. Test secrets are ROT13 encoded to prevent accidental triggers. Authentication to the Gemini API relies on user-provided API keys or Application Default Credentials, preventing hardcoded credentials within the application itself.
Updated: 2026-01-18GitHub→
0
0
Low Cost
nainaasharmaa icon

armoriq-mcp-server

by nainaasharmaa

Sec2

A basic banking server managing accounts, deposits, withdrawals, balances, and transaction history via a RESTful API.

Setup Requirements

  • ⚠️No User Management/Authentication: Accounts are identified solely by numerical IDs, allowing unrestricted access and manipulation without any security checks.
  • ⚠️Local SQLite Database: The system uses an ephemeral SQLite file (bank.db) for storage, unsuitable for persistent, concurrent, or production-grade financial operations.
  • ⚠️Basic Input Validation: The `amount` parameter for financial transactions lacks checks for non-positive values, potentially enabling unintended deposits/withdrawals or balance exploitation.
Review RequiredView Analysis
The server critically lacks any form of authentication and authorization, allowing any client to access and modify any account by knowing its ID. Input validation for `amount` in deposit/withdraw endpoints is insufficient, as it does not prevent negative values, which can lead to unintended balance manipulation (e.g., a negative deposit becomes a withdrawal). No bounds checking for transaction amounts is present. The use of a local SQLite database (bank.db) implies lack of encryption at rest and is not suitable for high-security or concurrent production environments.
Updated: 2026-01-19GitHub→
0
0
Medium Cost
bgbruno icon
Sec8

MCP server for managing and interacting with long-running shell processes (e.g., dev servers, watch modes) in real-time, providing output streaming and interactive control for AI agents.

Setup Requirements

  • ⚠️Requires Node.js >= 18.17.
  • ⚠️Requires native build tools (like Python, C++ compiler) for 'node-pty' dependency installation during `npm install`.
  • ⚠️Assumes a trusted MCP client and operating environment due to its arbitrary command execution capabilities.
Verified SafeView Analysis
The server's core function is to execute arbitrary shell commands (`cmd`, `args`) provided by the MCP client. This is an intended capability inherent to a 'shell server' and not a vulnerability in itself, but it grants significant power. If the MCP client or the AI agent interacting with it is untrusted or compromised, it can lead to arbitrary code execution on the host system. Communication occurs over standard I/O (stdio), not directly via network ports, meaning the security model relies on the integrity of the process launching and communicating with this server. No 'eval', obfuscation, or hardcoded secrets were found.
Updated: 2026-01-07GitHub→
0
0
Low Cost
kamikaz-e icon

MoonPhasesMCP

by kamikaz-e

Sec3

Provides a server with an API, likely for a game or application related to moon phases, potentially within a Minecraft context (given 'MCP').

Review RequiredView Analysis
CRITICAL: No source code was provided for analysis beyond the README.md file. Therefore, a comprehensive security audit for 'eval', obfuscation, network risks, hardcoded secrets, or malicious patterns cannot be performed. The assigned score reflects this critical lack of information, not an actual finding of vulnerabilities.
Updated: 2025-11-24GitHub→
0
0
High Cost
devnullvoid icon

mcp-ssh-session

by devnullvoid

Sec8

Enables AI agents to establish and manage persistent SSH sessions for remote command execution and file operations on servers and network devices.

Setup Requirements

  • ⚠️Requires SSH access to target hosts, including valid credentials (username, password, or SSH key).
  • ⚠️Python 3.10 or higher is required.
  • ⚠️When run in Docker, proper mounting of `~/.ssh` directory or individual SSH config/key files is critical for authentication.
Verified SafeView Analysis
The server implements several security measures including running as a non-root user in Docker, read-only mounting of SSH keys/config, output/file size limiting to prevent memory exhaustion, and using base64 for file content during `sudo tee` operations to mitigate shell injection risks. It also explicitly mentions and has fixed a prompt injection vulnerability. However, it uses Paramiko's `AutoAddPolicy` for host keys, which automatically accepts new host keys. While convenient for automation, this bypasses strict host key verification and can be susceptible to MITM attacks if the initial connection is compromised. This is a known trade-off clearly stated in the documentation.
Updated: 2025-12-12GitHub→
0
0
Low Cost
zv-louis icon

mcp7zop

by zv-louis

Sec8

A local MCP server for performing 7-Zip archive and file system operations.

Setup Requirements

  • ⚠️Requires Python 3.12+
  • ⚠️Requires 7-Zip (>=23.0) to be installed locally and configured (either via system PATH or a configuration file at `~/.mcp7zop/config.json`).
  • ⚠️Requires 'uv' package manager for installation.
Verified SafeView Analysis
The server interacts with the local 7-Zip command-line interface via `asyncio.create_subprocess_exec`. This method is generally secure against shell injection when commands and arguments are passed as a list. The primary security consideration is ensuring the configured 7-Zip executable path (either via PATH or `~/.mcp7zop/config.json`) points to a trusted binary, as a malicious configuration could lead to arbitrary command execution. Input paths are validated for existence before being passed to 7-Zip. No 'eval' or obvious obfuscation patterns were found.
Updated: 2026-01-10GitHub→
0
0
Low Cost
tejzpr icon
Sec10

This server provides a comprehensive suite of 70+ mathematical operations through the Model Context Protocol (MCP) for integration with AI assistants and compatible clients.

Setup Requirements

  • ⚠️Requires Go 1.23 or later if running directly from source/go modules.
  • ⚠️For HTTP transport, ensure port 8080 (or `MATH_PORT` specified) is open.
Verified SafeView Analysis
The source code appears to be very secure. It is written in Go, which provides strong type safety and memory management. Input validation is performed for all tool parameters (e.g., `RequireFloat`, `RequireInt`, checking for non-negative numbers where appropriate, division by zero). Critical numerical operations like factorial and Fibonacci use `math/big` for arbitrary precision integers and include resource limits. Error messages returned to the client are generic and do not expose internal stack traces. There are no indications of code injection vulnerabilities, `eval` usage, insecure deserialization, hardcoded secrets, or unintended system access. The server uses standard Go `math` and `math/cmplx` packages for computations and `mcp-go` for protocol handling.
Updated: 2026-01-17GitHub→
0
0
High Cost
chernistry icon

ohmyrepos

by chernistry

Sec8

The MCP Server integrates Oh My Repos' semantic search, RAG chat, and similarity tools with AI coding assistants like Claude Desktop and Cursor.

Setup Requirements

  • ⚠️Requires Python 3.11+
  • ⚠️Requires Qdrant instance (cloud or local Docker)
  • ⚠️LLM and Embedding model choices (e.g., Ollama locally or external APIs like Jina/OpenAI) incur varying costs and setup complexity. Ollama requires a local server running.
  • ⚠️Initial ingestion of repositories triggers embedding/LLM calls, which can be computationally intensive and costly depending on the number and size of repositories.
Verified SafeView Analysis
The server uses Pydantic's `SecretStr` and supports `env`, `keyring`, or `aws` as `SECRETS_BACKEND` for managing API keys and tokens, which is a good security practice. It also implements in-memory rate limiting and token budget tracking to prevent abuse and control costs. Network communication relies on external services (GitHub, LLMs, Qdrant) which inherently carry some risk, but the implementation appears to follow standard secure coding practices. No direct use of `eval` or obfuscation found.
Updated: 2025-11-22GitHub→
0
0
Medium Cost
ZCAD-Products icon

mcp-jobboss2-server

by ZCAD-Products

Sec9

Integrate AI assistants with JobBOSS2 ERP for managing manufacturing operations, including orders, customers, quotes, materials, and employee attendance.

Setup Requirements

  • ⚠️Requires a JobBOSS2 Cloud account with API access enabled (contact ECI Solutions to obtain).
  • ⚠️Requires JobBOSS2 API credentials (API Key, API Secret, and OAuth2 Token URL) which must be configured as environment variables.
Verified SafeView Analysis
The server uses standard OAuth2 for authentication, managing tokens securely. API calls are made using `axios` (TypeScript) or `httpx` (Python), both well-regarded libraries. The `apiCall` method correctly prefixes endpoints with `/api/v1/` and uses `encodeURIComponent` for path parameters, mitigating basic path traversal and injection risks. No direct `eval` or `child_process` execution, or hardcoded sensitive credentials were found. The `custom_api_call` tool allows arbitrary API interaction, which grants broad access and necessitates careful control over the AI's capabilities and trust in its use.
Updated: 2025-12-17GitHub→
0
0
Medium Cost
cnitlrt icon
Sec2

Provides an MCP server to enable AI agents or other clients to perform dynamic debugging using GDB/pwndbg.

Setup Requirements

  • ⚠️Requires Python 3.13 or higher.
  • ⚠️Requires GDB (preferably pwndbg) to be installed and accessible in the server's PATH.
  • ⚠️The `HOST` environment variable defaults to '0.0.0.0', which exposes the server to all network interfaces. This can be a security risk if not explicitly intended.
Review RequiredView Analysis
The `execute` tool allows arbitrary GDB/pwndbg commands to be executed. GDB includes a `shell` command that can execute arbitrary system commands on the host machine. This poses a critical security risk, as a malicious client or compromised LLM could leverage this to gain remote code execution with the privileges of the server process. If the server is exposed to untrusted networks or users, or runs with elevated privileges, this vulnerability is severe. There are no obvious hardcoded secrets or obfuscation, and environment variables are used for network configuration.
Updated: 2025-11-30GitHub→
0
0
Low Cost
metaif icon
Sec9

Provides a Dockerized wrapper for the Longport Model Context Protocol (MCP) server for easy deployment and management of Longport API services.

Setup Requirements

  • ⚠️Requires Docker Engine (20.10.0 or later)
  • ⚠️Requires Docker Compose (1.29.0 or later) for Docker Compose setup
  • ⚠️Requires Longport API credentials (APP_KEY, APP_SECRET, ACCESS_TOKEN) which must be obtained separately from Longport
Verified SafeView Analysis
The Docker wrapper itself follows good security practices: credentials are handled via environment variables with a clear warning not to commit the .env file, the service defaults to a 'read-only' mode, and there are no apparent 'eval' or malicious patterns in the provided scripts. Network binding is to all interfaces (0.0.0.0:8000), but the README correctly advises on firewall configuration for public exposure. The core 'longport-mcp' binary's source is not provided in this repo, so its internal security cannot be fully audited here.
Updated: 2025-11-20GitHub→
0
0
Low Cost
Manideep337 icon

add_mcp_server

by Manideep337

Sec1

This project provides a utility to assist in the setup and management of a Minecraft (MCP) server environment.

Setup Requirements

  • ⚠️Requires Java Runtime Environment (JRE) installed.
  • ⚠️May require specific system dependencies (e.g., `screen`, `curl`, `wget`).
  • ⚠️Requires an existing Minecraft server jar or components to manage.
Review RequiredView Analysis
CRITICAL: No source code was provided for analysis. Therefore, a comprehensive security audit could not be performed. It is impossible to verify the absence of 'eval', obfuscation, network risks, hardcoded secrets, or malicious patterns. Running unverified code poses significant security risks.
Updated: 2025-11-24GitHub→
PreviousPage 402 of 713Next