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(9120)

0
0
Medium Cost
ismaeldosil icon
Sec8

The BeneFisio Coordination MCP Server provides a centralized system for multi-agent coordination, managing tasks, communication, and system state using a SQLite database and exposing tools via the Model Context Protocol.

Setup Requirements

  • ⚠️Requires Node.js v20.0.0 or higher.
  • ⚠️Hardcoded paths for `finashopping-docs` (e.g., `/Users/admin/Projects/Precisa/finashopping/finashopping-docs`) exist in `src/tools/agents.ts` and `src/generators/markdown-generator.ts` and will need adjustment to match deployment environment unless the relative paths resolve correctly.
  • ⚠️Requires specific Claude Desktop configuration (`claude_desktop_config.json`) for full integration as outlined in the README.
Verified SafeView Analysis
The server appears well-designed with strong internal security practices, including explicit fixes for TOCTOU race conditions (BUG-001) using SQLite transactions and proper parameterization to prevent SQL injection. It utilizes environment variables for sensitive paths and webhook URLs, which is good practice. The HTTP health endpoint has broad CORS ('Access-Control-Allow-Origin: *'), which is acceptable for monitoring but should be noted. The primary external risk would be if environment variables for webhooks (`BACKUP_ALERT_WEBHOOK`, `MONITOR_ALERT_WEBHOOK`) are misconfigured or controlled by untrusted entities, potentially leading to data exfiltration or SSRF, though this assumes an compromised environment setup.
Updated: 2026-01-19GitHub→
0
0
Medium Cost
Akungapaul icon

wp-plugins-mcp

by Akungapaul

Sec8

Manages WordPress plugins, including listing, activating, deactivating, installing, updating, and deleting them, primarily for automation or agent-driven operations.

Setup Requirements

  • ⚠️Requires a running WordPress instance with REST API enabled and an Application Password configured for the specified user.
  • ⚠️WP-CLI must be installed and correctly configured (potentially with SSH access) on the WordPress server if WP-CLI dependent tools (install, delete, update) are to be used.
  • ⚠️Requires Node.js version 18 or higher based on SDK dependencies.
Verified SafeView Analysis
Sensitive WordPress and WP-CLI credentials are correctly externalized as environment variables. The server enables powerful plugin management operations (install, delete) which require careful external access control to the MCP server endpoint. No eval, code obfuscation, or immediately malicious patterns are present.
Updated: 2025-11-28GitHub→
0
0
Medium Cost
KhaiEr720 icon

AI-Skills

by KhaiEr720

Sec8

A framework and collection of tools for developing, validating, and managing modular AI skills to enhance AI agent capabilities.

Setup Requirements

  • ⚠️Requires Python 3 to run skill development and utility scripts.
  • ⚠️Requires the Pillow Python library for image generation.
  • ⚠️Requires the CairoSVG Python library for SVG icon input; otherwise, SVG must be manually converted to PNG.
Verified SafeView Analysis
The Python scripts primarily handle file system operations, YAML parsing, and image processing. The `generate_icons.py` script leverages third-party libraries (`Pillow`, optional `CairoSVG`) which could be vulnerable to malformed input, a common risk for media processing. However, no direct `eval`, `os.system` with unsanitized input, or hardcoded secrets are present in the provided code. The system is designed to be executed within an AI agent environment, implying potential external sandboxing.
Updated: 2026-01-19GitHub→
0
0
Low Cost
farhanrhine icon
Sec9

This server functions as an expense tracker application, providing both a REST API and an MCP (Model Context Protocol) interface for integration with Claude Desktop.

Setup Requirements

  • ⚠️Requires Python 3.11+
  • ⚠️Requires the 'uv' package manager for installation and execution
  • ⚠️For Claude Desktop integration, specific `fastmcp install` command and Claude Desktop restart are needed.
Verified SafeView Analysis
The server uses parameterized SQL queries for all database interactions, effectively preventing SQL injection vulnerabilities. No 'eval' or similar dangerous functions are used. No hardcoded credentials or sensitive information are present in the provided source code. The database is a local SQLite file, limiting external exposure.
Updated: 2025-11-29GitHub→
0
0
Medium Cost
Jaizkibel icon

mcp-server

by Jaizkibel

Sec1

Provides a Model Context Protocol server that offers AI assistants tools for web search, database interaction, and Java project source/Javadoc retrieval in a secure environment.

Setup Requirements

  • ⚠️Python 3.14+ required
  • ⚠️Brave Search API Key (Paid) required for web search functionality
  • ⚠️Requires PostgreSQL or SQL Server for database tools
  • ⚠️Requires `~/.mcp-server/config.yml` configuration file
  • ⚠️Java Development Kit (JDK) for Java build tools functionality
Review RequiredView Analysis
Critical SQL Injection vulnerability: The `execute_sql_statement` function directly executes user-provided SQL queries without parameterization, allowing for arbitrary SQL commands to be run on configured PostgreSQL or SQL Server databases. Network vulnerability: The HTTP client used for web searches and `http_get_request` disables SSL certificate validation (`verify=False`), making it vulnerable to Man-in-the-Middle attacks. Potential browser command injection/arbitrary file access: The `open_in_browser` tool uses `subprocess.Popen` with weakly validated URLs, potentially allowing opening of local files or execution of other commands.
Updated: 2026-01-05GitHub→
0
0
Low Cost
yumjbonddigit icon

mcp_server_sample

by yumjbonddigit

Sec4

Accessing financial data and news for bond issuers via BondDigit's MCP server.

Setup Requirements

  • ⚠️Requires Python 3 installed.
  • ⚠️Requires installing the 'mcp' Python package via pip.
  • ⚠️Requires obtaining an API token by contacting BondDigit marketing and manually inserting it into 'mcp_client.py'.
Verified SafeView Analysis
A hardcoded API token is required to be manually inserted into `mcp_client.py`, which is a significant security risk if the code is shared or committed to version control. The client interacts with an external API (`https://mcpapi.bonddigit.com/sse`) whose server-side code is not provided for audit.
Updated: 2025-11-26GitHub→
0
0
High Cost
Readify-App icon

scraping-mcp-server

by Readify-App

Sec2

Web scraping for both static and dynamic web pages, including site structure analysis, content extraction, and integration with WordPress APIs (Rakuraku Media School, Cloud GYM) and Google Sheets.

Setup Requirements

  • ⚠️Playwright browser installation is required (downloads Chromium, a potentially large component).
  • ⚠️Google Sheets API integration requires setting up a Google Service Account and placing its JSON key file (named `braided-circuit-465415-m6-1cbbf338d9f0.json`) in specific directories or providing its path via the `GOOGLE_APPLICATION_CREDENTIALS` environment variable.
  • ⚠️The server relies on `uv` for Python environment management, which is installed by the one-command script but would be a manual dependency if setting up from scratch.
  • ⚠️The server is specifically designed to run as an MCP server within Claude Desktop.
Review RequiredView Analysis
Hardcoded WordPress API credentials for 'Rakuraku Media School' are present in `server.py` (RAKURAKU_WP_USERNAME, RAKURAKU_WP_APP_PASSWORD). While environment variables can override these, their presence in source code is a significant security vulnerability. Additionally, the Google Sheets API requires a service account JSON file (`braided-circuit-465415-m6-1cbbf338d9f0.json`) whose name is hardcoded and searched in specific local paths, which could pose a risk if not managed securely.
Updated: 2026-01-12GitHub→
0
0
Low Cost
Sec5

Integrate a local Kali MCP (Machine Control Protocol) server with LM Studio to run AI models locally on Kali Linux.

Setup Requirements

  • ⚠️Requires Kali Linux (or a compatible Debian-based distribution) for `mcp-kali-server`.
  • ⚠️Manual replacement of `/absolute/path/to/mcp_server.py` with the actual path in `mcp.json` is critical and prone to user error.
  • ⚠️Significant local hardware resources (RAM, CPU, potentially GPU) are implicitly required for effective local LLM inference.
Review RequiredView Analysis
The setup relies on installing an 'mcp-kali-server' package from apt, whose internal workings are not detailed. Crucially, it involves cloning a GitHub repository from an unverified user ('Wh0am123') and then having LM Studio execute the 'mcp_server.py' script from that repository. This direct execution of external code introduces a significant supply chain risk; if the script or the apt package were malicious, the system could be compromised. While network traffic is local, the trust in the source code is a primary concern.
Updated: 2025-11-17GitHub→
0
0
Medium Cost
privacy-posture icon

gdpr-mcp

by privacy-posture

Sec9

A local Model Context Protocol (MCP) server for augmenting LLMs with GDPR document search capabilities using hybrid trigram and vector search.

Setup Requirements

  • ⚠️Requires Go 1.21+
  • ⚠️Requires GCC for SQLite (CGO) build dependencies
  • ⚠️OpenAI API Key is optional but recommended for semantic search, which incurs usage costs.
Verified SafeView Analysis
The server primarily operates via stdin/stdout for MCP communication, limiting direct network exposure. It uses SQLite with CGO, a standard but slightly more complex component. OpenAI API key is fetched from environment variables, which is good practice. No 'eval' or similar dynamic code execution or obfuscation is apparent. Network requests are limited to the optional OpenAI embeddings API. The 'SearchVectors' function reads all embeddings into memory for similarity calculation, which could be a denial-of-service risk for very large databases due to memory consumption, but this is an efficiency concern rather than a direct security vulnerability given the expected usage with a single GDPR document.
Updated: 2025-11-29GitHub→
0
0
Medium Cost
dnaerys icon

onekgpd-mcp

by dnaerys

Sec3

Provides natural language access and filtering to the 1000 Genomes Project dataset for genomic variant and sample information.

Setup Requirements

  • ⚠️Requires JRE 21 to run locally (Java 21)
  • ⚠️Depends on an external Dnaerys variant store service via gRPC (db.dnaerys.org:7443)
  • ⚠️Maximum 50 variants or items returned per 'select' call, requiring pagination for larger results
Review RequiredView Analysis
The gRPC client uses a `TrustManager` that explicitly trusts all certificates when SSL is enabled, effectively disabling proper TLS certificate validation. This makes the connection vulnerable to Man-in-the-Middle (MITM) attacks, where an attacker could intercept and tamper with communication without detection. When SSL is explicitly disabled, it uses insecure credentials. There are no other obvious signs of malicious patterns or `eval` usage in the provided source code.
Updated: 2026-01-18GitHub→
0
0
Low Cost
nohuiam icon
Sec8

Performs safe batch file operations (move, copy, delete, rename) with pre-validation, automatic backup, and rollback capabilities.

Setup Requirements

  • ⚠️Requires a Node.js environment with file system write permissions for its operations and local SQLite database storage.
  • ⚠️The 'better-sqlite3' dependency is a native module that might require build tools (e.g., Python, C++ compiler) on the host system if pre-built binaries are not available.
  • ⚠️Relies on a modern Node.js version due to 'type: module' and ES2022 target.
Verified SafeView Analysis
The server implements several safety features: input validation via Zod, rate limiting on HTTP endpoints, explicit CORS configuration, and a 'protected paths' list to prevent modifications to critical system directories. Batch operations are automatically backed up, and a rollback feature is provided. The InterLock UDP communication uses a whitelist for signals. However, the WebSocket and InterLock UDP broadcast mechanisms do not appear to have explicit authentication/authorization for receiving messages, which could be a concern if deployed in an untrusted network environment.
Updated: 2026-01-17GitHub→
0
0
Medium Cost
VN-78 icon
Sec7

This server provides tools for comprehensive analysis and reading of CSV files via the Model Context Protocol (MCP).

Setup Requirements

  • ⚠️Requires Python 3.14 or newer.
  • ⚠️Uses `uv` for dependency management, which needs to be installed.
  • ⚠️CSV files must exist locally on the server's filesystem and their paths must be provided to the tools; the server does not handle file uploads.
Review RequiredView Analysis
The `analyze_csv_data` and `read_csv_data` tools accept an arbitrary `file_path` argument and directly use it with `pandas.read_csv`. While basic checks for file existence and `.csv` extension are present, there is no explicit sanitization or restriction to a specific directory. This could potentially allow an attacker (via a compromised or malicious MCP client) to read arbitrary CSV files or even non-CSV files (if the `.csv` check is bypassed or a non-CSV file is named `.csv`) from the server's filesystem, leading to information disclosure or denial-of-service by attempting to read very large or special files. The server runs on localhost by default, which mitigates some external exposure, but internal misuse by an agent remains a risk.
Updated: 2025-11-19GitHub→
PreviousPage 282 of 760Next