Stop Searching. Start Trusting.

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

Browse Directory Why 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)

Medium Cost

tilt-mcp

by 0xBigBoss

Sec9

Integrates Tilt development workflows with AI assistants via the Model Context Protocol (MCP) for real-time status, resource management, and log access.

Setup Requirements

⚠️Requires Bun 1.3+ runtime to be installed.
⚠️Requires Tilt CLI v0.35.0 or later to be installed and in PATH.
⚠️The TILT_PORT environment variable is mandatory and must be configured (e.g., in .mcp.json or shell).

Verified SafeView Analysis

The server employs robust security practices. CLI interactions use `node:child_process.spawn` with argument arrays, explicitly avoiding shell interpolation to prevent command injection. All user inputs (resource names, filters, Tiltfile args, regex search queries/flags) are rigorously validated using Zod schemas with strong regex patterns to disallow malicious characters and patterns. Connection details (TILT_PORT, TILT_HOST) are configured via environment variables and not exposed as tool parameters, mitigating cross-instance attack vectors. The `tilt_dump` tool, which could expose large raw engine state, is intentionally not registered for MCP use. Regex flags in `tilt_logs` are validated, and the global flag is disallowed to prevent stateful matching issues. Overall, strong focus on preventing common injection and abuse vectors.

Updated: 2025-11-22GitHub→

Medium Cost

tamina-mcp

by buddl

Sec1

Enables AI assistants to search for local businesses and make reservations/bookings through the Tamina platform via the Model Context Protocol (MCP).

Setup Requirements

⚠️Requires an active account on Tamina.app and valid authentication credentials.
⚠️Users must provide a verified credit card on Tamina.app for contact information verification (no charge incurred).
⚠️Requires an AI platform that supports the Model Context Protocol (MCP).

Verified SafeView Analysis

The provided 'SOURCE CODE' is limited to the `README.md` file. Without access to the actual server-side code for the Tamina MCP server, it is impossible to perform a meaningful security audit for vulnerabilities such as `eval` usage, obfuscation, network risks, hardcoded secrets, or malicious patterns. The score reflects the complete lack of visibility into the server's implementation.

Updated: 2025-12-14GitHub→

Medium Cost

mcp-gcalendar

by moodjx

Sec9

Provides Google Calendar integration for AI assistants to manage events, check availability, and search calendars.

Setup Requirements

⚠️Python 3.13+ required.
⚠️Requires `uv` package manager for installation and running.
⚠️Requires a Google Cloud Project with Calendar API enabled and OAuth2 credentials configured.

Verified SafeView Analysis

The server implements strong security practices by using per-request OAuth2 tokens, meaning user credentials are not stored server-side. It explicitly disables automatic token refresh from the Google API client library to ensure the client is responsible for valid access tokens. Server-side OAuth client credentials are loaded securely from environment variables. Input validation is performed using Pydantic schemas and custom validation functions (e.g., for timezones and emails). There is no use of `eval` or obvious obfuscation. Comprehensive error handling is in place, providing informative but non-revealing messages for authentication failures.

Updated: 2025-12-10GitHub→

Medium Cost

mcp-toolsets

by Maximophone

Sec3

A modular Model Context Protocol (MCP) server that provides AI tools for filesystem operations, email, messaging, and productivity platforms (Notion, LinkedIn).

Setup Requirements

⚠️Requires manual configuration of the '.env' file for API keys/tokens for each enabled toolset (Gmail, Discord, Notion, LinkedIn), as the default 'MCP_API_KEY' is not secure for production.
⚠️Gmail integration setup involves creating a Google Cloud project, enabling the Gmail API, creating OAuth 2.0 Desktop app credentials, and downloading 'credentials.json' for the initial browser-based OAuth flow.
⚠️Discord integration requires creating a Discord bot, enabling 'Message Content Intent' and 'Server Members Intent' in the Discord Developer Portal, copying the bot token, and inviting the bot to servers with specific permissions.
⚠️LinkedIn integration uses an unofficial API which is prone to account restrictions. Authentication can be challenging, recommending browser cookie extraction (Brave, Chrome, Firefox, Edge) or manual 'li_at' and 'JSESSIONID' configuration over username/password due to 2FA/CAPTCHA risks.
⚠️Python dependencies must be installed via `pip install -r requirements.txt`.

Review RequiredView Analysis

The 'system' toolset includes highly dangerous capabilities like 'run_command' (executes arbitrary shell commands), 'execute_python' (executes arbitrary Python code), 'save_file' (writes to arbitrary paths within BASE_DIR), and 'persistent_shell' (maintains an interactive shell session). While these tools are explicitly marked 'safe=False' and require API key authentication, a compromised API key or misuse by an AI agent could lead to severe system compromise, data loss, or unauthorized access. File operations are sandboxed to 'MCP_BASE_DIR', which mitigates some direct filesystem risks but does not eliminate the danger of code execution. The use of pickle for Gmail tokens is generally safe as it's for self-generated, controlled data.

Updated: 2025-12-15GitHub→

Medium Cost

jobsearch-mcp-server

by hanjianchun

Sec1

This server provides AI-powered job search assistance by fetching job listings and matching them against a user's resume.

Setup Requirements

⚠️Requires manual removal and secure configuration of hardcoded API key and session cookies (CRITICAL security fix).
⚠️Requires ChromeDriver executable to be manually downloaded and its path configured (currently hardcoded for Windows).
⚠️Requires a valid DeepSeek/OpenAI API key for LLM services.

Review RequiredView Analysis

CRITICAL: The `llm.py` file contains a hardcoded OpenAI/DeepSeek API key. The `listjob.py` file contains a hardcoded user session cookie string for `zhipin.com` and a hardcoded proxy server address. These hardcoded credentials represent severe security vulnerabilities, exposing sensitive access tokens and potential user session data. The `chromedriver_path` is also hardcoded to a Windows-specific location, which is a misconfiguration risk.

Updated: 2025-12-11GitHub→

Medium Cost

github-trendings-mcp

by mainhanu

Sec8

An MCP server enabling AI assistants to fetch, track, and analyze trending GitHub repositories with language filtering and structured insights for technology trend analysis.

Setup Requirements

⚠️Requires global NPM installation (`npm install -g github-trending-mcp`).
⚠️Designed to integrate with AI assistants supporting Model Context Protocol (MCP), such as VS Code Copilot.
⚠️If behind a network proxy, requires `HTTPS_PROXY`/`HTTP_PROXY` environment variables or `--proxy` argument to be configured.

Verified SafeView Analysis

The server uses web scraping (`cheerio`, `undici`) to fetch data from GitHub Trending, which relies on GitHub's page structure remaining consistent. It reads proxy settings from standard environment variables (`HTTPS_PROXY`, `HTTP_PROXY`) or a command line argument (`--proxy`), which is common practice. File system operations are confined to a user-specific cache directory (`~/.github-trending-mcp/cache.json`) and do not pose a system-wide risk. No 'eval', hardcoded secrets, or obvious malicious patterns were found. The tool's reliance on scraping could make it brittle if GitHub's page structure changes.

Updated: 2026-01-19GitHub→

Low Cost

MyTaskly-mcp

by zaka265-star

Sec8

Provides a Model Context Protocol (MCP) server for task and note management, optimized for mobile UIs and voice interaction.

Setup Requirements

⚠️Requires a separate MyTaskly FastAPI backend server to be running and accessible.
⚠️The `JWT_SECRET_KEY` environment variable on the MCP server must EXACTLY match the `SECRET_KEY` used by the FastAPI backend server.
⚠️Requires Python 3.10 or higher.

Verified SafeView Analysis

The server implements a robust dual-JWT authentication mechanism, validating incoming MCP tokens against a specified audience and then generating new JWTs for the FastAPI backend using a shared secret key. This shared `JWT_SECRET_KEY` is a critical security dependency, as explicitly highlighted in the documentation; it must be securely managed and consistent across both services. The `X-API-Key` header provides an additional layer of authentication for the MCP server's communication with the FastAPI backend. Debug logging of JWT token details in `src/auth.py` is present, which should be disabled or carefully managed in production to prevent exposure of sensitive information. An older/deprecated file, `src/client.py`, contains a hardcoded `secret_key` for JWT generation; while the primary entry point (`main.py`) appears to use the refactored, secure client structure, the presence of this file is a potential vulnerability if it were to be inadvertently used.

Updated: 2026-01-19GitHub→

Low Cost

notes

by cardmagic

Sec4

Search, browse, create, and delete Apple Notes with fuzzy matching, designed for integration with Claude Code.

Setup Requirements

⚠️Requires macOS to function (reads Apple Notes database).
⚠️Requires manual 'Full Disk Access' for the terminal/IDE in System Settings to read Apple Notes data.
⚠️Optional: Install 'poppler' (`brew install poppler`) for PDF text extraction features.

Review RequiredView Analysis

The server uses `execSync` for AppleScript automation (create/delete notes) and PDF text extraction (`pdftotext`). While the `applescript.ts` file includes an `escapeAppleScript` function, direct execution of external commands with user-controlled input (even if sanitized) via `execSync` is inherently risky and could lead to arbitrary command injection if the escaping mechanism is flawed or bypassed. This is a critical security concern.

Updated: 2026-01-18GitHub→

High Cost

Quizard-MCP-Server

by Quizzardd

Sec8

This server functions as a Classroom Quiz Generator, leveraging AI to create, revise, and publish quizzes for educators by extracting content from learning materials and integrating with a backend classroom service.

Setup Requirements

⚠️Requires Google Cloud Project setup and service account credentials with Storage Object Viewer and OIDC token generation permissions.
⚠️A separate backend classroom service must be running and accessible at the specified BACKEND_BASE_URL.
⚠️Python dependencies (requests, google-cloud-storage, PyPDF2, python-dotenv, fastmcp, google-auth-oauthlib, google-auth) must be installed.

Verified SafeView Analysis

The server uses Google Cloud Storage for content retrieval and Google OIDC tokens for backend authentication, which are generally secure practices. It loads environment variables for sensitive paths. The `read_content_file_from_URL` function handles various URL schemes, including direct HTTP requests. While the intended use appears to be limited to Google Cloud Storage URLs by the tool descriptions, if the `file_url` input to this function were to be manipulated by an external, untrusted source, it could potentially lead to SSRF vulnerabilities, although the agent instructions for its use typically prevent this. No 'eval' or explicit shell execution is observed.

Updated: 2025-11-23GitHub→

Medium Cost

codemoder

by diodeinc

Sec5

Acts as an MCP proxy to enable AI models to write and execute JavaScript code, allowing them to chain multiple tool calls in a single execution.

Setup Requirements

⚠️Requires a downstream MCP server to proxy.
⚠️Requires Rust toolchain to build.
⚠️Interaction requires understanding of the MCP protocol for client integration.

Verified SafeView Analysis

The core functionality of codemoder is to execute arbitrary JavaScript code provided as input to its `execute_tools` tool. This is achieved using the `rquickjs` (QuickJS) runtime and its `eval` equivalent. While `rquickjs` generally provides a sandboxed environment, the JavaScript code has direct access to call any of the downstream MCP server's tools via the `tools` object. This design inherently poses a significant security risk: if a malicious actor or a compromised AI model provides malicious JavaScript code, it could execute arbitrary commands or logic through the downstream tools on the host system where codemoder is running. It is critical to only feed code from trusted sources.

Updated: 2026-01-18GitHub→

Low Cost

robinsonai-mcp-servers

by christcr2012

Sec7

Enhances Augment Code with a multi-server architecture, integrating local LLMs and a vast toolkit for autonomous, cost-effective AI development workflows.

Setup Requirements

⚠️Requires Ollama installation and pulling multiple large LLM models (up to 19GB each), leading to a significant disk space (60 GB) and RAM (16 GB minimum) commitment.
⚠️Manual configuration of Augment Code's VS Code `settings.json` is necessary for each of the five MCP servers, often requiring absolute paths which can be OS-specific.
⚠️Full functionality of the 'Robinson's Toolkit MCP' depends on providing various API keys (e.g., `GITHUB_TOKEN`, `VERCEL_TOKEN`, `NEON_API_KEY`) as environment variables.

Verified SafeView Analysis

The system integrates deeply with various external APIs (GitHub, Vercel, Neon, Stripe, etc.) and can modify local code. Users must securely manage API keys/tokens (via environment variables) and exercise caution when allowing agents to autonomously generate or modify code, especially in production environments. Local Ollama usage reduces external network exposure for core LLM tasks.

Updated: 2025-11-16GitHub→

Medium Cost

COMP4952_MCPWebAppServer

by BetterLordWilliam

Sec9

A multi-component .NET web application integrating an AI model (gpt-4o-mini) for content processing, featuring distinct API and MVC client components, likely for a 'beverage' system.

Setup Requirements

⚠️Requires Azure AI access/authentication (e.g., API Key, Azure AD) for the specified AI endpoint (https://models.inference.ai.azure.com).
⚠️This is a multi-project .NET solution, requiring individual compilation/running of components or a solution-level build/run command.
⚠️The MCPWebControllerApi uses a local SQLite database ('beverages.sqlite') which may require initial setup or migration.

Verified SafeView Analysis

No critical security risks like 'eval', obfuscation, or hardcoded sensitive secrets were found. The 'AllowedHosts: *' setting is common in development but should be restricted for production deployments.

Updated: 2025-11-28GitHub→

PreviousPage 466 of 760Next