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
openpharma-org icon

cdc-mcp

by openpharma-org

Sec4

Provides a unified interface to access a wide range of public health data from the CDC's Socrata Open Data API, covering disease prevalence, risk factors, and real-time surveillance.

Setup Requirements

  • ⚠️Requires Node.js version 18 or higher.
  • ⚠️Requires TypeScript to be installed and the project to be built (`tsc`) before running the generated JavaScript file.
  • ⚠️A `CDC_APP_TOKEN` environment variable is highly recommended for enhanced API rate limits (1,000 requests/hour vs. a shared pool), impacting performance without it.
Review RequiredView Analysis
The `search_dataset` method directly injects user-provided `where_clause` into a SoQL (Socrata Query Language) query without sanitization. This is a significant injection vulnerability, allowing for arbitrary manipulation of the query against the external CDC API, potentially leading to unintended data access, exfiltration, or denial of service on the Socrata platform. While it does not directly lead to RCE on the server, it is a critical flaw in data interaction.
Updated: 2025-12-22GitHub→
0
0
Medium Cost
zobra-auto icon

valeria-mcp-server

by zobra-auto

Sec8

Provide a Microservices Communication Platform (MCP) API for managing barber shop appointments, including catalog search, barber resolution, and Google Calendar integration for scheduling and availability checks.

Setup Requirements

  • ⚠️Requires a Google Cloud Project with the Calendar API enabled and a Service Account configured with appropriate permissions. Service Account credentials must be provided via `GOOGLE_SA_KEYFILE`, `GOOGLE_APPLICATION_CREDENTIALS_JSON`, or `GOOGLE_SA_JSON_BASE64` environment variables.
  • ⚠️Requires `data/barbers.json` and `data/business_hours.json` files to exist and be correctly formatted in the project root or specified via environment variables.
  • ⚠️An `API_KEY` environment variable is required for all API authentication.
Verified SafeView Analysis
The server implements API key authentication with `timingSafeCompare` to prevent timing attacks, and includes rate limiting. It securely loads Google Service Account credentials from environment variables or files. The `calendar.create` action includes a critical conflict check (`SLOT_OCCUPIED`) before adding events, preventing double-bookings. CORS is enabled globally, which might require careful configuration in a production environment to restrict origins.
Updated: 2025-11-24GitHub→
0
0
Medium Cost
labki-org icon

mw-mcp-server

by labki-org

Sec9

Integrates MediaWiki with large language models to provide semantic search, conversational AI, and structured wiki data access through LLM-callable tools.

Setup Requirements

  • ⚠️Requires OpenAI API Key (Paid Service).
  • ⚠️Requires PostgreSQL with pgvector extension (handled by provided Docker image).
  • ⚠️The 'WIKI_CREDS' environment variable requires a JSON string with specific secrets for each wiki, which can be complex to set up for multi-tenant environments.
Verified SafeView Analysis
The server employs robust security practices including: bidirectional JWT authentication with strong validation (scopes, issuers, audiences, `wiki_id`), Pydantic models for strict input/output validation (e.g., `extra="forbid"`), explicit tool dispatching with allow-listing, and safeguards against path traversal. SQLAlchemy ORM is used for database interactions mitigating common SQL injection risks. A global exception handler prevents sensitive information leakage. The primary security dependency lies in the MediaWiki Semantic MediaWiki (SMW) extension's handling of raw `#ask` queries forwarded by the server, though the server adds validation for known properties/categories to mitigate LLM hallucinations. No direct `eval` or `exec` on user input is observed.
Updated: 2026-01-18GitHub→
0
0
Medium Cost
gouthamgo icon

mcp-server

by gouthamgo

Sec9

Provides Model Context Protocol (MCP) server implementations to integrate AI clients (like Claude) with various external services, including a virtual try-on shopping assistant, Salesforce Marketing Cloud, and a comprehensive ecommerce suite.

Setup Requirements

  • ⚠️Requires Node.js v18+ (v20+ recommended) for Node.js servers or Python 3.8+ for Python servers, crucial for `fetch` API support.
  • ⚠️Requires manually populating `.env` files with API keys and other credentials (e.g., Salesforce credentials, HuggingFace token) for specific tools.
  • ⚠️Requires absolute paths for server command and arguments when configuring with Claude Desktop or other MCP clients.
  • ⚠️In-memory data stores in demo ecommerce servers must be replaced with a real database for production use.
Verified SafeView Analysis
The project extensively documents security best practices in `SECURITY.md`, covering input validation, secrets management (environment variables, `.gitignore`), API security (rate limiting, HTTPS), network security (SSRF, domain allowlisting), and deployment security (Docker, non-root users). Tool execution includes generic `try-catch` blocks and uses `McpError` to prevent raw error leakage. However, actual security relies on users implementing the documented best practices and replacing in-memory data stores with secure databases in production, as some demo servers currently use in-memory state.
Updated: 2025-11-18GitHub→
0
0
High Cost
No-Smoke icon
Sec8

Provides semantic search and code vectorization capabilities using Qdrant vector database and various embedding providers, suitable for AI-assisted development and knowledge base management.

Setup Requirements

  • ⚠️Requires Qdrant and Ollama Docker containers (or local installations) to be running.
  • ⚠️The 'nomic-embed-text' Ollama model (or specified embedding model) must be pulled if using Ollama.
  • ⚠️Requires API keys (e.g., OPENAI_API_KEY, COHERE_API_KEY, VOYAGE_API_KEY) if not using Ollama as the embedding provider.
Verified SafeView Analysis
The server includes path validation (`validatePath`) to prevent directory traversal attacks when indexing codebases. For HTTP transport, the README provides critical security warnings regarding deployment (HTTPS, authentication, firewalls, reverse proxy). Rate limiting is implemented for the HTTP server to mitigate abuse. The `containsSecrets` function is a feature to detect potential secrets within *indexed user codebases*, not within the server's own codebase, which is a good safety measure. No 'eval' or obvious hardcoded secrets were found in the provided source.
Updated: 2026-01-17GitHub→
0
0
Low Cost
bobtherdsman icon

RDSMCP

by bobtherdsman

Sec8

Assess SQL Server instances for migration compatibility to AWS RDS.

Setup Requirements

  • ⚠️Requires Microsoft ODBC Driver 18 for SQL Server (Windows-specific, external download).
  • ⚠️Requires specific SQL Server permissions (VIEW SERVER STATE, VIEW ANY DEFINITION, access to system databases, sp_MSforeachdb).
  • ⚠️Requires Python 3.8 or higher.
Verified SafeView Analysis
The tool connects to specified SQL Server instances, requiring network access. It uses `Trusted_Connection=yes` or explicit credentials. The connection string uses `Encrypt=yes;TrustServerCertificate=yes`, where `TrustServerCertificate=yes` bypasses certificate validation, potentially allowing MITM attacks if not used in a secure, trusted environment. The core SQL query utilizes `sp_MSforeachdb`, which can be a SQL injection risk if the inner queries were dynamically constructed from untrusted user input; however, in this implementation, the inner queries are fixed and use `?` only for database names derived internally from `sys.databases`, mitigating this specific risk. No hardcoded credentials or `eval` statements were found.
Updated: 2025-12-02GitHub→
0
0
Medium Cost
reouno icon

gemini-mcp

by reouno

Sec9

An MCP (Model Context Protocol) server that provides a standardized interface for accessing Google Gemini AI models for text generation.

Setup Requirements

  • ⚠️Requires a Google Gemini API Key (paid service, obtained from AI Studio).
  • ⚠️Initial deployment to Google Cloud Run requires significant manual GCP project setup (creating project, linking billing, gcloud CLI authentication).
  • ⚠️Automated deployment via GitHub Actions requires setting up multiple GitHub Secrets with values derived from GCP resources.
Verified SafeView Analysis
The server correctly uses environment variables for sensitive API keys and validates input with Zod schemas, mitigating common injection risks. Deployment to Google Cloud Run leverages Workload Identity Federation for secure GitHub Actions integration, limiting permissions to the specific repository. No 'eval' or direct malicious patterns were found. The use of 'any' casts for Google Gen AI SDK responses is a minor type safety concern but not a security vulnerability.
Updated: 2025-11-19GitHub→
0
0
Low Cost
olahiselius icon

PigelloMockAPI

by olahiselius

Sec8

This Model Context Protocol (MCP) server integrates with a Pigello Mock API to manage property information, cases, tenants, and components, primarily for AI assistant interaction via the Model Context Protocol.

Setup Requirements

  • ⚠️Requires the Pigello Mock API (a separate .NET project) to be running on 'http://localhost:5059' (or 'https://localhost:7293' for HTTPS) before starting the MCP server.
  • ⚠️Manual JSON configuration is required in VS Code or Codespaces (under 'mcpServers') to integrate this MCP server with GitHub Copilot.
  • ⚠️This is an internal prototype; production readiness would require further hardening beyond the mock setup.
Verified SafeView Analysis
The server design as a proxy to a mock API is generally safe for development/prototyping. For production, careful configuration of CORS is necessary to restrict origins. The base URL for the mock API is configurable via environment variables, which is good practice. No 'eval', obfuscation, or obvious hardcoded sensitive secrets were found in the provided code snippets.
Updated: 2025-11-19GitHub→
0
0
Low Cost
WonkyBilly icon
Sec9

Provides real-time air temperature data for Singapore as a Model Context Protocol (MCP) tool.

Verified SafeView Analysis
The server makes a `fetch` request to `https://api-open.data.gov.sg/v2/real-time/api/air-temperature`. While direct external API calls are a potential network risk, the URL is hardcoded and appears to be a legitimate government service. No `eval`, hardcoded secrets, or unusual malicious patterns were found. Uses `process.env.PORT` for configuration, which is good practice.
Updated: 2026-01-18GitHub→
0
0
Medium Cost
u9401066 icon
Sec8

Integrates PubMed literature search capabilities into AI agents, particularly for Microsoft Copilot Studio, enabling comprehensive scientific research and information retrieval within Microsoft 365 applications.

Setup Requirements

  • ⚠️Requires an NCBI Email for API access (free but essential for NCBI's E-utilities policy).
  • ⚠️For Microsoft Copilot Studio integration, a publicly accessible HTTPS URL with a valid SSL certificate is required (often achieved via Nginx reverse proxy or temporary tunnels like ngrok).
  • ⚠️Python 3.8 or higher is required to run the server.
Verified SafeView Analysis
The server uses environment variables for sensitive credentials (NCBI API Key) and supports deployment with Nginx for HTTPS termination, rate limiting, and security headers. The internal HTTP API (`/api/*` endpoints) for MCP-to-MCP communication uses `allow_origins=['*']`, which is a common pattern for trusted internal integrations or services behind a proxy. However, this requires careful configuration of the ingress layer (e.g., Nginx) in production environments to prevent unintended external exposure. The server generally binds its internal HTTP API to `127.0.0.1` by default, reducing direct external exposure.
Updated: 2026-01-19GitHub→
0
0
Medium Cost
Zutavern icon
Sec8

Enables AI assistants to interact with the Belbo Booking API for managing appointments, customers, services, and other business operations.

Setup Requirements

  • ⚠️Requires Node.js 18 or higher to run.
  • ⚠️Requires a Belbo account with API access enabled and an API token.
  • ⚠️Configuration needs to be provided via environment variables (BELBO_BUSINESS, BELBO_TOKEN) or dynamically through the 'belbo_configure' tool.
Verified SafeView Analysis
The server correctly uses HTTPS for all API communication and explicitly avoids logging API tokens to console output. API tokens are passed as query parameters, which is a common pattern for the Belbo API but generally less secure than 'Bearer' tokens in headers as they can persist in some server logs or proxy caches. The server itself does not introduce arbitrary code execution risks like 'eval'. Input validation for tool arguments is expected to be handled by the MCP SDK based on the defined 'inputSchema', and the internal Belbo client performs checks for configuration existence.
Updated: 2025-12-14GitHub→
0
0
Medium Cost
f114112140-star icon

MCP-Server

by f114112140-star

Sec3

A server to integrate Google Gemini AI with Unity, allowing scene generation based on two camera inputs and outputting a JSON array of scene objects.

Setup Requirements

  • ⚠️Requires manual installation of dependencies listed in `requirement.txt`.
  • ⚠️The Google Gemini API key is hardcoded in `mcp_server.py` and must be manually replaced within the source code.
  • ⚠️The AI prompt for scene generation is hardcoded and requires modifying `mcp_server.py` to change.
Review RequiredView Analysis
The Google API key is hardcoded directly into `mcp_server.py`, posing a severe security risk as it can be easily exposed. The `CORSMiddleware` is configured to allow all origins (`allow_origins=["*"]`), which is generally unsafe for production environments as it can lead to Cross-Site Request Forgery (CSRF) vulnerabilities.
Updated: 2025-11-25GitHub→
PreviousPage 471 of 760Next